adding baires dev interview problem

Author: ivanpenaloza

src/my_project/interviews/baires_dev_interview/dataset/imbalanced_solution.py

@@ -0,0 +1,106 @@
+import pandas as pd
+import xgboost as xgb
+from sklearn.model_selection import StratifiedKFold, GridSearchCV
+from sklearn.metrics import f1_score, make_scorer
+from sklearn.impute import SimpleImputer
+from imblearn.over_sampling import SMOTE
+from imblearn.pipeline import Pipeline
+
+# Set the seed
+seed = 1234
+
+# Load the dataset
+train_data = pd.read_csv("train.csv")
+
+# Remove IDs
+train_data = train_data.drop(columns=['ID'])
+
+# Identify imbalanced dataset
+train_data["Target"].value_counts()
+'''
+>>> train_data["Target"].value_counts()
+Target
+0    6420
+1     316
+Name: count, dtype: int64
+'''
+
+# Impute missing values with the mean
+imputer = SimpleImputer(strategy='mean')
+train_data.iloc[:, :] = imputer.fit_transform(train_data)
+
+# Separate features and target
+X = train_data.drop(columns=["Target"])
+y = train_data["Target"]
+
+# Define the XGBoost model
+model = xgb.XGBClassifier(use_label_encoder=False, eval_metric="logloss")
+
+# Define the hyperparameter grid
+param_grid = {
+    'model__subsample': [0.6, 0.8],
+    'model__colsample_bytree': [0.6, 0.8],
+    'model__n_estimators': [50, 100, 200],
+    'model__max_depth': [3, 5, 7],
+    'model__learning_rate': [0.01, 0.1, 0.2]    
+}
+
+# Define F1 scorer
+f1_scorer = make_scorer(f1_score, pos_label=1)
+
+# Create a pipeline with SMOTE and the XGBoost model
+pipeline = Pipeline([
+    ('smote', SMOTE(random_state=seed)),
+    ('model', model)
+])
+
+# Perform 5-fold stratified cross-validation with grid search
+skf = StratifiedKFold(n_splits=5, shuffle=True, random_state=seed)
+grid_search = GridSearchCV(
+    estimator=pipeline,
+    param_grid=param_grid,
+    scoring=f1_scorer,
+    cv=skf,
+    verbose=1,
+    n_jobs=-1
+)
+
+# Fit the grid search
+grid_search.fit(X, y)
+
+# Print the best parameters and the best F1 score
+print("Best parameters: {}".format(grid_search.best_params_))
+print("Best F1 score: {}".format(grid_search.best_score_))
+
+score = 100*grid_search.best_score_
+'''
+This is the solution you are lookin at since test.csv does not have 
+Target column
+'''
+print('Solution: {}'.format(score))
+
+# Train the best model on the full training set
+best_model = grid_search.best_estimator_
+best_model.fit(X, y)
+
+# Load the test dataset
+test_data = pd.read_csv("test.csv")
+
+# Remove the 'ID' column
+test_ids = test_data['ID']  # Save the IDs for the output file
+test_data = test_data.drop(columns=['ID'])
+
+# Predict using the trained model
+predictions = best_model.predict(test_data)
+
+# Create a DataFrame for the predictions
+output = pd.DataFrame({
+    'ID': test_ids,
+    'Target': predictions
+})
+
+# Save the predictions to a CSV file
+output.to_csv("predictions.csv", index=False)
+
+print("Predictions saved to predictions.csv")
+