基于个性化自适应的网络混合式信息推荐  

Personalized Adaptive Network Hybrid Information Recommendation

作　　者：刘颖[1] 韦妙[1] LIU Ying;WEI Miao(Hubei University of Technology,Wuhan Hubei 430000,China)

机构地区：[1]湖北工业大学,湖北武汉430000

出　　处：《计算机仿真》2021年第4期399-402,416,共5页Computer Simulation

基　　金：湖北省教育科学规划重点课题(2017GA018)。

摘　　要：针对传统网络混合式信息推荐满意度和精度低的问题,提出了基于个性化自适应的网络混合式信息推荐。首先采用模块化矩阵对子因素序列进行模块化处理,计算出每一个子因素序列对应的关联系数均值,利用耦合度函数研究网络混合式信息之间的耦合关系,完成网络混合式信息的模块化处理。然后通过判断网络混合式信息的权重值是否高于提前设定好的门限值,计算网络混合式信息的权重,通过定义网络混合式信息之间的相似度,得到信息集中网络混合式信息的相似度分布情况;利用个性化自适应学习改进了网络混合式信息的兴趣取向,通过计算用户的偏好相似度,设计网络混合式信息推荐算法,实现网络混合式信息的推荐。实验结果表明,所设计方法在网络混合式信息推荐的准确率、召回率和覆盖率上分别提高了3.91%、3.45%和4.84%,使网络混合式信息的推荐性能明显提高。Traditional network hybrid information recommendation has low satisfaction and accuracy. In this regard, a method of network hybrid information recommendation based on personalized adaptive was put forward in this paper.First of all, the modular matrix was adopted to process the sub-factor sequence and calculate the mean value of the correlation coefficient corresponding to each sub-factor sequence.The coupling degree function was utilized to study the coupling relationship between network hybrid information, thus completing the modular processing of network hybrid information.Secondly, the difference between the weight value of network hybrid information and the set threshold value was judged to calculate the weight of network hybrid information, and then, based on the definition of similarity between network hybrid information, the similarity distribution state of network hybrid information in information set was obtained.Finally, the personalized adaptive learning method was used to improve the interest orientation of network hybrid information. According to the calculation results of user preference similarity, the recommendation algorithm of network hybrid information was designed, achieving the recommendation of network hybrid information.The results show that the proposed method significantly improves the recommendation performance of network hybrid information, and the accuracy rate, recall rate, and coverage rate are increased by 3.91%,3.45%,and 4.84%,respectively.

关 键 词：个性化 自适应 网络混合式信息 信息推荐 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]