nlp計數法應用於PTB資料集範例詳解

2022-04-05 19:00:12

PTB資料集

內容如下：

一行儲存一個句子；將稀有單詞替換成特殊字元 < unk > ；將具體的數位替換成“N”

 we 're talking about years ago before anyone heard of asbestos having any questionable properties 
 there is no asbestos in our products now 
 neither <unk> nor the researchers who studied the workers were aware of any research on smokers of the kent cigarettes 
 we have no useful information on whether users are at risk said james a. <unk> of boston 's <unk> cancer institute 
 dr. <unk> led a team of researchers from the national cancer institute and the medical schools of harvard university and boston university

ptb.py

使用PTB資料集：

由下面這句話，可知用PTB資料集時候，是把所有句子首尾連線了。

words = open(file_path).read().replace('n', '<eos>').strip().split()

ptb.py起到了下載PTB資料集，把資料集存到資料夾某個位置，然後對資料集進行提取的功能，提取出corpus, word_to_id, id_to_word。

import sys
import os
sys.path.append('..')
try:
    import urllib.request
except ImportError:
    raise ImportError('Use Python3!')
import pickle
import numpy as np
url_base = 'https://raw.githubusercontent.com/tomsercu/lstm/master/data/'
key_file = {
    'train':'ptb.train.txt',
    'test':'ptb.test.txt',
    'valid':'ptb.valid.txt'
}
save_file = {
    'train':'ptb.train.npy',
    'test':'ptb.test.npy',
    'valid':'ptb.valid.npy'
}
vocab_file = 'ptb.vocab.pkl'
dataset_dir = os.path.dirname(os.path.abspath(__file__))
def _download(file_name):
    file_path = dataset_dir + '/' + file_name
    if os.path.exists(file_path):
        return
    print('Downloading ' + file_name + ' ... ')
    try:
        urllib.request.urlretrieve(url_base + file_name, file_path)
    except urllib.error.URLError:
        import ssl
        ssl._create_default_https_context = ssl._create_unverified_context
        urllib.request.urlretrieve(url_base + file_name, file_path)
    print('Done')
def load_vocab():
    vocab_path = dataset_dir + '/' + vocab_file
    if os.path.exists(vocab_path):
        with open(vocab_path, 'rb') as f:
            word_to_id, id_to_word = pickle.load(f)
        return word_to_id, id_to_word
    word_to_id = {}
    id_to_word = {}
    data_type = 'train'
    file_name = key_file[data_type]
    file_path = dataset_dir + '/' + file_name
    _download(file_name)
    words = open(file_path).read().replace('n', '<eos>').strip().split()
    for i, word in enumerate(words):
        if word not in word_to_id:
            tmp_id = len(word_to_id)
            word_to_id[word] = tmp_id
            id_to_word[tmp_id] = word
    with open(vocab_path, 'wb') as f:
        pickle.dump((word_to_id, id_to_word), f)
    return word_to_id, id_to_word
def load_data(data_type='train'):
    '''
        :param data_type: 資料的種類：'train' or 'test' or 'valid (val)'
        :return:
    '''
    if data_type == 'val': data_type = 'valid'
    save_path = dataset_dir + '/' + save_file[data_type]
    word_to_id, id_to_word = load_vocab()
    if os.path.exists(save_path):
        corpus = np.load(save_path)
        return corpus, word_to_id, id_to_word
    file_name = key_file[data_type]
    file_path = dataset_dir + '/' + file_name
    _download(file_name)
    words = open(file_path).read().replace('n', '<eos>').strip().split()
    corpus = np.array([word_to_id[w] for w in words])
    np.save(save_path, corpus)
    return corpus, word_to_id, id_to_word
if __name__ == '__main__':
    for data_type in ('train', 'val', 'test'):
        load_data(data_type)

使用ptb.py

corpus儲存了單詞ID列表，id_to_word 是將單詞ID轉化為單詞的字典，word_to_id 是將單詞轉化為單詞ID的字典。

使用ptb.load_data()載入資料。裡面的引數 ‘train’、‘test’、‘valid’ 分別對應訓練用資料、測試用資料、驗證用資料。

import sys
sys.path.append('..')
from dataset import ptb
corpus, word_to_id, id_to_word = ptb.load_data('train')
print('corpus size:', len(corpus))
print('corpus[:30]:', corpus[:30])
print()
print('id_to_word[0]:', id_to_word[0])
print('id_to_word[1]:', id_to_word[1])
print('id_to_word[2]:', id_to_word[2])
print()
print("word_to_id['car']:", word_to_id['car'])
print("word_to_id['happy']:", word_to_id['happy'])
print("word_to_id['lexus']:", word_to_id['lexus'])

