NLTK07《Python自然语言处理》code06 学习分类文本

学习分类文本

# -*- coding: utf-8 -*-# win10 python3.5.3/python3.6.1 nltk3.2.4# 《Python自然语言处理》 06 学习分类文本# pnlp06.py 本部分代码有几个训练比较耗时，大约需要60+分钟# 6.1 监督式分类def gender_features(word):    return {'last_letter':word[-1]}res = gender_features('Shrek')print(res) # {'last_letter': 'k'}from nltk.corpus import namesimport randomimport nltknames = ([(name, 'male') for name in names.words('male.txt')] +         [(name, 'female') for name in names.words('female.txt')])random.shuffle(names)featuresets = [(gender_features(n), g) for (n, g) in names]train_set, test_set = featuresets[500:], featuresets[:500]classifier = nltk.NaiveBayesClassifier.train(train_set)res = classifier.classify(gender_features('Neo'))print(res) # maleres = classifier.classify(gender_features('Trinity'))print(res) # femaleres = nltk.classify.accuracy(classifier, test_set)print(res) # 0.73classifier.show_most_informative_features(5)# Most Informative Features#              last_letter = 'a'            female : male   =     38.1 : 1.0#              last_letter = 'k'              male : female =     30.9 : 1.0#              last_letter = 'f'              male : female =     17.4 : 1.0#              last_letter = 'p'              male : female =     11.9 : 1.0#              last_letter = 'v'              male : female =     10.6 : 1.0from nltk.classify import apply_featurestrain_set = apply_features(gender_features, names[500:])test_set = apply_features(gender_features, names[:500])# 选择正确的特征def gender_features2(name):    features = {}    features["firstletter"] = name[0].lower()    features["lastletter"] = name[-1].lower()    for letter in 'abcdefghijklmnopgrstuvwxyz':        features["count(%s)" % letter] = name.lower().count(letter)        features["has(%s)" % letter] = (letter in name.lower())    return featuresres = gender_features2('John')print(res) # {'firstletter': 'j', 'lastletter': 'n', 'count(a)': 0,...featuresets = [(gender_features2(n), g) for (n, g) in names]train_set, test_set = featuresets[500:], featuresets[:500]classifier = nltk.NaiveBayesClassifier.train(train_set)res = nltk.classify.accuracy(classifier, test_set)print(res) # 0.762# 训练集、开发测试集、测试集train_names = names[1500:]devtest_names = names[500:1500]test_names = names[:500]train_set = [(gender_features(n), g) for (n, g) in train_names]devtest_set = [(gender_features(n), g) for (n, g) in devtest_names]test_set = [(gender_features(n), g) for (n, g) in test_names]classifier = nltk.NaiveBayesClassifier.train(train_set)res = nltk.classify.accuracy(classifier, devtest_set)print(res) # 0.753errors = []for (name, tag) in devtest_names:    guess = classifier.classify(gender_features(name))    if guess != tag:        errors.append((tag, guess, name))for (tag, guess, name) in sorted(errors):    print('correct=%-8s guess=%-8s name=%-30s' % (tag, guess, name))# correct=female   guess=male     name=Abigael# correct=female   guess=male     name=Adriaens# ...def gender_featuress(word):    return {'suffix1':word[-1:],            'suffix2':word[-2:]}train_set = [(gender_features(n), g) for (n, g) in train_names]devtest_set = [(gender_features(n), g) for (n, g) in devtest_names]classifier = nltk.NaiveBayesClassifier.train(train_set)res = nltk.classify.accuracy(classifier, devtest_set)print(res) # 0.771# 文档分类import random, nltkfrom nltk.corpus import movie_reviewsdocuments = [(list(movie_reviews.words(fileid)), category)             for category in movie_reviews.categories()             for fileid in movie_reviews.fileids(category)]random.shuffle(documents)all_words = nltk.FreqDist(w.lower() for w in movie_reviews.words())word_features = list(all_words.keys())[:2000]def document_features(document):    document_words = set(document)    features = {}    for word in word_features:        features['contains(%s)' % word] = (word in document_words)    return featuresres = document_features(movie_reviews.words('pos/cv957_8737.txt'))print(res) # {'contains(plot)': True, 'contains(:)': True, ...featuresets = [(document_features(d), c) for (d, c) in documents]train_set, test_set = featuresets[100:], featuresets[:100]classifier = nltk.NaiveBayesClassifier.train(train_set)res = nltk.classify.accuracy(classifier, test_set)print(res) # 