Keras implementation of an LSTM neural network to classify and predict the MINST dataset

這篇將會針對經常用於時間序列數據進行分類的神經網路LSTM對2D數據進行分類，例如手寫字母的圖像來進行討論。

首先將MINST數據集60,000張手寫數字(0至9)圖像用於28 x 28像素網格上做訓練，以及10,000張用於測試的圖像。因此，每個數字由784條信息表示。在這個測試中，將每個圖像分解為28個steps，並為LSTM網絡提供28個features，看它是否仍然可以學習。有效地將784像素的圖像分解為28個time-step的“movie”（每個frame包含28條訊息）。

在下面的第一次嘗試中，將完全沒有優化的網絡做學習能夠達到98.89％的準確度。

# James implementation of minst database

# import required to process csv files with pandas

import pandas as pd

# for array manipulation

import numpy as np

# for normalizing the data

from sklearn.preprocessing import MinMaxScaler

# allows onehotencoding

from sklearn.preprocessing import OneHotEncoder

 

# read the training csv file # data from 60000 images originally represented as a 28 x 28 pixel grid

# originally obtained from http://yann.lecun.com/exdb/mnist/

mnist_training_data = pd.read_csv(r'C:\Users\james\Anaconda3JamesData\mnist_train.csv', header=None)

 

# each row of data consists of

# the first element [0] is the label of the actual number (0 through 9)

# the following 784 element [1:785] is the actual number originally respresented as a 28 x 28 pixel grid

 

# we want to create training data, and label data for our LSTM to train from

mnist_training_data_values = mnist_training_data.iloc[:, 1:785].values

mnist_training_data_labels = mnist_training_data.iloc[:, 0].values

 

# we will normalize the training data (from 0 to 1 using MinMaxScaler)

# normalise the training data

scaler = MinMaxScaler(feature_range = (0.0, 1.0))

mnist_training_data_values = scaler.fit_transform(mnist_training_data_values.astype('float64'))

 

# the output of the network values (the labels) are 0, 1, 2, 3, 4, 5, 6, 7, 8, 9

 

# Onehotencoding expects a 2D array

mnist_training_data_labels = mnist_training_data_labels.reshape (60000,1)

 

# create the encoder

encoder = OneHotEncoder(sparse=False, categories='auto')

 

# encode the training labels

mnist_training_data_labels = encoder.fit_transform(mnist_training_data_labels)

 

# reshape mnist_training_data_values into 3d array

mnist_training_data_values = mnist_training_data_values.reshape(60000 , 28, 28)

 

# effectively I will feed each image into the LSTM as 28 rows of data

# with 28 steps - so effectively preceptually changing the data into a

# 28 step time sequence (with 28 features per step)

 

#print("mnist_training_data_values shape:", mnist_training_data_values.shape)

#print ("mnist_training_data_labels shape: ",mnist_training_data_labels.shape)

#print (mnist_training_data_labels )

 

from keras.models import Sequential

from keras.layers import LSTM

from keras.layers import Dense

from keras.optimizers import Adam

#from keras.layers import Dropout

 

# each sequence (representing one number from the minst database) has 28 steps and 28 features

model = Sequential()

model.add(LSTM(50,return_sequences=True,input_shape=(28, 28)))

model.add(LSTM(50))

model.add(Dense(10, activation='softmax'))

model.compile(Adam(lr=0.001), loss='categorical_crossentropy', metrics=['acc'])

 

history = model.fit(mnist_training_data_values , mnist_training_data_labels , validation_split=0.1, shuffle=True, batch_size= 28,epochs=500, verbose=2)

 

# display diagnostics of the training

import matplotlib.pyplot as plt 

%matplotlib inline

 

plt.figure(figsize=(13,8)) 

plt.plot(history.history['loss'], color='blue')

plt.plot(history.history['val_loss'], color='orange')

plt.plot(history.history['acc'], color='red')

plt.plot(history.history['val_acc'], color='green')

plt.title('model loss during training')

plt.ylabel('loss')

plt.xlabel('epoch')

plt.legend(['loss', 'val_loss', 'acc',' val_acc'], loc='upper left')

plt.show()

 

 

# prediction model

mnist_testing_data = pd.read_csv(r'C:\Users\james\Anaconda3JamesData\mnist_test.csv', header=None)

 

# process the data as before

mnist_testing_data_values = mnist_testing_data.iloc[:, 1:785].values

mnist_testing_data_labels = mnist_testing_data.iloc[:, 0].values

 

# using the same normalisation as I used on the training data

# hence using 'transform' and not 'fit_transform'

mnist_testing_data_values = scaler.transform(mnist_testing_data_values.astype('float64'))

 

length_of_testing_data = len(mnist_testing_data_values)

 

# reshape the testing data values to the same shape as the training values

mnist_testing_data_values = mnist_testing_data_values.reshape(length_of_testing_data, 28, 28)

 

predict = model.predict(mnist_testing_data_values)

 

# decode the onehotencoded prediction results

one_hot_decoded_data = encoder.inverse_transform(predict)

 

#length = len(one_hot_decoded_data) #the decoded predict model

one_hot_decoded_data = np.array(one_hot_decoded_data)

mnist_testing_data_labels = np.array(mnist_testing_data_labels)

 

#print (one_hot_decoded_data.shape)

#print (mnist_testing_data_labels.shape)

 

# reshape the one_hot_decoded_data

one_hot_decoded_data = one_hot_decoded_data.reshape(1,-1)

mnist_testing_data_labels = mnist_testing_data_labels.reshape(1,10000)

 

#print (one_hot_decoded_data.shape)

#print (mnist_testing_data_labels.shape)

 

# accuracy analytics

correct= 0

for i in range(0,10000):

    if (one_hot_decoded_data[0,i] == mnist_testing_data_labels[0,i]):

        print ("sample number: ",i," test: ",mnist_testing_data_labels[0,i], " predicted: ",one_hot_decoded_data[0,i], " CORRECT")

        correct+=1

    else:

        print ("sample number: ",i," test: ",mnist_testing_data_labels[0,i], " predicted: ",one_hot_decoded_data[0,i], " FALSE")

 

print ('Accuracy: %f' % ((correct/10000)*100))

參考
https://www.drjamesfroggatt.com/python-and-neural-networks/keras-implementation-of-an-lstm-neural-network-to-classify-and-predict-the-minst-dataset/