結果：

corpus size: 929589
corpus[:30]: [ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
 24 25 26 27 28 29]
id_to_word[0]: aer
id_to_word[1]: banknote
id_to_word[2]: berlitz
word_to_id['car']: 3856
word_to_id['happy']: 4428
word_to_id['lexus']: 7426
Process finished with exit code 0

計數方法應用於PTB資料集

其實和不用PTB資料集的區別就在於這句話。

corpus, word_to_id, id_to_word = ptb.load_data('train')

下面這句話起降維的效果

word_vecs = U[:, :wordvec_size]

整個程式碼其實耗時最大的是在下面這個函數上：

W = ppmi(C, verbose=True)

完整程式碼：

import sys
sys.path.append('..')
import numpy as np
from common.util import most_similar, create_co_matrix, ppmi
from dataset import ptb
window_size = 2
wordvec_size = 100
corpus, word_to_id, id_to_word = ptb.load_data('train')
vocab_size = len(word_to_id)
print('counting  co-occurrence ...')
C = create_co_matrix(corpus, vocab_size, window_size)
print('calculating PPMI ...')
W = ppmi(C, verbose=True)
print('calculating SVD ...')
#try:
    # truncated SVD (fast!)
print("ok")
from sklearn.utils.extmath import randomized_svd
U, S, V = randomized_svd(W, n_components=wordvec_size, n_iter=5,
                             random_state=None)
#except ImportError:
    # SVD (slow)
    # U, S, V = np.linalg.svd(W)
word_vecs = U[:, :wordvec_size]
querys = ['you', 'year', 'car', 'toyota']
for query in querys:
    most_similar(query, word_to_id, id_to_word, word_vecs, top=5)

下面這個是用普通的np.linalg.svd(W)做出的結果。

[query] you
 i: 0.7016294002532959
 we: 0.6388039588928223
 anybody: 0.5868048667907715
 do: 0.5612815618515015
 'll: 0.512611985206604
[query] year
 month: 0.6957005262374878
 quarter: 0.691483736038208
 earlier: 0.6661213636398315
 last: 0.6327787041664124
 third: 0.6230476498603821
[query] car
 luxury: 0.6767407655715942
 auto: 0.6339930295944214
 vehicle: 0.5972712635993958
 cars: 0.5888376235961914
 truck: 0.5693157315254211
[query] toyota
 motor: 0.7481387853622437
 nissan: 0.7147319316864014
 motors: 0.6946366429328918
 lexus: 0.6553674340248108
 honda: 0.6343469619750977

下面結果，是用了sklearn模組裡面的randomized_svd方法，使用了亂數的 Truncated SVD，僅對奇異值較大的部分進行計算，計算速度比常規的 SVD 快。

calculating SVD ...
ok
[query] you
 i: 0.6678948998451233
 we: 0.6213737726211548
 something: 0.560122013092041
 do: 0.5594725608825684
 someone: 0.5490139126777649
[query] year
 month: 0.6444296836853027
 quarter: 0.6192560791969299
 next: 0.6152222156524658
 fiscal: 0.5712860226631165
 earlier: 0.5641934871673584
[query] car
 luxury: 0.6612467765808105
 auto: 0.6166062355041504
 corsica: 0.5270425081253052
 cars: 0.5142025947570801
 truck: 0.5030257105827332
[query] toyota
 motor: 0.7747215628623962
 motors: 0.6871038675308228
 lexus: 0.6786072850227356
 nissan: 0.6618651151657104
 mazda: 0.6237337589263916
Process finished with exit code 0

以上就是nlp計數法應用於PTB資料集範例詳解的詳細內容，更多關於nlp計數法應用於PTB資料集的資料請關注it145.com其它相關文章！

nlp計數法應用於PTB資料集範例詳解

目錄

PTB資料集

ptb.py

使用ptb.py

計數方法應用於PTB資料集

熱門文章