主要是通过mnist了解kaggle的操作细节,最终这里的结果为:

kaggle竞赛_mnist_10%-LMLPHP

引入必须的库

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
import seaborn as sns #专门用于数据可视化的
%matplotlib inline np.random.seed(2) from sklearn.model_selection import train_test_split
from sklearn.metrics import confusion_matrix
import itertools from keras.utils.np_utils import to_categorical # convert to one-hot-encoding
from keras.models import Sequential
from keras.layers import Dense, Dropout, Flatten, Conv2D, MaxPool2D
from keras.optimizers import RMSprop
from keras.preprocessing.image import ImageDataGenerator
from keras.callbacks import ReduceLROnPlateau

sns设定格式

sns.set(style='white', context='notebook', palette='deep')
 

读取数据,两个kaggle上下载的csv已经提前放置在了固定的地方

 
train = pd.read_csv("./input/train.csv")
test = pd.read_csv("./input/test.csv")
 
 

数据的显示和预处理

 
Y_train = train["label"]  #获得label
X_train = train.drop(labels = ["label"],axis = 1) #获得label以外的东西,也即是数据
del train #没用了
g = sns.countplot(Y_train)
 
 
aaarticlea/png;base64,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" alt="" />
 

空数据检查

X_train.isnull().any().describe()   #isnull是所有空数据,any是进行与运算,describe其实是用来查看第一个数据是什么的
Out[42]:
count       784unique        1top       Falsefreq        784dtype: object
test.isnull().any().describe()    #通过这里的检查,可以发现所有数据的isnull都为false,也就是所有的地方都是有数据的
Out[43]:
count       784unique        1top       Falsefreq        784dtype: object
 

将mnist图片转换为浮点类型

X_train = X_train / 255.0
test = test / 255.0
 

将mnist图片转化为图片的大小

X_train = X_train.values.reshape(-1,28,28,1)   #这里就是按照28*28的图片大小进行压缩
test = test.values.reshape(-1,28,28,1) #这个地方的这种写法,能够成功地将图片转化成28*28 * N的格式
 

准备训练数据

Y_train = to_categorical(Y_train, num_classes = 10) #onhot
 

即使是有test了,也要进行train的数据分割

random_seed = 2
X_train, X_val, Y_train, Y_val = train_test_split(X_train, Y_train, test_size = 0.1, random_state=random_seed)
 

打印结果,查看是否正确

g = plt.imshow(X_train[1][:,:,0])
 
 
aaarticlea/png;base64,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" alt="" />
 

开始导入cnn

model = Sequential()                   #序贯,关于模型的选择我还没有什么想法,我不知道从模型方面考虑怎样才能从0.9961继续往上增长

model.add(Conv2D(filters = 32, kernel_size = (5,5),padding = 'Same', activation ='relu', input_shape = (28,28,1)))
model.add(Conv2D(filters = 32, kernel_size = (5,5),padding = 'Same', activation ='relu'))
model.add(MaxPool2D(pool_size=(2,2)))
model.add(Dropout(0.25)) model.add(Conv2D(filters = 64, kernel_size = (3,3),padding = 'Same', activation ='relu'))
model.add(Conv2D(filters = 64, kernel_size = (3,3),padding = 'Same', activation ='relu'))
model.add(MaxPool2D(pool_size=(2,2), strides=(2,2)))
model.add(Dropout(0.25)) model.add(Flatten())
model.add(Dense(256, activation = "relu"))
model.add(Dropout(0.5))
model.add(Dense(10, activation = "softmax"))

训练准备

optimizer = RMSprop(lr=0.001, rho=0.9, epsilon=1e-08, decay=0.0)
model.compile(optimizer = optimizer , loss = "categorical_crossentropy", metrics=["accuracy"])

学习率自动降低

learning_rate_reduction = ReduceLROnPlateau(monitor='val_acc', patience=3, verbose=1, factor=0.5, min_lr=0.00001)
In [52]:
epochs = 28 # 训练次数
batch_size = 256

数据增量方法

datagen = ImageDataGenerator(
featurewise_center=False, # set input mean to 0 over the dataset
samplewise_center=False, # set each sample mean to 0
featurewise_std_normalization=False, # divide inputs by std of the dataset
samplewise_std_normalization=False, # divide each input by its std
zca_whitening=False, # apply ZCA whitening
rotation_range=10, # randomly rotate images in the range (degrees, 0 to 180)
zoom_range = 0.1, # Randomly zoom image
width_shift_range=0.1, # randomly shift images horizontally (fraction of total width)
height_shift_range=0.1, # randomly shift images vertically (fraction of total height)
horizontal_flip=False, # randomly flip images
vertical_flip=False) # randomly flip images datagen.fit(X_train)
In [54]:
#开始训练
history = model.fit_generator(datagen.flow(X_train,Y_train, batch_size=batch_size),
epochs = epochs, validation_data = (X_val,Y_val),
verbose = 2, steps_per_epoch=X_train.shape[0] // batch_size
, callbacks=[learning_rate_reduction])

绘制曲线

fig, ax = plt.subplots(2,1)
ax[0].plot(history.history['loss'], color='b', label="Training loss")
ax[0].plot(history.history['val_loss'], color='r', label="validation loss",axes =ax[0])
legend = ax[0].legend(loc='best', shadow=True) ax[1].plot(history.history['acc'], color='b', label="Training accuracy")
ax[1].plot(history.history['val_acc'], color='r',label="Validation accuracy")
legend = ax[1].legend(loc='best', shadow=True)
aaarticlea/png;base64,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" alt="" />
 

得出结果

results = model.predict(test)
results = np.argmax(results,axis = 1)
results = pd.Series(results,name="Label")
submission = pd.concat([pd.Series(range(1,28001),name = "ImageId"),results],axis = 1)
submission.to_csv("mnist_kaggle_jsxyhelu.csv",index=False)

最后要把 mnist_kaggle_jsxyhelu.csv上传上去

 
kaggle竞赛_mnist_10%-LMLPHP

05-07 15:32