El modelo base es una regresión logística dentro de un Pipeline (Pedregosa et al., 2011).
Se evalúa frente a tres referencias de dificultad creciente.
| Referencia | Qué representa |
|---|---|
DummyClassifier | predice siempre la clase mayoritaria |
| Persistencia | dentro de quince días el río estará como hoy |
| Curva de gasto sólido | el método estándar del dominio |
La persistencia es la referencia exigente en series fluviales. Con una autocorrelación de 0,92 a quince días, un modelo puede alcanzar un acierto elevado limitándose a repetir el estado actual, de modo que superarla es lo que acredita capacidad predictiva.
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import scipy.stats as st
import json, sys
sys.path.insert(0, ".")
from src.config import (SEED, DATOS_PROC, FIGURAS, ESTACION_OBJETIVO,
FECHA_CORTE_TEST, GAP_DIAS, HORIZONTE,
fijar_semillas, estilo_graficos)
fijar_semillas(SEED)
estilo_graficos()
dfx = pd.read_csv(DATOS_PROC / "features.csv", parse_dates=["fecha"])
with open(DATOS_PROC / "parametros.json", encoding="utf-8") as f:
par = json.load(f)
Q_EF = par["Q_EF"]
LIM = tuple(par["LIM"])
A_OF, B_OF = par["A_OF"], par["B_OF"]
BASE_COLS, DOM_COLS = par["BASE_COLS"], par["DOM_COLS"]
TODAS = BASE_COLS + DOM_COLS
print(f"{len(dfx):,} filas, {len(TODAS)} predictores")
print(f"Q_ef = {Q_EF:,.0f} m³/s banda = [{LIM[0]:,.0f} - {LIM[1]:,.0f}]")31,475 filas, 43 predictores
Q_ef = 9,954 m³/s banda = [8,461 - 11,448]
from sklearn.pipeline import Pipeline
from sklearn.impute import SimpleImputer
from sklearn.preprocessing import StandardScaler
from sklearn.linear_model import LogisticRegression
from sklearn.dummy import DummyClassifier
from sklearn.model_selection import TimeSeriesSplit, cross_validate, learning_curve
from sklearn.metrics import (accuracy_score, precision_score, recall_score, f1_score,
roc_auc_score, average_precision_score, confusion_matrix,
brier_score_loss, classification_report,
RocCurveDisplay, PrecisionRecallDisplay, ConfusionMatrixDisplay)
from sklearn.calibration import calibration_curve
corte = pd.Timestamp(FECHA_CORTE_TEST)
inicio_test = corte + pd.Timedelta(days=GAP_DIAS)
dm = dfx.dropna(subset=["y"])
TR = dm[dm.fecha < corte]
TE = dm[dm.fecha >= inicio_test]
Xtr, ytr = TR[TODAS], TR.y.astype(int)
Xte, yte = TE[TODAS], TE.y.astype(int)
print(f"Train: {len(Xtr):,} obs. prevalencia {ytr.mean():.2%}")
print(f"Test: {len(Xte):,} obs. prevalencia {yte.mean():.2%}")Train: 25,364 obs. prevalencia 25.67%
Test: 6,066 obs. prevalencia 24.86%
# La validacion cruzada respeta la estructura temporal
cv = TimeSeriesSplit(n_splits=5, gap=GAP_DIAS)
print(f"TimeSeriesSplit(n_splits=5, gap={GAP_DIAS})")
for i, (a, b) in enumerate(cv.split(Xtr), 1):
print(f" pliegue {i}: train n = {len(a):>6,} validación n = {len(b):>6,}")TimeSeriesSplit(n_splits=5, gap=30)
pliegue 1: train n = 4,199 validación n = 4,227
pliegue 2: train n = 8,426 validación n = 4,227
pliegue 3: train n = 12,653 validación n = 4,227
pliegue 4: train n = 16,880 validación n = 4,227
pliegue 5: train n = 21,107 validación n = 4,227
def modelo_base():
return Pipeline([
("imp", SimpleImputer(strategy="median")),
("esc", StandardScaler()),
("clf", LogisticRegression(max_iter=5000, class_weight="balanced", random_state=SEED)),
])
class Persistencia:
"""El estado de hoy se mantiene dentro de h dias."""
