Python代码 PX化工数据时LSTM GRU 预测5天后的数值 回归随机森林.zip

import joblib import numpy as np import pandas as pd from tensorflow.python.keras.callbacks import EarlyStopping from tensorflow.python.keras.models import Model from tensorflow.python.keras.layers import LSTM, GRU from tensorflow.python.keras import Input from sklearn.svm import LinearSVR from sklearn.model_selection import train_test_split from sklearn.preprocessing import StandardScaler from sklearn.neighbors import KNeighborsRegressor from sklearn.metrics import mean_squared_error from sklearn.linear_model import LinearRegression, Ridge from sklearn.ensemble import RandomForestRegressor from tensorflow.keras.layers import Dense, Activation from tensorflow.keras.models import Sequential data = pd.read_excel("PX数据.xlsx") dfs_dict = {} for i in range(1, 6): for j in range(1, 6): for k in range(1, 6): x = np.zeros((len(data.index) - 5, 3)) y = np.zeros((len(data.index) - 5, 1)) for l in range(len(data.index) - 5): x[l, 0] = data.iloc[l + 5 - i, 1] x[l, 1] = data.iloc[l + 5 - j, 2] x[l, 2] = data.iloc[l + 5 - k, 3] y[l, 0] = data.iloc[l + 5, 3] x = pd.DataFrame(x) y = pd.DataFrame(y) dfs_dict[str(i) + str(j) + str(k)] = (x, y) pd.concat([x, y], axis=1).to_csv("all data/" + str(i) + str(j) + str(k) + ".csv") mse_array = np.zeros((len(dfs_dict.keys()), 8)) root_path = "" for i, (k, (x, y)) in enumerate(zip(dfs_dict.keys(), dfs_dict.values())): ss_x = StandardScaler() x = pd.DataFrame(ss_x.fit_transform(x)) ss_y = StandardScaler() y = pd.DataFrame(ss_y.fit_transform(y)) x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.1) knn = KNeighborsRegressor() knn.fit(x_train, y_train) y_pred = knn.predict(x_test) mse = mean_squared_error(y_test, y_pred) print(mse) mse_array[i, 0] = mse joblib.dump(knn, root_path + "knn" + k + ".model") lr = LinearRegression() lr.fit(x_train, y_train) y_pred = lr.predict(x_test) mse = mean_squared_error(y_test, y_pred) print(mse) mse_array[i, 1] = mse joblib.dump(lr, root_path + "lr" + k + ".model") ridge = Ridge() ridge.fit(x_train, y_train) y_pred = ridge.predict(x_test) mse = mean_squared_error(y_test, y_pred) print(mse) mse_array[i, 2] = mse joblib.dump(ridge, root_path + "ridge" + k + ".model") rf = RandomForestRegressor() rf.fit(x_train, y_train) y_pred = rf.predict(x_test) mse = mean_squared_error(y_test, y_pred) print(mse) mse_array[i, 3] = mse joblib.dump(rf, root_path + "rf" + k + ".model") svm = LinearSVR() svm.fit(x_train, y_train) y_pred = svm.predict(x_test) mse = mean_squared_error(y_test, y_pred) print(mse) mse_array[i, 4] = mse joblib.dump(svm, root_path + "svm" + k + ".model") es = EarlyStopping(monitor="val_loss", patience=5) model = Sequential() model.add(Dense(3, input_dim=3)) model.add(Activation('relu')) model.add(Dense(8)) model.add(Activation('relu')) model.add(Dense(8)) model.add(Activation('relu')) model.add(Dense(1)) model.compile(optimizer='rmsprop', loss='mae', metrics=['mse']) model.fit(x_train, y_train, epochs=200, batch_size=16, validation_split=0.2, callbacks=[es]) y_pred = model.predict(x_test) mse = mean_squared_error(y_test, y_pred) print(mse) mse_array[i, 5] = mse model.save(root_path + "bpnn" + k + ".h5") x_train = np.array(x_train).reshape((x_train.shape[0], 1, x_train.shape[1])) x_test = np.array(x_test).reshape((x_test.shape[0], 1, x_test.shape[1])) inp = Input(shape=(1, 3)) x = LSTM(1, input_shape=(1, 3), return_sequences=True)(inp) x = Dense(8, activation='relu')(x) x = Dense(8, activation='relu')(x) output = Dense(1)(x) model = Model(inp, output) model.compile(optimizer='rmsprop', loss='mae', metrics=['mse']) model.fit(x_train, y_train, epochs=200, batch_size=16, validation_split=0.2, callbacks=[es]) y_pred = model.predict(x_test) y_pred = y_pred.reshape((y_pred.shape[0], y_pred.shape[1])) mse = mean_squared_error(y_test, y_pred) print(mse) mse_array[i, 6] = mse model.save(root_path + "lstm" + k + ".h5") inp = Input(shape=(1, 3)) x = GRU(100, input_shape=(1, 3), return_sequences=True)(inp) x = Dense(8, activation='relu')(x) x = Dense(8, activation='relu')(x) output = Dense(1)(x) model = Model(inp, output) model.compile(optimizer='rmsprop', loss='mae', metrics=['mse']) model.fit(x_train, y_train, epochs=200, batch_size=16, validation_split=0.2, callbacks=[es]) y_pred = model.predict(x_test) y_pred = y_pred.reshape((y_pred.shape[0], y_pred.shape[1])) mse = mean_squared_error(y_test, y_pred) print(mse) mse_array[i, 7] = mse model.save(root_path + "gru" + k + ".h5") mse_array = pd.DataFrame(mse_array, index=dfs_dict.keys(), columns=["KNN", "LinearRegression", "RidgeRegression", "RandomForestRegression", "SVMRegression", "NormalNeuralNetwork", "LSTM", "GRU"]) mse_array.to_csv("result.csv")