0.79classifier.show_most_informative_features(5)# Most Informative Features#        contains(martian) = True              neg : pos    =      7.7 : 1.0#      contains(atrocious) = True              neg : pos    =      7.1 : 1.0#  contains(unimaginative) = True              neg : pos    =      7.1 : 1.0#         contains(turkey) = True              neg : pos    =      6.8 : 1.0#     contains(schumacher) = True              neg : pos    =      6.7 : 1.0# 词性标注import nltkfrom nltk.corpus import brownsuffix_fdist = nltk.FreqDist()for word in brown.words():    word = word.lower()    suffix_fdist[word[-1:]] += 1    suffix_fdist[word[-2:]] += 1    suffix_fdist[word[-3:]] += 1common_suffixes = list(suffix_fdist.keys())[:100]print(common_suffixes) # ['e', 'he', 'the', 'n', 'on', 'ton', 'y', 'ty',...def pos_features(word):    features = {}    for suffix in common_suffixes:        features['endswith(%s)' % suffix] = word.lower().endswith(suffix)    return featurestagged_words = brown.tagged_words(categories='news')featuresets = [(pos_features(n), g) for (n, g) in tagged_words]size = int(len(featuresets) * 0.1)train_set, test_set = featuresets[size:], featuresets[:size]classifier = nltk.DecisionTreeClassifier.train(train_set)res = nltk.classify.accuracy(classifier, test_set)print(res) # 0.5689706613625062res = classifier.classify(pos_features('cats'))print(res) # NNSprint(classifier.pseudocode(depth=4))"""if endswith(the) == False:   if endswith(,) == False:     if endswith(s) == False:       if endswith(.) == False: return '.'      if endswith(.) == True: return '.'    if endswith(s) == True:       if endswith(was) == False: return 'PP$'      if endswith(was) == True: return 'BEDZ'  if endswith(,) == True: return ','if endswith(the) == True: return 'AT'"""# 探索上下文语境def pos_features(sentence, i):    features = {"suffix(1)":sentence[i][-1:],                "suffix(2)":sentence[i][-2:],                "suffix(3)":sentence[i][-3:]}    if i == 0:        features["prev-word"] = "<START>"    else:        features["prev-word"] = sentence[i-1]    return featuresres = pos_features(brown.sents()[0], 8)print(res) # {'suffix(1)': 'n', 'suffix(2)': 'on', 'suffix(3)': 'ion', 'prev-word': 'an'}tagged_sents = brown.tagged_sents(categories='news')featuresets = []for tagged_sent in tagged_sents:    untagged_sent = nltk.tag.untag(tagged_sent)    for i, (word, tag) in enumerate(tagged_sent):        featuresets.append((pos_features(untagged_sent, i), tag))size = int(len(featuresets) * 0.1)train_set, test_set = featuresets[size:], featuresets[:size]classifier = nltk.NaiveBayesClassifier.train(train_set)res = nltk.classify.accuracy(classifier, test_set)print(res) # 0.7891596220785678# 序列分类def pos_features(sentence, i, history):    features = {"suffix(1)": sentence[i][-1:],                "suffix(2)": sentence[i][-2:],                "suffix(3)": sentence[i][-3:]}    if i == 0:        features["prev-word"] = "<STRAT>"        features["prev-tag"] = "<START>"    else:        features["prev-word"] = sentence[i-1]        features["prev-tag"] = sentence[i-1]    return featuresimport nltkfrom nltk.corpus import brownclass ConsecutivePosTagger(nltk.TaggerI):    def __init__(self, train_sents):        train_set = []        for tagged_sent in train_sents:            untagged_sent = nltk.tag.untag(tagged_sent)            history = []            for i, (word, tag) in enumerate(tagged_sent):                featureset = pos_features(untagged_sent, i, history)                train_set.append((featureset, tag))                history.append(tag)            self.classifier = nltk.NaiveBayesClassifier.train(train_set)    def tag(self, sentence):        history = []        for i, word in enumerate(sentence):            featureset = pos_features(sentence, i, history)            tag = self.classifier.classify(featureset)            history.append(tag)        return zip(sentence, history)tagged_sents = brown.tagged_sents(categories='news')size = int(len(tagged_sents)*0.1)train_sents, test_sents = tagged_sents[size:], tagged_sents[:size]tagger = ConsecutivePosTagger(train_sents)res = tagger.evaluate(test_sents)print(res) # 0.7965693092257765# 其他序列分类方法# 6.2 监督式分类的举例# 句子分类import nltksents = nltk.corpus.treebank_raw.sents()tokens = []boundaries = set()offset = 0for sent in nltk.corpus.treebank_raw.sents():    tokens.extend(sent)    offset += len(sent)    boundaries.add(offset - 1)def punct_features(tokens, i):    return {'next-word-capitalized': tokens[i+1][0].isupper(),            'prevword': tokens[i-1].lower(),            'punct': tokens[i],            'prev-word-is-one-char': len(tokens[i-1])==1}featuresets = [(punct_features(tokens, i), (i in boundaries))               for i in range(1, len(tokens) - 1)               if tokens[i] in '.?!']size = int(len(featuresets) * 0.1)train_set, test_set = featuresets[size:], featuresets[:size]classifier = nltk.NaiveBayesClassifier.train(train_set)res = nltk.classify.accuracy(classifier, test_set)print(res) # 0.936026936026936def segment_sentences(words):    start = 0    sents = []    for i, word in words:        if word in '.?!' and classifier.classify(punct_features(words, i)) == True:            sents.append(words[start:i+1])            start = i + 1    if start < len(words):        sents.append(words[start:])        return sents# 识别对话行为类型posts = nltk.corpus.nps_chat.xml_posts()[:10000]def dialogure_act_features(post):    features = {}    for word in nltk.word_tokenize(post):        features['contains(%s)' % word.lower()] = True    return featuresfeaturesets = [(dialogure_act_features(post.text), post.get('class')) for post in posts]size = int(len(featuresets) * 0.1)train_set, test_set = featuresets[size:], featuresets[:size]classifier = nltk.NaiveBayesClassifier.train(train_set)res = nltk.classify.accuracy(classifier, test_set)print(res) # 0.668# 识别文字蕴涵(Recognizing textual entailment, RTE)def rte_features(rtepair):    extractor = nltk.RTEFeatureExtractor(rtepair)    features = {}    features['word_overlap'] = len(extractor.overlap('word'))    features['word_hyp_extra'] = len(extractor.hyp_extra('word'))    features['ne_overlap'] = len(extractor.overlap('ne'))    features['ne_hyp_extra'] = len(extractor.hyp_extra('ne'))    return featuresrtepair = nltk.corpus.rte.pairs(['rte3_dev.xml'])[33]extractor = nltk.RTEFeatureExtractor(rtepair)print(extractor.text_words) # {'four', 'republics', 'association', 'Iran', 'Co', ...print(extractor.hyp_words) # {'SCO.', 'China', 'member'}print(extractor.overlap('word')) # set()print(extractor.overlap('ne')) # {'China'}print(extractor.hyp_extra('word')) # {'member'}# 扩展到大型数据集# 6.3 评估# 测试集import random, nltkfrom nltk.corpus import browntagged_sents = list(brown.tagged_sents(categories='news'))random.shuffle(tagged_sents)size = int(len(tagged_sents)*0.1)train_set, test_set = tagged_sents[size:], tagged_sents[:size]file_ids = brown.fileids(categories='news')size = int(len(file_ids)*0.1)train_set = brown.tagged_sents(file_ids[size:])test_set = brown.tagged_sents(file_ids[:size])train_set = brown.tagged_sents(categories='news')test_set = brown.tagged_sents(categories='fiction')# 精确度# classifier = nltk.NaiveBayesClassifier.train(train_set)# print('Accuracy: %4.2f' % nltk.classify.accuracy(classifier, test_set))# 精确度和召回率# 精确度(Precision)，表示发现的项目中有多少是相关的# 召回率(Recall),表示相关的项目中发现了多少# F-度量值(F-Measure|F-Score)，(2*Precision*Recall)/(Precision+Recall)# 混淆矩阵def tag_list(tagged_sents):    return [tag for sent in tagged_sents for (word, tag) in sent]def apply_tagger(tagger, corpus):    return [tagger.tag(nltk.tag.untag(sent)) for sent in corpus]sents = brown.tagged_sents(categories='editorial')t0 = nltk.DefaultTagger('NN')t1 = nltk.UnigramTagger(sents, backoff=t0)t2 = nltk.BigramTagger(sents, backoff=t1)gold = tag_list(sents)test = tag_list(apply_tagger(t2, sents))cm = nltk.ConfusionMatrix(gold, test)print(cm)# 交叉验证# 6.4 决策树# 商和信息增益import mathdef entropy(labels):    freqdist = nltk.FreqDist(labels)    probs = [freqdist.freq(l) for l in nltk.FreqDist(labels)]    return -sum([p * math.log(p, 2) for p in probs])print(entropy(['male', 'male', 'male', 'male'])) # -0.0print(entropy(['male', 'female', 'male', 'male'])) # 0.8112781244591328print(entropy(['female', 'male', 'female', 'male'])) # 1.0print(entropy(['female', 'female', 'male', 'female'])) # 0.8112781244591328print(entropy(['female', 'female', 'female', 'female'])) # -0.0# 6.5 朴素贝叶斯分类器# 潜在概率模型# 零计数和平滑# 非二元特征# 独立的朴素贝叶斯# 双重计数的原因# 6.6 最大熵分类器# 6.7 为语言模式建模