集成特征选择算法和LightGBM融合的分类模型

A Classification Model Based on Integrated Feature Selection Algorithm and LightGBM

实体零售网点具有数据特征种类多、数据结构复杂的特点。为了对实体零售网点进行分类分级管理,本文采用改进的集成特征选择算法ISFA和LightGBM算法对零售网点进行数据特征优化选择和分类分级。首先采用集成特征选择方法对零售网点的特征进行筛选,然后用LightGBM算法对筛选出的特征子集进行分类和预测。以网点基本数据和统计数据为对象进行对比实验。结果表明,对于不同网点的分类目标,使用本文的方法可取得良好的效果。

Physical retail outlets have the characteristics of many kinds of data and complex data structure.In order to carry out classification and classification management of physical retail outlets,this paper uses the improved integrated feature selection algorithm ISFA and lightgbm algorithm to optimize the data feature,classification of retail outlets.Firstly,the integrated feature selection method is used to screen the features of retail outlets,and then the selected feature subset is classified and predicted by lightgbm algorithm.Take the basic data and statistical data of outlets as the object of comparative experiment.The results show that for the classification objectives of different outlets,the method in this paper can achieve good results.