数据
3万文本,train val test 6 2 2.
工具、手法
pytorch、sklearn、gensim的word2vec。
word2vec嵌入句子进行表示,padding后,用LSTM+linear对句序列向量分类。
代码
import jieba
import xgboost as xgb
from sklearn.model_selection import train_test_split
import numpy as np
from gensim.models import Word2Vec
# reorganize data
def get_split_sentences(file_path):
res_sen=[]
with open(file_path) as f:
for line in f:
split_query=jieba.lcut(line.strip())
res_sen.append(split_query)
return res_sen
label2_sentences=get_split_sentences('label2.csv')
label0_sentences=get_split_sentences('label0.csv')
label1_sentences=get_split_sentences('label1.csv')
all_sentences=[]
all_sentences.extend(label0_sentences)
all_sentences.extend(label1_sentences)
all_sentences.extend(label2_sentences)
# set params
emb_size=128
win=3
model=Word2Vec(sentences=all_sentences,vector_size=emb_size,window=win,min_count=1)
# retrieve word embeddings
w2vec=model.wv
# assemble sentence embeddings
def assemble_x(w2vec:dict,sentences):
sen_vs=[]
for sen in sentences:
v=np.vstack([w2vec[w] for w in sen])
v_len=v.shape[0]
sen_v=np.concatenate((v,np.zeros((max_len-v_len,emb_size)))) if v_len
结果
ACC: 0.4303
macro:
Recall: 0.3333
F1-score: 0.2006
Precision: 0.1434
micro:
Recall: 0.4303
F1-score: 0.4303
Precision: 0.4303
小结
效果非常差,原因主要有
- padding的0向量过于多了,导致模型得到的大部分都是0向量;
- 并未对lstm做任何参数调整(懒