import xlrd import jieba import sys import importlib import os #python内置的包，用于进行文件目录操作，我们将会用到os.listdir函数 import pickle #导入cPickle包并且取一个别名pickle #持久化类 import random import numpy as np import matplotlib.pyplot as plt from mpl_toolkits.mplot3d import Axes3D from pylab import mpl from sklearn.naive_bayes import MultinomialNB # 导入多项式贝叶斯算法包 from sklearn import svm from sklearn import metrics from sklearn.datasets.base import Bunch from sklearn.feature_extraction.text import TfidfVectorizer importlib.reload(sys) #把内容和类别转化成一个向量的形式 trainContentdatasave=[] #存储所有训练和测试数据的分词 testContentdatasave=[] trainContentdata = [] testContentdata = [] trainlabeldata = [] testlabeldata = [] #导入文本描述的训练和测试数据 def importTrainContentdata(): file = '20180716_train.xls' wb = xlrd.open_workbook(file) ws = wb.sheet_by_name("Sheet1") for r in range(ws.nrows): col = [] for c in range(1): col.append(ws.cell(r, c).value) trainContentdata.append(col) def importTestContentdata(): file = '20180716_test.xls' wb = xlrd.open_workbook(file) ws = wb.sheet_by_name("Sheet1") for r in range(ws.nrows): col = [] for c in range(1): col.append(ws.cell(r, c).value) testContentdata.append(col) #导入类别的训练和测试数据 def importTrainlabeldata(): file = '20180716_train_label.xls' wb = xlrd.open_workbook(file) ws = wb.sheet_by_name("Sheet1") for r in range(ws.nrows): col = [] for c in range(1): col.append(ws.cell(r, c).value) trainlabeldata.append(col) def importTestlabeldata(): file = '20180716_test_label.xls' wb = xlrd.open_workbook(file) ws = wb.sheet_by_name("Sheet1") for r in range(ws.nrows): col = [] for c in range(1): col.append(ws.cell(r, c).value) testlabeldata.append(col) """ def importClassSet(): file = 'ClassSet.xls' wb = xlrd.open_workbook(file) ws = wb.sheet_by_name("Sheet1") for r in range(ws.nrows): col = [] for c in range(ws.ncols): col.append(ws.cell(r, c).value) ClassSet.append(col) """ def buildtrainbunch(bunch_path): bunch = Bunch(label=[],contents=[]) for item1 in trainlabeldata: bunch.label.append(item1) for item2 in trainContentdata: item2=str(item2) item2 = item2.replace(" ", "") item2 = item2.replace(" ", "") content_seg=jieba.cut(item2) save2='' for item3 in content_seg: if len(item3) > 1 and item3!=' ': trainContentdatasave.append(item3) save2=save2+","+item3 bunch.contents.append(save2) with open(bunch_path, "wb") as file_obj: pickle.dump(bunch, file_obj) print("构建训练数据文本对象结束！！！") def buildtestbunch(bunch_path): bunch = Bunch(label=[],contents=[]) for item1 in testlabeldata: bunch.label.append(item1) for item2 in testContentdata: item2=str(item2) item2 = item2.replace(" ", "") item2 = item2.replace(" ", "") content_seg=jieba.cut(item2) save2='' for item3 in content_seg: if len(item3) > 1 and item3!