def __init__(self, lim): self.lim = lim
def fit(self, X, y=None): return self
def predict(self, X): return X[ESTACION_OBJETIVO].between(*self.lim).astype(int).values
def predict_proba(self, X):
p = self.predict(X).astype(float)
return np.column_stack([1 - p, p])
class GastoSolido:
"""Clasifica por la carga estimada con la curva oficial."""
def __init__(self, a, b, lim): self.a, self.b, self.lim = a, b, lim
def fit(self, X, y=None): return self
def predict(self, X):
qs = self.a * X[ESTACION_OBJETIVO].clip(lower=1) ** self.b
lo, hi = self.a * self.lim[0]**self.b, self.a * self.lim[1]**self.b
return ((qs >= lo) & (qs <= hi)).astype(int).values
def predict_proba(self, X):
p = self.predict(X).astype(float)
return np.column_stack([1 - p, p])
def evaluar(modelo, Xa, ya, Xb, yb, nombre, cols):
es_dominio = isinstance(modelo, (Persistencia, GastoSolido))
modelo.fit(Xa if es_dominio else Xa[cols], ya)
Xp = Xb if es_dominio else Xb[cols]
yp = modelo.predict(Xp)
ys = modelo.predict_proba(Xp)[:, 1]
return {"modelo": nombre,
"accuracy": accuracy_score(yb, yp),
"precision": precision_score(yb, yp, zero_division=0),
"recall": recall_score(yb, yp, zero_division=0),
"f1": f1_score(yb, yp, zero_division=0),
"roc_auc": roc_auc_score(yb, ys),
"pr_auc": average_precision_score(yb, ys),
"brier": brier_score_loss(yb, np.clip(ys, 0, 1))}, yp, ys4.1 Desempeño en el conjunto de prueba¶
XtrF = TR[TODAS + [ESTACION_OBJETIVO]]
XteF = TE[TODAS + [ESTACION_OBJETIVO]]
modelos = {
"dummy": DummyClassifier(strategy="most_frequent", random_state=SEED),
"persistencia": Persistencia(LIM),
"gasto sólido": GastoSolido(A_OF, B_OF, LIM),
"regresión logística": modelo_base(),
}
resultados, preds = [], {}
for nombre, m in modelos.items():
r, yp, ys = evaluar(m, XtrF, ytr, XteF, yte, nombre, TODAS)
resultados.append(r)
preds[nombre] = (yp, ys)
comp = pd.DataFrame(resultados).set_index("modelo")
print(comp.round(4).to_string())
print()
print(f"Ganancia de PR-AUC sobre el dummy: "
f"{comp.loc['regresión logística','pr_auc'] - comp.loc['dummy','pr_auc']:+.4f}")
print(f"Ganancia de PR-AUC sobre la persistencia: "
f"{comp.loc['regresión logística','pr_auc'] - comp.loc['persistencia','pr_auc']:+.4f}") accuracy precision recall f1 roc_auc pr_auc brier
modelo
dummy 0.7514 0.0000 0.0000 0.0000 0.5000 0.2486 0.2486
persistencia 0.8826 0.7639 0.7639 0.7639 0.8429 0.6423 0.1174
gasto sólido 0.8826 0.7639 0.7639 0.7639 0.8429 0.6423 0.1174
regresión logística 0.8820 0.6947 0.9370 0.7979 0.9588 0.8717 0.0899
Ganancia de PR-AUC sobre el dummy: +0.6231
Ganancia de PR-AUC sobre la persistencia: +0.2295
4.2 Validación cruzada y contraste de sobreajuste¶
cvres = cross_validate(modelo_base(), Xtr, ytr, cv=cv,
scoring=["accuracy", "f1", "roc_auc", "average_precision"],
return_train_score=True)
tabla_cv = pd.DataFrame({
"pliegue": range(1, 6),
"train_roc_auc": cvres["train_roc_auc"],
"val_roc_auc": cvres["test_roc_auc"],
"val_f1": cvres["test_f1"],
"val_pr_auc": cvres["test_average_precision"]})
print(tabla_cv.round(4).to_string(index=False))
print(f"\nROC-AUC en validación: {cvres['test_roc_auc'].mean():.4f} ± {cvres['test_roc_auc'].std():.4f}")C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
