#rf = RandomForestRegressor(max_depth=10,n_estimators=10) #rf.fit(inputs, outputs) dt = DecisionTreeRegressor(max_depth=20) dt.fit(inputs, outputs) ml = dt print(ml.score(inputs[:1000,:],outputs[:1000,:])) print(ml.predict(inputs[:10,:])) print(outputs[:10,:]) #rf.output_coef_path = '' #rf_sk2f(rf) dt.output_coef_path = '' dt_sk2f(dt) data_dir = 'one_million_data' i_exact_f = np.load(data_dir+'/exact_f.npy') i_exact_angle = np.load(data_dir+'/exact_angle.npy') i_exact_centroid = np.load(data_dir+'/exact_centroid.npy') i_delta_angle = np.load(data_dir+'/delta_angle.npy') i_initial_angle = np.load(data_dir+'/initial_angle.npy') i_exact_f = i_exact_f.reshape([len(i_exact_f),1]) i_inputs = np.hstack((i_initial_angle,i_exact_f)) i_outputs = i_delta_angle.copy() print(ml.score(i_inputs,i_outputs))
rf.fit(inputs, outputs)
sco = open('rf_score.dat', 'w')
sc1 = rf.score(inputs, outputs)
sc2 = rf.score(i_inputs, i_outputs)
sco.write(str(sc1))
sco.write(str(sc2))
sco.close()
rf.output_coef_path = 'batch/'
rf_sk2f(rf)
sco = open('dt_score.dat', 'w')
sc1 = dt.score(inputs, outputs)
sc2 = dt.score(i_inputs, i_outputs)
sco.write(str(sc1))
sco.write(str(sc2))
sco.close()
dt.output_coef_path = 'batch/'
dt_sk2f(dt)
os.system('mv ' + rf.output_coef_path + 'dt_coef.dat ' +
rf.output_coef_path + 'dt_' + str(de) + '.dat')
os.system('mv ' + rf.output_coef_path + 'rf_coef.dat ' +
rf.output_coef_path + 'rf_' + str(de) + '.dat')
os.system('mv ' + 'dt_score.dat ' + rf.output_coef_path + 'dt_' + str(de) +
'score' + '.dat')
os.system('mv ' + 'rf_score.dat ' + rf.output_coef_path + 'rf_' + str(de) +
'score' + '.dat')
print('Depth ' + str(de) + 'success. \n')
