def evaluate_list(model): evaluate_list = [] #4. 평가, 예측 y_predict = model.predict(x_pred) print(y_predict.shape) print(y_pred.shape) # r2_list r2 = r2_score(y_pred, y_predict) evaluate_list.append(r2) print('r2 : ', r2) # rmse_list rmse = mse_(y_pred, y_predict, squared=False) evaluate_list.append(rmse) print('rmse : ', rmse) # mae_list mae = mae_(y_pred, y_predict) evaluate_list.append(mae) print('mae : ', mae) # mse_list mse = mse_(y_pred, y_predict, squared=True) evaluate_list.append(mse) print('mse : ', mse) return evaluate_list
callbacks=[reduce_lr, cp]) # from tensorflow.keras.models import load_model # model = load_model('./data/hdf5/1_0del.hdf5', custom_objects={'leaky_relu': tf.nn.leaky_relu}) # 4. 평가, 예측 loss, mae = model.evaluate(x_pred, y_pred, batch_size=1024) y_predict = model.predict(x_pred) from sklearn.metrics import mean_squared_error as mse_ from sklearn.metrics import mean_absolute_error as mae_ # r2_list r2 = r2_score(y_pred, y_predict) print('r2 : ', r2) # rmse_list rmse = mse_(y_pred, y_predict, squared=False) print('rmse : ', rmse) # mae_list mae = mae_(y_pred, y_predict) print('mae : ', mae) # mse_list mse = mse_(y_pred, y_predict, squared=True) print('mse : ', mse) import matplotlib.pyplot as plt fig = plt.figure(figsize=(12, 4)) chart = fig.add_subplot(1, 1, 1) chart.plot(y_pred, marker='o', color='blue', label='실제값') chart.plot(y_predict, marker='^', color='red', label='예측값') plt.legend(loc='best')
#coding:utf-8 import numpy as np import os import pandas as pd from sklearn.metrics import mean_squared_error as mse_ #path source file1 = "E:\\houseprice\\33803.csv" file2 = "E:\\houseprice\\19502.csv" save = "E:\\houseprice\\result.csv" #read data data1 = pd.read_csv(file1, header=None) data2 = pd.read_csv(file2, header=None) #distance dist1 = 33803 dist2 = 19501 dist = np.sqrt(mse_(data1, data2)) #余弦定理 cos1 = (dist1 * dist1 + dist * dist - dist2 * dist2) / (2 * dist1 * dist) cos2 = (dist2 * dist2 + dist * dist - dist1 * dist1) / (2 * dist2 * dist) w1 = dist1 * cos1 / dist w2 = dist2 * cos2 / dist result = np.array(data1) * w2 + np.array(data2) * w1 result = pd.DataFrame(result) result.to_csv(save, index=None, header=None)
def RMSE(y_test, y_predict): return np.sqrt(mse_(y_test, y_predict))
#coding:utf-8 import numpy as np import pandas as pd from scipy import interpolate as inp from sklearn.metrics import mean_absolute_error as mae_ from sklearn.metrics import mean_squared_error as mse_ #read data file = "rst\\rst_cnn+mlp_300.csv" data = pd.read_csv(file, header=None) data = np.array(data) data1 = data for i in range(6): d = data[i * 96 * 21:(i + 1) * 96 * 21, :] label = d[:, 0] pred = d[:, 1] x = range(96 * 21) m = inp.UnivariateSpline(x, pred, s=len(pred) / 4) y_spline = m(x) data1[i * 96 * 21:(i + 1) * 96 * 21, 1] = y_spline print(i + 1) print(mse_(label, pred)) print(mse_(label, y_spline)) print(mae_(data1[:, 0], data1[:, 1])) data1 = pd.DataFrame(data1) data1.to_csv("rst\\rst_cnn+mlp_adjustedhist_spline.csv", header=None, index=None)