基于改进神经网络算法的英语数字资源个性化推荐方法  

Research on Personalized Recommendation of English Digital Resources Based on Improved Neural Network Algorithm

作　　者：强薇[1] Qiang Wei(Suzhou Industrial Park Institute of Service Outsourcing,Suzhou 215123)

机构地区：[1]苏州工业园区服务外包职业学院,江苏苏州215123

出　　处：《中阿科技论坛（中英文）》2023年第10期90-94,共5页China-Arab States Science and Technology Forum

基　　金：2022年度高校哲学社会科学研究项目“以产出导向为理论基础的以赛促学高职英语教学模式研究”(2022SJYB1605)。

摘　　要：为了有效解决英语数字资源的个性化推荐结果不准确等问题,文章提出一种基于改进神经网络算法的英语数字资源个性化推荐方法。计算时间序列内滑动窗口内的数据均值,获取起始序列向量;将用户行为加以分类处理,形成多个规格一致的时间片,采用取样法对用户群体进行统计,得到各类型用户的行为状态定性;将平均查询频率作为标准,观察用户的查询行为特征,得到用户行为特征挖掘结果。在改进神经网络中引入元数据概念,构建以数字资源为基础的英语数字资源本体,对用户偏好以及英语数字资源本体双重聚类后,匹配类之间的拟合关系,确定最终推荐的英语数字资源。应用结果表明,该方法可有效提升英语数字资源个性化推荐的质量,减少个性化推荐耗时。To address the problem of inaccurate personalized recommendation of English digital resources,this article proposes a new method based on an improved neural network algorithm.Firstly,the mean value of the data within the sliding window of the time series is calculated to obtain the starting sequence vector;Then,user behavior is classified and processed to form multiple time slices with consistent specifications,and sampling method is adopted to statistically analyze user groups to obtain the qualitative behavior status of each type of user;With average query frequency as the criterion,the user's query behavior are obser ved to obtain user behavior features.Finally,the concept of metadata is introduced into the improved neural network to construct an English digital resource ontology.After double clustering of user preferences as well as the ontology,the fitting relationship between classes are matched to determine the recommended English digital resource.The results show that this method is less time-consuming and can effectively boost the accuracy of personalized recommendation.

关 键 词：改进神经网络算法 英语数字资源 个性化推荐 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]