手写数字识别¶
基于内置数据¶
In [11]:
#导入一些相关的第三方库
import matplotlib.pyplot as plt
from sklearn import datasets, svm, metrics
In [12]:
# 导入手写数据集
digits = datasets. load_digits()
images_and_labels = list(zip(digits. images, digits.target))
for index, (image, label) in enumerate(images_and_labels[:4]):
plt.subplot (2, 4, index + 1)
plt.axis('off')
plt. imshow(image, cmap=plt.cm.gray_r, interpolation='nearest')
plt.title('Training: %i' % label)
n_samples = len(digits. images)
data = digits. images.reshape((n_samples, -1))
In [13]:
# 创建一个分类器：支持向量机分类器
classifier = svm.SVC(gamma=0.001)
In [14]:
# 采用数据的前半段数据用来训练分类器模型
classifier.fit(data[:n_samples // 2], digits.target[:n_samples // 2])
Out[14]:
In [15]:
# 预测另外一半数据
expected = digits.target[n_samples // 2:]
predicted= classifier.predict(data[n_samples // 2:])
print("Classification report for classifier %s:\n%s\n"
% (classifier, metrics.classification_report(expected, predicted)))
print("Confusion matrix:\n%s"% metrics.confusion_matrix(expected, predicted))
images_and_predictions = list(zip(digits.images[n_samples // 2:],predicted))
for index,(image,prediction) in enumerate(images_and_predictions[:4]):
plt. subplot (2, 4, index + 5)
plt.axis('off')
plt. imshow(image, cmap=plt.cm.gray_r, interpolation='nearest')
plt.title('Prediction: %i' % prediction)
plt. show()
基于kaggle数据¶
In [16]:
#读取数字图片数据，分别截取图片数据和数据标签
%matplotlib inline
import pandas as pd
def opencsv () : # open with pandas
data = pd.read_csv('data/digit_train.csv')
datal = pd.read_csv('data/digit_test.csv')
train_data = data.values[0:,1:]
train_label = data.values[0:,0]
test_data = datal.values[0:,0:]
print('Data Load Done!')
return train_data, train_label, test_data
train_data, train_label, test_data = opencsv()
# Train_data 中存储了训练集的784个特征，Test_data存储了测试集的784个特征，train_lable则存储了训练集的标签
# 可以看出这道题是典型的监督学习问题
In [17]:
#查看前70幅图
import matplotlib.pyplot as plt
from numpy import*
print(shape(train_data),shape(test_data)) #-j///fj;J/l1f420001'-a l/fiifJ/l1.f28000-f-
def showPic(data):
plt.figure(figsize=(l2, 12))
for digit_num in range(O, 70):
plt. subplot(7, 10,digit_num+l)
grid_data = data[digit_num].reshape(28,28) # reshape from id to 2d pixel array
plt. imshow(grid_data, interpolation = "none", cmap = "afmhot")
plt.xticks([])
plt.yticks([])
plt.tight_layout()
showPic(train_data)
In [23]:
#看看第一幅图像的数据
print(train_data[0,].reshape(28,28))
数据清洗¶
从图和数值上看出，值在0-255范围变化，即每个特征都是连续的值，想想这样连续的值对我们后期的特征选择是重要还是不重要的？
我们观察发现，在0与>0的边界，其值都不很高，（有种写字笔水在纸张上晕开了的感觉？）
所以这里可以有三种处理：
（1）不对图像进行任何处理
（2）对图像进行二值化，0即0，>0即1
（3）对图像进行二值化，设置一个阈值，大于这个阈值，才为1，否则为0
很显然，（2）（3）种方法会造成原始信息的丢失，但是这两种方法对我们后面的工作是起到正面作用还是负面作用呢？稍安勿躁，后面再来探讨。
In [ ]:
def DataClean(data,epsilon): # normalize data
m, n = shape(data)
ret = zeros((m,n))
for i in range(m):
for j in range(n):
if data[i, j] > epsilon:
ret [i, j] = 1
else:
ret[i, j] = 0
return ret
In [ ]:
#不降维基本跑不出来
from sklearn import svm
from datetime import datetime
from sklearn.model_selection import cross_val_score
start= datetime.now()
model = svm.SVC(kernel='rbf', C=10)
metric = cross_val_score(model, train_data,train_label,cv=5,scoring='accuracy').mean()
end= datetime.now()
print('CV use: %f' %((end-start). seconds))
print('Offline Accuracy is' % metric)
下来就要考虑一个问题：如何从这784维中找出我们需要的维度呢？或者，如何对这784维进行投影，得到一个维度比较低的空间呢？
引入三种方法：
Principal Component Analysis ( PCA ) - Unsupervised, linear method
Linear Discriminant Analysis (LDA) - Supervised, linear method
t-distributed Stochastic Neighbour Embedding (t-SNE) - Nonlinear, probabilistic method
特征降维：主成分分析（PCA）¶
In [ ]:
from sklearn.decomposition import PCA
def getncomponent(inputdata):
pca = PCA()
pca.fit(inputdata)
EV_List = pca.explained_variance_
EVR_List = []
for j in range(len(EV_List)):
EVR_List.append(EV_List[j]/EV_List[0])
for j in range(len(EVR_List)):
if(EVR_List[j]< 0.10):
print('Recommend %d:' %j)
return j
getncomponent(train_data)
模型选择¶
分析题目可以知道，这是一个典型的多分类问题，所以可以用很多方法
比如knn 这个太慢了我就不尝试了
比如svm
比如LR
比如Random Forest
比如Decision Tree
比如GBDT
In [19]:
#降维
pca = PCA(n_components=22,whiten=True)
train_x = pca.fit_transform(train_data)
In [20]:
#模型测试函数
from datetime import datetime
def modeltest(train_x,train_label,model):
start = datetime.now()
metric = cross_val_score(model,train_x,train_label,cv=5,scoring='accuracy').mean()
end = datetime.now()
print('CV use: %f' %((end-start).seconds))
print('Offline Accuracy is %f' % (metric))
In [22]:
from sklearn import svm
from sklearn.model_selection import cross_val_score
SVM_model = svm.SVC(kernel='rbf', C=10)
print('PCA+SVM')
modeltest(train_x,train_label,SVM_model)
from sklearn.linear_model import LogisticRegression
LR_model = LogisticRegression()
print('PCA+LR')
modeltest(train_x,train_label,LR_model)
from sklearn.ensemble import RandomForestClassifier
RF_model = RandomForestClassifier(n_estimators=100)
print('PCA+RF')
modeltest(train_x,train_label,RF_model)
最后选择22维的PCA+SVM¶
In [ ]:
pea = PCA(n_components=22,whiten=True)
train_x =pca.fit_transform(train_data)
test_x = pca.transform(test_data)
modeltest(train_x,train_label,SVM_model)
resultname = 'PCA_SVM'
start = datetime.now()
SVM_model.fit(train_x,train_label)
end = datetime.now()
print('train time used:%f' % (end-start).seconds)
test_y = SVM_model.predict(test_x)
end = datetime.now()
print('predict time used:%f' % (end-start). seconds)
pred = [[index+ 1, x] for index, x in enumerate(test_y)]
savetxt(resultname+' .csv', pred, delimiter=',', fmt='%d,%d', header='Imageld,Label' ,comments='')
In [ ]: