%0 Journal Article
%T 面向短视频账号浅度安全等级识别方法——短视频潜在优质账号的初步划分<br>Short Video Account Shallow Security Level Identification Method—Preliminary Division of Potential High-Quality Accounts for Short Videos
%A 吴志浩
%A 郭晓军
%A 闫宇辰
%A 李金涛
%J Hans Journal of Data Mining
%P 165-172
%@ 2163-1468
%D 2023
%I Hans Publishing
%R 10.12677/HJDM.2023.132016
%X 针对短视频账号信息内容浅度安全等级分类识别的短视频官方未认证账号进行初级的潜在优质用户划分进行研究。由于账号信息内容浅度安全等级分类识别的特征值相对不多，而且本文要的需求是保证其识别的准确率，故改进了KNN算法，提出选择使用GM-KNN模型来对其进行分类识别。通过网格搜索算法确定了K值的最优参数，将不同距离进行对比，选择了该模型的最优距离曼哈顿距离，实现了潜在优质账号的初步划分。经过实验对比，GM-KNN模型准确率在短视频潜在优质账号的初步划分中均优于其他对比算法。<br />
The research is on the primary potential high-quality user division of short video official unauthenticated accounts identified by shallow security level classification of short video account information content. Since there are relatively few eigenvalues for the classification and identification of shallow security levels of account information, and the requirement of this paper is to ensure the accuracy of its identification, the KNN algorithm is improved, and the GM-KNN model is proposed to be used for classification and identification. The optimal parameter of the K value was determined by the grid search algorithm, and the different distances were compared, and the optimal distance of the model was selected, the Manhattan distance, and the preliminary division of potential high-quality accounts was realized. After experimental comparison, the accuracy of GM-KNN model is better than other comparative algorithms in the preliminary division of potential high-quality accounts of short videos.
%K 账号安全类型识别，KNN算法，网格搜索算法，曼哈顿距离
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=64200