Random XGBoost Model Generation with Parameter Search

This function uses the XGBoost library to generate a random XGBoost model. It performs a random search over the parameters to find the best model.

Technology Stack : XGBoost, NumPy

Code Type : The type of code

Code Difficulty : Intermediate

                
                    
import xgboost as xgb
import numpy as np

def random_xgb_model(X, y, num_round=100):
    """
    Generates a random XGBoost model on the given dataset.
    """
    # Define the parameters for the random search
    param_dist = {
        'max_depth': [3, 5, 7, 9],
        'min_child_weight': [1, 3, 5],
        'gamma': [0.1, 0.3, 0.5],
        'subsample': [0.5, 0.7, 1.0],
        'colsample_bytree': [0.5, 0.7, 1.0]
    }
    
    # Initialize the XGBoost classifier
    xgb_clf = xgb.XGBClassifier(use_label_encoder=False, eval_metric='logloss')
    
    # Perform random search
    best_score = 0
    best_params = {}
    for i in range(10):
        # Sample random parameters
        sample_params = {key: np.random.choice(value) for key, value in param_dist.items()}
        xgb_clf.set_params(**sample_params)
        
        # Train the model
        xgb_clf.fit(X, y)
        
        # Evaluate the model
        score = xgb_clf.score(X, y)
        if score > best_score:
            best_score = score
            best_params = sample_params
    
    # Return the best parameters
    return best_params

# JSON representation of the code