import pandas as pd
import matplotlib.pyplot as plt
from time import time
pt_ = time()
print("Starting Grid Search Algo for catboost...")
param_grid = {'learning_rate' : [0.001, 0.01, 0.1],
'l2_leaf_reg' : [0.1, 1, 3, 10],
'depth' : [3, 6, 9, 12],
'iterations' : [100, 1000, 3000, 7000, 11000]}
model = CatBoost()
X = pd.read_csv(config_dict['PROCESSED_DATA'])
y = X[TARGET_COLUMN]
X = Pool(X.drop(TARGET_COLUMN, 1), y)
grid_search_result = model.grid_search(param_grid,
X,
cv=5,
partition_random_seed=0,
calc_cv_statistics=True,
stratified=True,
train_size=0.8,
verbose=True,
plot=True)
pd.DataFrame.from_dict(grid_search_result).to_csv('grid_params.csv',index=None)
print(f"Parameter Search Completed! Time Taken : {time()-pt_}s")
print("Best params found:")
