التعلم الآلي والنماذج التنبؤية للمؤسسات بلغتي Python وR

from sklearn.datasets import load_breast_cancer
from sklearn.ensemble import HistGradientBoostingClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.model_selection import StratifiedKFold
from sklearn.model_selection import cross_validate
from sklearn.pipeline import make_pipeline
from sklearn.preprocessing import StandardScaler

X, y = load_breast_cancer(return_X_y=True)
cv = StratifiedKFold(5, shuffle=True, random_state=42)
metrics = ["roc_auc", "balanced_accuracy", "f1"]

# Benchmark a flexible model against a transparent baseline
models = {
    "logit": make_pipeline(StandardScaler(),
                           LogisticRegression(max_iter=1000)),
    "gbm": HistGradientBoostingClassifier(random_state=42),
}
for name, m in models.items():
    r = cross_validate(m, X, y, cv=cv, scoring=metrics)
    out = [f"{k}={r['test_'+k].mean():.3f}" for k in metrics]
    print(name, *out)

library(tidymodels)
set.seed(42)

# Synthetic data: 1,000 cases, binary outcome
n <- 1000
df <- tibble(x1 = rnorm(n), x2 = rnorm(n), x3 = runif(n))
df$y <- factor(rbinom(n, 1, plogis(df$x1 - df$x2 * df$x3)))
folds <- vfold_cv(df, v = 5, strata = y)

# Transparent baseline vs. flexible model (ranger engine)
models <- list(
  logit = logistic_reg(),
  rf = rand_forest(trees = 500, mode = "classification"))

res <- workflow_set(list(base = y ~ .), models) |>
  workflow_map("fit_resamples", resamples = folds,
    metrics = metric_set(roc_auc, bal_accuracy),
    seed = 42)
rank_results(res, rank_metric = "roc_auc")