面向水质分类的分组降维核朴素贝叶斯模型  

Study on Naive Bayes Model of Grouping Dimensionality Reduction Kernel for Water Quality Classification

作　　者：万玥 赖会霞[1] 钱伟[2] 张仕[1] WAN Yue;LAI Huixia;QIAN Wei;ZHANG Shi(College of Computer and Cyber Security,Fujian Normal University,Fuzhou,China,350117;School of Geographical Sciences,Fujian Normal University,Fuzhou,China,350117)

机构地区：[1]福建师范大学计算机与网络空间安全学院,福州350117 [2]福建师范大学地理科学学院,福州350117

出　　处：《福建电脑》2024年第3期18-23,共6页Journal of Fujian Computer

基　　金：福建省自然科学基金项目(No.2020J01161);福建省科技厅对外合作项目(No.2023I0013)资助。

摘　　要：为提高贝叶斯模型在实际数据中处理相关性维度时的效率、精度,并保持其可解释性,本文提出一种改进的核朴素贝叶斯模型。首先通过关联规则挖掘相关维度子集,然后对这些子集进行降维处理,利用降维后的数据构建核朴素贝叶斯模型。实际应用的结果显示,该模型减少了降维对数据的影响,在保留有效信息的同时,提高了模型的可解释性。To improve the efficiency and accuracy of Bayesian models in handling correlation dimensions in actual data,and maintain their interpretability,this paper proposes an improved kernel naive Bayesian model.Firstly,relevant dimension subsets are mined through association rules,and then these subsets are subjected to dimensionality reduction processing to construct a kernel naive Bayesian model using the reduced data.The practical application results show that the model reduces the impact of dimensionality reduction on data,while retaining effective information and improving the interpretability of the model.

关 键 词：分组降维 核朴素贝叶斯 贝叶斯网络 水质评估 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]