from os import chdir
chdir("C:/Users/pnlawlor/Google Drive/Research/Projects/Mad_Kegel")
import numpy as np
from fit_GLM import fit_logistic_GLM, plot_logistic_fit, fit_RF, fit_linear_model
X_data = np.random.rand(200,200)
y_data2 = np.round(X_data[:,0])
y_data = y_data2.ravel()
#models, scores, C, kf = fit_logistic_GLM(X_data, y_data,num_cv=10,verbose=False,plot_results=True)
#
#models2, scores2, num_estimators2, kf2 = fit_RF(X_data,y_data,num_estimators=1000,num_cv=10)
models3 = fit_linear_model(X_data,y_data)
#season_labels = season_labels[data_points]
results = (.5*(np.sign(y_data)+1)).astype(int)
X_data = pp.scale(X_data)
#models3, R2_3, loss_scores3, coef3, prob3, kf3, group_keys3 = fit_linear_model(
# X_data,y_data,
# results,keys,
# labels=season_labels)
#models3, R2_3, loss_scores3, coef3, prob3, kf3, group_keys3 = fit_linear_model(
# X_data,y_data,
# results,keys,num_cv=10)
models3, R2_3, loss_scores3, coef3, prob3, group_keys3 = fit_linear_model(
X_data,y_data,
results,keys,num_cv=1)
# first 1/3 win pct,
# mid 1/3 win pct,
# last 1/3 win pct,
# last 1/6 win pct,
# away win pct,
# ppg,
# ppg against,
# avg pt differential,
# stdv points for*losing team,
# stdv points against*losing team
save_model = False
