# -*- coding: utf-8 -*- """ Created on Thu Jul 20 09:59:50 2017@author: Administrator """import pandas as pd inputfile = 'input.xlsx' #excel輸入 outputfile = 'output.xls' #excel輸出 modelfile = 'modelweight.model' #神經(jīng)網(wǎng)絡權重保存 data = pd.read_excel(inputfile,index='Date',sheetname=0) #pandas以DataFrame的格式讀入excel表 feature = ['F1','F2','F3','F4'] #影響因素四個 label = ['L1'] #標簽一個，即需要進行預測的值 data_train = data.loc[range(0,20)].copy() #標明excel表從第0行到520行是訓練集#2 數(shù)據(jù)預處理和標注 data_mean = data_train.mean() data_std = data_train.std() data_train = (data_train - data_mean)/data_std #數(shù)據(jù)標準化 x_train = data_train[feature].as_matrix() #特征數(shù)據(jù) y_train = data_train[label].as_matrix() #標簽數(shù)據(jù)#3 建立一個簡單BP神經(jīng)網(wǎng)絡模型 from keras.optimizers import SGD,adamfrom keras.models import Sequential from keras.layers.core import Dense, Activation model = Sequential() #層次模型 model.add(Dense(12,input_dim=4,init='uniform')) #輸入層，Dense表示BP層 model.add(Activation('relu')) #添加激活函數(shù) model.add(Dense(1,input_dim=12)) #輸出層 model.compile(loss='mean_squared_error', optimizer='adam') #編譯模型 model.fit(x_train, y_train, nb_epoch = 1000, batch_size = 6) #訓練模型1000次 model.save_weights(modelfile) #保存模型權重#4 預測，并還原結果。 x = ((data[feature] - data_mean[feature])/data_std[feature]).as_matrix() data[u'L1_pred'] = model.predict(x) * data_std['L1'] + data_mean['L1']#5 導出結果 data.to_excel(outputfile) #6 畫出預測結果圖 import matplotlib.pyplot as plt p = data[['L1','L1_pred']].plot(subplots = True, style=['b-o','r-*']) plt.show()

總結

以上是生活随笔為你收集整理的keras 实现BP神经网络的全部內(nèi)容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網(wǎng)站內(nèi)容還不錯，歡迎將生活随笔推薦給好友。