pliegue train_roc_auc val_roc_auc val_f1 val_pr_auc
1 0.9835 0.9191 0.7637 0.7392
2 0.9669 0.9077 0.7523 0.7661
3 0.9601 0.8992 0.7100 0.7297
4 0.9608 0.9659 0.8127 0.8712
5 0.9626 0.9360 0.7659 0.8274
ROC-AUC en validación: 0.9256 ± 0.0236
# Prueba t pareada sobre los cinco pliegues
dif = cvres["train_roc_auc"] - cvres["test_roc_auc"]
t_stat, p_val = st.ttest_rel(cvres["train_roc_auc"], cvres["test_roc_auc"])
ic = st.t.interval(0.95, len(dif) - 1, loc=dif.mean(), scale=st.sem(dif))
print(f"Diferencia media train - validación: {dif.mean():+.4f}")
print(f"t = {t_stat:.3f} p = {p_val:.4f}")
print(f"IC 95%: [{ic[0]:+.4f}, {ic[1]:+.4f}]")
print("Lectura:", "diferencia significativa" if p_val < 0.05 else "sin evidencia de sobreajuste")Diferencia media train - validación: +0.0412
t = 3.068 p = 0.0374
IC 95%: [+0.0039, +0.0785]
Lectura: diferencia significativa
La prueba detecta una diferencia estadísticamente significativa entre entrenamiento y validación, de modo que conviene separar la significancia de la magnitud del efecto.
La diferencia media de área bajo la curva ROC ronda los cuatro puntos porcentuales, por debajo del rango de 5 a 10 % que las notas del curso consideran aceptable y lejos del 15 a 20 % que señalaría sobreajuste preocupante. La prueba opera sobre diferencias muy consistentes entre pliegues, con dispersión reducida, y con datos de esta naturaleza casi cualquier diferencia sistemática resulta significativa. De ahí que se reporte también el intervalo de confianza, cuyo límite inferior queda próximo a cero.
La curva de aprendizaje y el desempeño sobre el conjunto de prueba, un periodo futuro de dieciséis años nunca visto, completan el diagnóstico.
4.3 Curvas de evaluación¶
yp_l, ys_l = preds["regresión logística"]
fig, ax = plt.subplots(2, 2, figsize=(12.5, 9))
for nombre in ["regresión logística", "persistencia", "gasto sólido"]:
RocCurveDisplay.from_predictions(yte, preds[nombre][1], name=nombre, ax=ax[0, 0])
ax[0, 0].plot([0, 1], [0, 1], "k--", lw=.8)
ax[0, 0].set_title("(a) Curva ROC")
for nombre in ["regresión logística", "persistencia", "gasto sólido"]:
PrecisionRecallDisplay.from_predictions(yte, preds[nombre][1], name=nombre, ax=ax[0, 1])
ax[0, 1].axhline(yte.mean(), color="k", ls="--", lw=.8, label=f"azar ({yte.mean():.2f})")
ax[0, 1].set_title("(b) Curva precisión-exhaustividad")
ax[0, 1].legend(fontsize=7)
ConfusionMatrixDisplay(confusion_matrix(yte, yp_l),
display_labels=["no $Q_{ef}$", "$Q_{ef}$"]).plot(
ax=ax[1, 0], cmap="Blues", colorbar=False)
ax[1, 0].set_title("(c) Matriz de confusión")
pt, pp = calibration_curve(yte, np.clip(ys_l, 0, 1), n_bins=10, strategy="quantile")
ax[1, 1].plot(pp, pt, "o-", label="regresión logística")
ax[1, 1].plot([0, 1], [0, 1], "k--", lw=.8, label="calibración perfecta")
ax[1, 1].set(xlabel="Probabilidad predicha", ylabel="Frecuencia observada",
title="(d) Curva de calibración")
ax[1, 1].legend(fontsize=8)
plt.tight_layout()
plt.savefig(FIGURAS / "evaluacion.png", dpi=150)
plt.show()
print(classification_report(yte, yp_l, target_names=["no Q_ef", "Q_ef"], digits=3))
