""" 针对爬虫类型数据单独训练模型 """ import os import sys import json import optuna from sklearn.linear_model import LogisticRegression sys.path.append(os.getcwd()) import numpy as np import pandas as pd import lightgbm as lgb from scipy.stats import randint as sp_randint from scipy.stats import uniform as sp_uniform from sklearn.model_selection import RandomizedSearchCV, train_test_split from sklearn.metrics import roc_auc_score from bayes_opt import BayesianOptimization class LightGBM(object): """ LightGBM model for classification """ def __init__(self, flag, dt): self.label_encoder = LabelEncoder() self.my_c = [ "channel", "out_user_id", "mode", "out_play_cnt", "out_like_cnt", "out_share_cnt", "lop", "duration", "tag1", "tag2", "tag3" ] self.str_columns = ["channel", "mode", "out_user_id", "tag1", "tag2", "tag3"] self.float_columns = [ "out_play_cnt", "out_like_cnt", "out_share_cnt", "lop", "duration" ] self.split_c = 0.7 self.yc = 0.8 self.model = "models/lightgbm_0401_spider.bin" self.flag = flag self.dt = dt def read_data(self): """ Read data from local :return: """ path = "data/train_data/spider_data_240401.json" df = pd.read_json(path) df = df.dropna(subset=['label']) labels = df['label'] temp = sorted(labels) yc = temp[int(len(temp) * self.yc)] labels = [0 if i < yc else 1 for i in labels] features = df.drop("label", axis=1) for key in self.float_columns: features[key] = pd.to_numeric(features[key], errors="coerce") for key in self.str_columns: features[key] = self.label_encoder.fit_transform(features[key]) return features, labels def best_params(self): X, y = self.read_data() X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42) lgbM = lgb.LGBMClassifier(objective='binary') # 设置搜索的参数范围 param_dist = { 'num_leaves': sp_randint(20, 40), 'learning_rate': sp_uniform(0.001, 0.1), 'feature_fraction': sp_uniform(0.5, 0.9), 'bagging_fraction': sp_uniform(0.5, 0.9), 'bagging_freq': sp_randint(1, 10), 'min_child_samples': sp_randint(5, 100), } # 定义 RandomizedSearchCV rsearch = RandomizedSearchCV(estimator=lgbM, param_distributions=param_dist, n_iter=100, cv=3, scoring='roc_auc', random_state=42, verbose=2) # 开始搜索 rsearch.fit(X_train, y_train) # 打印最佳参数和对应的AUC得分 print("Best parameters found: ", rsearch.best_params_) print("Best AUC found: ", rsearch.best_score_) # 使用最佳参数在测试集上的表现 best_model = rsearch.best_estimator_ y_pred = best_model.predict_proba(X_test)[:, 1] auc = roc_auc_score(y_test, y_pred) print("AUC on test set: ", auc) def train_model(self): """ Load dataset :return: """ x, y = self.read_data() train_size = int(len(x) * self.split_c) X_train, X_test = x[:train_size], x[train_size:] Y_train, Y_test = y[:train_size], y[train_size:] train_data = lgb.Dataset( X_train, label=Y_train, categorical_feature=["channel", "mode", "out_user_id", "tag1", "tag2", "tag3"], ) test_data = lgb.Dataset(X_test, label=Y_test, reference=train_data) params = { 'num_leaves': 31, 'learning_rate': 0.00020616904432655601, 'feature_fraction': 0.6508847259863764, 'bagging_fraction': 0.7536774652478249, 'bagging_freq': 6, 'min_child_samples': 99, 'num_threads': 16 } # 训练模型 num_round = 100 print("开始训练......") bst = lgb.train(params, train_data, num_round, valid_sets=[test_data]) bst.save_model(self.model) print("模型训练完成✅") def evaluate_model(self): """ 评估模型性能 :return: """ fw = open("summary_tag_03{}_spider.txt".format(self.dt), "a+", encoding="utf-8") # 测试数据 with open("data/produce_data/x_data_total_return_predict_{}_spider.json".format(self.dt)) as f1: x_list = json.loads(f1.read()) # 测试 label with open("data/produce_data/y_data_total_return_predict_{}_spider.json".format(self.dt)) as f2: Y_test = json.loads(f2.read()) Y_test = [0 if i <= 19 else 1 for i in Y_test] X_test = pd.DataFrame(x_list, columns=self.my_c) for key in self.str_columns: X_test[key] = self.label_encoder.fit_transform(X_test[key]) for key in self.float_columns: X_test[key] = pd.to_numeric(X_test[key], errors="coerce") bst = lgb.Booster(model_file=self.model) y_pred = bst.predict(X_test, num_iteration=bst.best_iteration) temp = sorted(list(y_pred)) yuzhi = temp[int(len(temp) * 0.7) - 1] y_pred_binary = [0 if i <= yuzhi else 1 for i in list(y_pred)] # 转换为二进制输出 score_list = [] for index, item in enumerate(list(y_pred)): real_label = Y_test[index] score = item prid_label = y_pred_binary[index] print(real_label, "\t", prid_label, "\t", score) fw.write("{}\t{}\t{}\n".format(real_label, prid_label, score)) score_list.append(score) print("预测样本总量: {}".format(len(score_list))) data_series = pd.Series(score_list) print("统计 score 信息") print(data_series.describe()) # 评估模型 accuracy = accuracy_score(Y_test, y_pred_binary) print(f"Accuracy: {accuracy}") fw.close() def feature_importance(self): """ Get the importance of each feature :return: """ lgb_model = lgb.Booster(model_file=self.model) importance = lgb_model.feature_importance(importance_type='split') feature_name = lgb_model.feature_name() feature_importance = sorted(zip(feature_name, importance), key=lambda x: x[1], reverse=True) # 打印特征重要性 for name, imp in feature_importance: print(name, imp) if __name__ == "__main__": # i = int(input("输入 1 训练， 输入 2 预测:\n")) # if i == 1: # f = "train" # dt = "whole" # L = LightGBM(flag=f, dt=dt) # L.train_model() # elif i == 2: # f = "predict" # dt = int(input("输入日期， 16-21:\n")) # L = LightGBM(flag=f, dt=dt) # L.evaluate_model() # L.feature_importance() L = LightGBM("train", "whole") L.best_params() # study = optuna.create_study(direction='maximize') # study.optimize(L.bays_params, n_trials=100) # print('Number of finished trials:', len(study.trials)) # print('Best trial:', study.best_trial.params)