=' ': testContentdatasave.append(item3) save2=save2+","+item3 bunch.contents.append(save2) with open(bunch_path, "wb") as file_obj: pickle.dump(bunch, file_obj) print("构建测试数据文本对象结束！！！") #读取停用词 def _readfile(path): with open(path, "rb") as fp: content = fp.read() return content # 读取bunch对象 def _readbunchobj(path): with open(path, "rb") as file_obj: bunch = pickle.load(file_obj) return bunch # 写入bunch对象 def _writebunchobj(path, bunchobj): with open(path, "wb") as file_obj: pickle.dump(bunchobj, file_obj) def vector_space(stopword_path,bunch_path,space_path): stpwrdlst = _readfile(stopword_path).splitlines()#读取停用词 bunch = _readbunchobj(bunch_path)#导入分词后的词向量bunch对象 #构建tf-idf词向量空间对象 tfidfspace = Bunch(label=bunch.label,tdm=[], vocabulary={}) ''' 权重矩阵tdm，其中，权重矩阵是一个二维矩阵，tdm[i][j]表示，第j个词（即词典中的序号）在第i个类别中的IF-IDF值 ''' #使用TfidVectorizer初始化向量空间模型 vectorizer = TfidfVectorizer(stop_words=stpwrdlst, sublinear_tf=True, max_df=0.5, min_df=0.0001,use_idf=False,max_features=10000) #print(vectorizer) #文本转为词频矩阵，单独保存字典文件 tfidfspace.tdm = vectorizer.fit_transform(bunch.contents) tfidfspace.vocabulary = vectorizer.vocabulary_ #创建词袋的持久化 _writebunchobj(space_path, tfidfspace) print("if-idf词向量空间实例创建成功！！！") def testvector_space(stopword_path,bunch_path,space_path,train_tfidf_path): stpwrdlst = _readfile(stopword_path).splitlines()#把停用词变成列表 bunch = _readbunchobj(bunch_path) tfidfspace = Bunch(label=bunch.label,tdm=[], vocabulary={}) ''' tdm存放的是计算后得到的TF-IDF权重矩阵. vocabulary是词向量空间的索引，例如，如果我们定义的词向量空间是（我，喜欢，相国大人），那么vocabulary就是这样一个索引字典 vocabulary={"我":0,"喜欢":1,"相国大人":2}，你可以简单的理解为：vocabulary就是词向量空间的坐标轴，索引值相当于表明了第几个维度。 ''' #导入训练集的TF-IDF词向量空间 ★★ trainbunch = _readbunchobj(train_tfidf_path) tfidfspace.vocabulary = trainbunch.vocabulary ''' 关于参数，你只需要了解这么几个就可以了： stop_words: 传入停用词，以后我们获得vocabulary_的时候，就会根据文本信息去掉停用词得到 vocabulary: 之前说过，不再解释。 sublinear_tf: 计算tf值采用亚线性策略。比如，我们以前算tf是词频，现在用1+log(tf)来充当词频。 smooth_idf: 计算idf的时候log(分子/分母)分母有可能是0，smooth_idf会采用log(分子/(1+分母))的方式解决。默认已经开启，无需关心。 norm: 归一化，我们计算TF-IDF的时候，是用TF*IDF，TF可以是归一化的，也可以是没有归一化的，一般都是采用归一化的方法，默认开启. max_df: 有些词，他们的文档频率太高了（一个词如果每篇文档都出现，那还有必要用它来区分文本类别吗？当然不用了呀），所以，我们可以 设定一个阈值，比如float类型0.5（取值范围[0.0,1.0]）,表示这个词如果在整个数据集中超过50%的文本都出现了，那么我们也把它列 为临时停用词。当然你也可以设定为int型，例如max_df=10,表示这个词如果在整个数据集中超过10的文本都出现了，那么我们也把它列 为临时停用词。 min_df: 与max_df相反，虽然文档频率越低，似乎越能区分文本，可是如果太低，例如10000篇文本中只有1篇文本出现过这个词，仅仅因为这1篇 文本，就增加了词向量空间的维度，太不划算。 当然，max_df和min_df在给定vocabulary参数时，就失效了。 ''' vectorizer = TfidfVectorizer(stop_words=stpwrdlst, sublinear_tf=True, max_df=0.7, vocabulary=trainbunch.vocabulary, min_df=0.001) #print(vectorizer) tfidfspace.tdm = vectorizer.fit_transform(bunch.contents) _writebunchobj(space_path, tfidfspace) print("if-idf词向量空间实例创建成功！！！") def