precision recall f1-score support
no Q_ef 0.976 0.864 0.917 4558
Q_ef 0.695 0.937 0.798 1508
accuracy 0.882 6066
macro avg 0.836 0.900 0.857 6066
weighted avg 0.906 0.882 0.887 6066
4.4 Intervalos de confianza¶
def ic_bootstrap(yt, ys, metrica, n_boot=2000, alfa=0.05):
rng = np.random.default_rng(SEED)
yt, ys = np.asarray(yt), np.asarray(ys)
vals = []
for _ in range(n_boot):
i = rng.integers(0, len(yt), len(yt))
if len(np.unique(yt[i])) < 2:
continue
vals.append(metrica(yt[i], ys[i]))
lo, hi = np.percentile(vals, [100*alfa/2, 100*(1-alfa/2)])
return np.mean(vals), lo, hi
f1_umbral = lambda a, b: f1_score(a, (np.asarray(b) >= 0.5).astype(int), zero_division=0)
filas = []
for nombre in modelos:
_, ys = preds[nombre]
for met, fn in [("ROC-AUC", roc_auc_score), ("PR-AUC", average_precision_score),
("F1", f1_umbral)]:
m, lo, hi = ic_bootstrap(yte, ys, fn)
filas.append({"modelo": nombre, "metrica": met,
"estimacion": m, "IC95_inf": lo, "IC95_sup": hi})
pd.DataFrame(filas).round(4)Loading...
4.5 Curva de aprendizaje y residuos¶
from statsmodels.graphics.tsaplots import plot_acf
from statsmodels.tsa.stattools import acf
tam, sc_tr, sc_va = learning_curve(modelo_base(), Xtr, ytr, cv=cv,
scoring="average_precision",
train_sizes=np.linspace(0.15, 1.0, 8))
fig, ax = plt.subplots(1, 2, figsize=(13, 4.2))
ax[0].plot(tam, sc_tr.mean(1), "o-", label="entrenamiento")
ax[0].plot(tam, sc_va.mean(1), "o-", label="validación")
ax[0].fill_between(tam, sc_tr.mean(1)-sc_tr.std(1), sc_tr.mean(1)+sc_tr.std(1), alpha=.15)
ax[0].fill_between(tam, sc_va.mean(1)-sc_va.std(1), sc_va.mean(1)+sc_va.std(1), alpha=.15)
ax[0].set(xlabel="Tamaño del conjunto de entrenamiento", ylabel="PR-AUC",
title="(a) Curva de aprendizaje")
ax[0].legend()
residuo = yte.values - np.clip(ys_l, 0, 1)
plot_acf(residuo, lags=60, ax=ax[1])
ax[1].set_title("(b) ACF de los residuos")
plt.tight_layout()
plt.savefig(FIGURAS / "aprendizaje_residuos.png", dpi=150)
plt.show()
ar = acf(residuo, nlags=30, fft=True)
print(f"ACF de residuos: lag1 = {ar[1]:.4f} | lag7 = {ar[7]:.4f} | lag30 = {ar[30]:.4f}")C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'pto_salgar_t21' 'pto_salgar_d3' 'pto_salgar_rel'
'san_roque_t10' 'san_roque_d3' 'san_roque_rel' 'tres_cruces_t16'
'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
C:\Users\Qclem\miniconda3\envs\ml_venv\lib\site-packages\sklearn\impute\_base.py:555: UserWarning: Skipping features without any observed values: ['coyongal_t7' 'coyongal_d3' 'coyongal_rel' 'el_banco_t12' 'el_banco_d3'
'el_banco_rel' 'san_roque_t10' 'san_roque_d3' 'san_roque_rel'
'tres_cruces_t16' 'tres_cruces_d3' 'tres_cruces_rel']. At least one non-missing value is needed for imputation with strategy='median'.
warnings.warn(
ACF de residuos: lag1 = 0.9142 | lag7 = 0.5273 | lag30 = 0.2130
4.6 Coeficientes del modelo¶
mdl = modelo_base().fit(Xtr, ytr)
coef = pd.DataFrame({"variable": TODAS, "coeficiente": mdl.named_steps["clf"].coef_[0]})
coef["odds_ratio"] = np.exp(coef.coeficiente)
coef.reindex(coef.coeficiente.abs().sort_values(ascending=False).index).head(15).round(4)Loading...
