對於乙個簡單的文字情感分類來說,其實就是乙個二分類,這篇部落格主要講述的是使用scikit-learn
來做文字情感分類。分類主要分為兩步:1)訓練,主要根據訓練集來學習分類模型的規則。2)分類,先用已知的測試集評估分類的準確率等,如果效果還可以,那麼該模型對無標註的待測樣本進行**。
下面實現了svm,nb,邏輯回歸,決策樹,邏輯森林,knn 等幾種分類方法,主要**如下:
#coding:utf-8
from matplotlib import pyplot
import scipy as sp
import numpy as np
from sklearn.cross_validation import train_test_split
from sklearn.feature_extraction.text import countvectorizer
from sklearn.feature_extraction.text import tfidfvectorizer
from sklearn.metrics import precision_recall_curve
from sklearn.metrics import classification_report
from numpy import *
#*****===svm*****===#
def svmclass(x_train, y_train):
from sklearn.svm import svc
#調分類器
clf = svc(kernel = 'linear',probability=true)#default with 'rbf'
clf.fit(x_train, y_train)#訓練,對於監督模型來說是 fit(x, y),對於非監督模型是 fit(x)
return clf
#*****nb*****====#
def nbclass(x_train, y_train):
from sklearn.*****_bayes import multinomialnb
clf=multinomialnb(alpha=0.01).fit(x_train, y_train)
return clf
#*****===logistic regression*****===#
def logisticclass(x_train, y_train):
from sklearn.linear_model import logisticregression
clf = logisticregression(penalty='l2')
clf.fit(x_train, y_train)
return clf
#*****===knn*****===#
def knnclass(x_train,y_train):
from sklearn.neighbors import kneighborsclassifier
clf=kneighborsclassifier()
clf.fit(x_train,y_train)
return clf
#*****===decision tree *****===#
def dccisionclass(x_train,y_train):
from sklearn import tree
clf=tree.decisiontreeclassifier()
clf.fit(x_train,y_train)
return clf
#*****===random forest classifier *****===#
def random_forest_class(x_train,y_train):
from sklearn.ensemble import randomforestclassifier
clf= randomforestclassifier(n_estimators=8)#引數n_estimators設定弱分類器的數量
clf.fit(x_train,y_train)
return clf
#*****===準確率召回率 *****===#
def precision(clf):
doc_class_predicted = clf.predict(x_test)
print(np.mean(doc_class_predicted == y_test))#**結果和真實標籤
#準確率與召回率
precision, recall, thresholds = precision_recall_curve(y_test, clf.predict(x_test))
answer = clf.predict_proba(x_test)[:,1]
report = answer > 0.5
print(classification_report(y_test, report, target_names = ['neg', 'pos']))
print("--------------------")
from sklearn.metrics import accuracy_score
print('準確率: %.2f' % accuracy_score(y_test, doc_class_predicted))
if __name__ == '__main__':
data=
labels=
with open ("train2.txt","r")as file:
for line in file:
line=line[0:1]
with open("train2.txt","r")as file:
for line in file:
line=line[1:]
x=np.array(data)
labels=np.array(labels)
labels=[int (i)for i in labels]
movie_target=labels
#轉換成空間向量
count_vec = tfidfvectorizer(binary = false)
#載入資料集,切分資料集80%訓練,20%測試
x_train, x_test, y_train, y_test= train_test_split(x, movie_target, test_size = 0.2)
x_train = count_vec.fit_transform(x_train)
x_test = count_vec.transform(x_test)
print('**************支援向量機************ ')
precision(svmclass(x_train, y_train))
print('**************樸素貝葉斯************ ')
precision(nbclass(x_train, y_train))
print('**************最近鄰knn************ ')
precision(knnclass(x_train,y_train))
print('**************邏輯回歸************ ')
precision(logisticclass(x_train, y_train))
print('**************決策樹************ ')
precision(dccisionclass(x_train,y_train))
print('**************邏輯森林************ ')
precision(random_forest_class(x_train,y_train))
結果如下:
aspect level 的文字情感分類試驗結果1
前段時間準備了資料,試了一下 基於attention model的aspect level文字情感分類 用python keras實現 這篇文章裡面的模型。結果和文章裡差不多,驗證集準確率在75 80 左右。但仔細去看模型 的結果,這個資料其實並不好。剔除掉單個aspect的句子,多aspect句子...
基於LSTM分類文字情感分析
文字情感分析作為nlp的常見任務,具有很高的實際應用價值。本文將採用lstm模型,訓練乙個能夠識別文字postive,neutral,negative三種情感的分類器。本文的目的是快速熟悉lstm做情感分析任務,所以本文提到的只是乙個baseline,並在最後分析了其優劣。對於真正的文字情感分析,在...
文字情感分類(二) 深度學習模型
原文寫的不錯 源 有改動。coding utf 8 created on sep 6,2016 author zhangdapeng from future import absolute import 匯入3.x的特徵函式 from future import print function imp...