|
from sklearn.linear_model import LogisticRegression |
|
from sklearn.ensemble import RandomForestClassifier |
|
from sklearn.metrics import accuracy_score, confusion_matrix |
|
from sklearn.model_selection import train_test_split |
|
import pandas as pd |
|
|
|
|
|
def train_model(df: pd.DataFrame, target_column: str, model_type: str = "logistic_regression"): |
|
""" |
|
Trains a model on the dataset using the specified model type. |
|
|
|
Args: |
|
- df (pd.DataFrame): The dataset. |
|
- target_column (str): The target column for prediction. |
|
- model_type (str): Type of model ('logistic_regression' or 'random_forest'). |
|
|
|
Returns: |
|
- model: The trained model. |
|
""" |
|
X = df.drop(columns=[target_column]) |
|
y = df[target_column] |
|
|
|
|
|
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42) |
|
|
|
if model_type == "logistic_regression": |
|
model = LogisticRegression() |
|
elif model_type == "random_forest": |
|
model = RandomForestClassifier(n_estimators=100) |
|
else: |
|
raise ValueError(f"Unsupported model type: {model_type}") |
|
|
|
|
|
model.fit(X_train, y_train) |
|
|
|
|
|
y_pred = model.predict(X_test) |
|
accuracy = accuracy_score(y_test, y_pred) |
|
cm = confusion_matrix(y_test, y_pred) |
|
|
|
print(f"Model Accuracy: {accuracy}") |
|
print(f"Confusion Matrix:\n{cm}") |
|
|
|
return model |
|
|
|
|
|
def evaluate_model(model, X_test: pd.DataFrame, y_test: pd.Series): |
|
""" |
|
Evaluates a trained model using custom metrics. |
|
|
|
Args: |
|
- model: The trained model. |
|
- X_test (pd.DataFrame): The test feature data. |
|
- y_test (pd.Series): The true labels. |
|
|
|
Returns: |
|
- dict: Dictionary containing custom evaluation metrics. |
|
""" |
|
from sklearn.metrics import classification_report |
|
|
|
y_pred = model.predict(X_test) |
|
report = classification_report(y_test, y_pred, output_dict=True) |
|
|
|
return report |