4.7 Sensibilidad al horizonte¶
El caudal diario de un río grande es muy persistente, de modo que predecir a un día vista se resuelve repitiendo el estado actual. El caudal de hoy está legítimamente disponible hoy, así que no constituye fuga, pero a horizonte corto el ejercicio pierde contenido. El análisis siguiente cuantifica cómo se endurece el problema al alejarse en el tiempo.
filas = []
for h in (1, 7, 15, 30, 60):
z = dfx.copy()
z["y"] = z[ESTACION_OBJETIVO].between(*LIM).astype(float).shift(-h)
z = z.dropna(subset=["y"])
a_ = z[z.fecha < corte]
b_ = z[z.fecha >= inicio_test]
yb = b_.y.astype(int)
for nombre, cols in [("solo lineales en Q", BASE_COLS), ("con dominio", TODAS)]:
m = modelo_base().fit(a_[cols], a_.y.astype(int))
s = m.predict_proba(b_[cols])[:, 1]
filas.append({"horizonte": f"t+{h}", "predictores": nombre,
"PR_AUC": average_precision_score(yb, s)})
pm = Persistencia(LIM)
filas.append({"horizonte": f"t+{h}", "predictores": "persistencia",
"PR_AUC": average_precision_score(yb, pm.predict_proba(b_)[:, 1])})
sens = (pd.DataFrame(filas)
.pivot(index="horizonte", columns="predictores", values="PR_AUC")
.reindex([f"t+{h}" for h in (1, 7, 15, 30, 60)]))
print(sens.round(4).to_string())predictores con dominio persistencia solo lineales en Q
horizonte
t+1 0.9960 0.9621 0.4814
t+7 0.9548 0.8011 0.4834
t+15 0.8717 0.6423 0.4868
t+30 0.6286 0.4395 0.4762
t+60 0.4832 0.2705 0.4591
fig, ax = plt.subplots(figsize=(8, 4.2))
for c in sens.columns:
ax.plot([1, 7, 15, 30, 60], sens[c], "o-", label=c)
ax.set(xlabel="Horizonte de predicción (días)", ylabel="PR-AUC",
title="Dificultad del problema según el horizonte")
ax.legend()
ax.grid(alpha=.3)
plt.tight_layout()
plt.savefig(FIGURAS / "horizonte.png", dpi=150)
plt.show()
4.8 Verificación del desempeño¶
Un desempeño alto admite explicaciones distintas del aprendizaje, por lo que se revisan las posibles fuentes de inflación.
| Verificación | Resultado |
|---|---|
| Fuga de datos | Se auditó en 3.5: el objetivo está en y todo predictor usa información disponible hasta . La serie de transporte quedó excluida por depender de forma determinista del caudal |
| Comparación con línea base | Se realizó contra tres referencias, dos de ellas propias del dominio hidrológico |
| Estructura temporal en la validación | La partición es cronológica, con intervalo de separación de treinta días, y la validación cruzada emplea TimeSeriesSplit |
| Complejidad del conjunto | El registro abarca 86 años y siete estaciones, con régimen bimodal y autocorrelación elevada |
| Dificultad del problema | A horizonte corto resulta trivial, de ahí la elección de quince días y el reporte de la curva completa frente al horizonte |
4.9 Limitación de forma funcional del modelo base¶
El objetivo es una banda de caudales. La regresión logística modela el logit como función monótona de cada predictor, de modo que solo puede afirmar que a mayor caudal corresponde mayor probabilidad. La clase positiva exige en cambio que el caudal se mantenga dentro de un intervalo, condición no monótona que una frontera lineal no representa.
La solución aplicada sustituye el caudal por la distancia al centro de la banda. La transformación vuelve monótono el problema, puesto que a menor distancia corresponde mayor probabilidad, y lo pone al alcance del modelo lineal. La tabla de sensibilidad al horizonte cuantifica la ganancia que aporta esa variable.
La limitación reside en la forma funcional del modelo base. El siguiente entregable debería explorar algoritmos capaces de trazar fronteras no monótonas, entre ellos k-NN, los árboles de decisión y las máquinas de vectores de soporte con núcleo radial. La autocorrelación remanente en los residuos apunta en esa dirección (Müller y Guido, 2016).