estudIA-MCP

from sklearn.ensemble import RandomForestRegressor from sklearn.model_selection import train_test_split from sklearn.metrics import mean_squared_error import pandas as pd import joblib df = pd.read_csv('business_growth_dataset.csv') # Normalize 'formalization_level' column df = df.drop('sector', axis=1) df = df.drop('business_id', axis=1) df = df.drop('sales_growth_last_6m', axis=1) df = df.drop('formalization_level', axis=1) # formalization_map = { # 'Informal': 0, # 'Formal': 1 # } # df['formalization_level'] = df['formalization_level'].map(formalization_map) credit_availabe = { 'No': 0, 'Si': 1, } df['access_to_credit'] = df['access_to_credit'].map(credit_availabe) # If you want to consider 3 possible inputs for 'formalization_level', ensure only these values exist # df = df[df['formalization_level'].isin([0, 1, 2])] # Sample data X = df.drop('growth_potential', axis=1) # Features y = df['growth_potential'] # Target variable # Split data into training and testing sets X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42) # Initialize and train the RandomForestRegressor regressor = RandomForestRegressor(n_estimators=100, random_state=42) regressor.fit(X_train, y_train) # Save the trained model to a file joblib.dump(regressor, 'business_growth_predictor.pkl') # Make predictions y_pred = regressor.predict(X_test) # Evaluate the model mse = mean_squared_error(y_test, y_pred) print(f"Mean Squared Error: {mse}")