train_lgb_classifer.py

# -*- coding:utf-8 -*-
'''
-------------------------------------------------
   Description :  DNN   trainer
   Author :       machinelp
   Date :         2020-06-06
-------------------------------------------------

'''

import json 
import numpy as np
from textmatch.models.text_embedding.model_factory_sklearn import ModelFactory

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn import metrics
from sklearn.preprocessing import LabelEncoder,OneHotEncoder
from textmatch.models.ml.lgb import LGB
from textmatch.models.ml.lr import LR
from textmatch.models.ml.gbdt import GBDT
from textmatch.models.ml.gbdt_lr import GBDTLR
from textmatch.models.ml.xgb import XGB


if __name__ == '__main__':
    doc_dict = {"0":"我去玉龙雪山并且喜欢玉龙雪山玉龙雪山", "1":"我在玉龙雪山并且喜欢玉龙雪山", "2":"我在九寨沟", "3":"你好"}   #["我去玉龙雪山并且喜欢玉龙雪山玉龙雪山","我在玉龙雪山并且喜欢玉龙雪山","我在九寨沟"]
    #doc_dict = {"0":"This is the first document.", "1":"This is the second second document.", "2":"And the third one."}
    #query = "This is the second second document."
    query = [ "我在玉龙雪山并且喜欢玉龙雪山", "我在玉龙雪山并且喜欢玉龙雪山", "我在玉龙雪山并且喜欢玉龙雪山","我在玉龙雪山并且喜欢玉龙雪山", "我在九寨沟,很喜欢", "我在九寨沟,很喜欢","我在九寨沟,很喜欢", "我在九寨沟,很喜欢", "我在九寨沟,很喜欢","我在九寨沟,很喜欢", "我在九寨沟,很喜欢", "我在玉龙雪山并且喜欢玉龙雪山"]
    train_labels = [0,0,0,0,1,1,1,1,1,1,1,0]
    
    # 基于bow
    mf = ModelFactory( match_models=['bow', 'tfidf', 'ngram_tfidf', 'albert'] )
    #mf.init(words_dict=doc_dict, update=True)
    mf.init(update=False)
    train_sample = []
    for per_query in query:
        bow_pre = mf.predict_emb(per_query)
        # print ('pre>>>>>', bow_pre)
        per_train_sample=[]
        for per_v in bow_pre.values():
           per_train_sample.extend( per_v )
        train_sample.append(per_train_sample)
    print ('train_sample, train_labels', train_sample, train_labels) 
    #print ('train_sample:::::', len(train_sample[0])) 
    train_x = np.array( train_sample[:10] )
    train_y = train_labels[:10]
    val_x = np.array(  train_sample[10:12] )
    val_y = train_labels[10:12]
    print ('val_y:', val_y)

    
    lgb = LGB()
    lgb.fit( train_x, train_y )
    res = lgb.predict(val_x)
    print ('>>>>', res)