# -*- coding: utf-8 -*-
"""
Created on Fri Aug 10 10:12:56 2018
@author: HP
"""
from NLP import NLP
import pandas as pd
common_words_filename=r'D:\Mainland\Campus Life\ZXtrainee\data\common_words.txt'
model_place='D:\\Mainland\\Campus Life\\ZXtrainee\\data\\zhwiki_latest\\model\\'
save_result_place='D:\\Mainland\\Campus Life\\ZXtrainee\\data\\'
min_count=3
size=100
saveflag=0
def calculate_nlpv(min_count,size,saveflag):
nlp=NLP()
nlp.load_model(model_place+str(min_count)+'_'+str(size)+'\\wiki_nlp_'+str(min_count)+'_'+str(size)+'.model')
words=[]
fp=open(common_words_filename,'r',encoding='utf-8')
for line in fp.readlines():
try:
temp=nlp.model.wv.__getitem__(line.strip('\n')) # 检查词是否在word2vec模型中
words.append(line.strip('\n'))
except:
continue
try:
vector=nlp.nlp_vector(words)
except:
vector=nlp.safe_nlp_vector(words)
if saveflag:
vector_transform = vector.T
names=[]
for index in nlp.Label_index:
names.append(index+'w2v')
for index in nlp.Label_index:
names.append(index+'wn')
text = pd.DataFrame(columns = names, data = vector_transform)
#生成的csv文件的地址
text.to_csv(save_result_place+'common_words_vector_'+str(min_count)+'_'+str(size)+'.csv')
calculate_nlpv(min_count, size, saveflag=1)
calculate_nlpv(3,200, saveflag=1)
calculate_nlpv(3,300, saveflag=1)
calculate_nlpv(4,500, saveflag=1)
calculate_nlpv(5,100, saveflag=1)
calculate_nlpv(5,200, saveflag=1)
calculate_nlpv(5,300, saveflag=1)
calculate_nlpv(5,400, saveflag=1)