metrics_result(actual, predict): # metrics.f1_score(y_test, y_pred, average='weighted', labels=np.unique(y_pred)) print('精度:{0:.3f}'.format(metrics.precision_score(actual, predict,average='weighted', labels=np.unique(predict)))) print('召回:{0:0.3f}'.format(metrics.recall_score(actual, predict,average='weighted', labels=np.unique(predict)))) print('f1-score:{0:.3f}'.format(metrics.f1_score(actual, predict, average='weighted', labels=np.unique(predict)))) #准确率和召回率是相互影响的，理想情况下是二者都高，但是一般情况下准确率高，召回率就低；召回率高，准确率就低 if __name__=="__main__": importTrainContentdata() importTestContentdata() importTrainlabeldata() importTestlabeldata() #导入分词后的词向量bunch对象 train_bunch_path ="F:/goverment/ArticleMining/trainbunch.bat"#Bunch保存路径 test_bunch_path ="F:/goverment/ArticleMining/testbunch.bat" stopword_path ="F:/goverment/ArticleMining/hlt_stop_words.txt" train_space_path = "F:/goverment/ArticleMining/traintfdifspace.dat" test_space_path = "F:/goverment/ArticleMining/testtfdifspace.dat" #对训练和测试集进行bunch操作 buildtrainbunch(train_bunch_path) buildtestbunch(test_bunch_path) vector_space(stopword_path,train_bunch_path,train_space_path) testvector_space(stopword_path,test_bunch_path,test_space_path,train_space_path) #导入训练和测试数据集 train_set=_readbunchobj(train_space_path) test_set=_readbunchobj(test_space_path) print(train_set.tdm) ''' mm=0 ii=0 jj=0 for i in range(3142): for j in range(3142): if train_set.tdm[i][j] >mm: mm=train_set.tdm[i][j] ii=i jj=j print(ii) print(jj) ''' #test_set.tdm #train_set.label # 训练分类器：输入词袋向量和分类标签，alpha:0.001 alpha越小，迭代次数越多，精度越高 #低召回、F1： 0.75 rbf:0.59 0.8 rbf 0.578 #c0.75 poly 66.5 精度:0.665 gamma=10 召回:0.330 f1-score:0.416 #C=0.7, kernel='poly', gamma=10 召回:0.331 f1-score:0.417 # alpha:0.001 alpha 越小，迭代次数越多，精度越高 ''' clf = MultinomialNB(alpha=0.052).fit(train_set.tdm, train_set.label) #clf = svm.SVC(C=0.7, kernel='poly', gamma=10, decision_function_shape='ovr') clf.fit(train_set.tdm, train_set.label) predicted=clf.predict(test_set.tdm) tv = TfidfVectorizer() train_data = tv.fit_transform(X_train) test_data = tv.transform(X_test) lr = LogisticRegression(C=3) lr.fit(train_set.tdm, train_set.label) predicted=lr.predict(test_set.tdm) print(lr.score(test_set.tdm, test_set.label)) #print(test_set.tdm) ''' clf = SVC(C=1500) clf.fit(train_set.tdm, train_set.label) predicted=clf.predict(test_set.tdm) print(clf.score(test_set.tdm, test_set.label)) ''' from sklearn.neighbors import KNeighborsClassifier knnclf = KNeighborsClassifier(n_neighbors=9)#default with k=5 knnclf.fit(train_set.tdm,train_set.label) predicted = knnclf.predict(test_set.tdm) ''' a=[] b=[] for i in range(len(predicted)): b.append((int)(float(predicted[i]))) a.append(int(test_set.label[i][0])) f=open('F:/goverment/ArticleMining/predict.txt', 'w') for i in range(len(predicted)): f.write(str(b[i])) f.write(' ') f.write("写好了") f.close() #for i in range(len(predicted)): #print(b[i]) metrics_result(a, b)