基于协同过滤和室内三维定位的智慧校园精细化学生动态信息推荐研究  

Research on Fine Students'Dynamic Information Recommendation in Smart Campus Based on Collaborative Filtering and Indoor 3D Positioning

作　　者：周勇 李莉[2] 吴瑕 狄宏林 ZHOU Yong;LI Li;WU Xia;DI Honglin(Dongguan Open University,Dongguan 523000,China;Huizhou University,Huizhou 516000,China)

机构地区：[1]东莞开放大学,广东东莞523000 [2]惠州学院

出　　处：《通化师范学院学报》2024年第6期66-73,共8页Journal of Tonghua Normal University

基　　金：广东省哲学社会科学规划2022年度常规项目(GD22CJY08);2022年度广东远程开放教育科研基金项目重点课题(YJ2220);广东省2023年度教育科学规划课题(高等教育专项)(2023GXJK505);广东省2023年度省级一流本科课程“运筹学”(粤教高函[2023]33号);2023年度广东省高校思想政治教育课题(2023GXSZ067);惠州市2023年度基础教育教育科学研究课题(2023hzkt056)。

摘　　要：智慧校园作为一种全新的教育环境,由于其内部资源数量庞大,易出现信息过载的问题.当前的无差别信息推荐技术容易忽视学生的个体差异,难以满足不同年级和专业在校学生的具体需求.为了解决此类问题,研究首先利用基于改进K-means聚类算法的室内三维定位算法对学生位置进行定位,然后通过基于权重矩阵的协同过滤算法对在校学生需求进行精细化与动态化的信息推荐.结果表明,室内三维定位算法的定位准确率与误差分别为91.13%、0.930 1 m,明显优于改进前的室内三维定位算法.且当相关相似性与Tanimoto系数的调节参数β取值为0.4时,基于权重矩阵的协同过滤算法的MAE值普遍低于0.51.研究提出的室内三维定位算法与协同过滤算法性能优越,在智慧校园背景下,对于不同学生的需求能够做出适当推荐.As a new educational environment,the smart campus often has the problem of information over⁃load because of its huge amount of internal resources.However,the current undifferentiated information recommendation technology is easy to ignore the individual differences of students,and it is difficult to meet the specific needs of students in different grades and majors.In order to solve such problems,the research first uses the indoor three-dimensional positioning algorithm based on the improved K-means clustering algorithm to locate the students′positions,and then uses the collaborative filtering algorithm based on the weight matrix to recommend the students′demands for refined and dynamic information.The research results show that the positioning accuracy and error of the indoor 3D positioning algorithm are 91.13%and 0.9301 m,respectively,which is obviously better than the indoor three-dimensional position⁃ing algorithm before improvement.And when the adjustment parameter of correlation similarity and Tanimoto coefficient is 0.4,the MAE value of collaborative filtering algorithm based on weight matrix is generally lower than 0.51.To sum up,the indoor three-dimensional positioning algorithm and collabora⁃tive filtering algorithm proposed in this study have superior performance,and can make appropriate recommendations for different students′needs in the background of smart campus.

关 键 词：权重矩阵 协同过滤算法 K-MEANS聚类算法 信息推荐 智慧校园 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]