基于JSP网页技术的高校PU系统功能可视化界面设计

The visual interface design of PU system function in colleges and universities based on JSP web page technology

针对高职院校第二课堂活动开展过程中存在创业项目推荐效果差的问题,提出基于JSP网页技术的高校PU系统功能可视化界面设计。首先设计一个JSP网页技术的高校PU系统,然后以创业项目的文本信息作为基础数据,提出基于深度神经网络和矩阵分解的创业项目推荐算法(DNN-MF)从而通过PU系统实现大学生创业项目的准确推荐。结果表明,在推荐算法方面、不同数量特征方面,提出的DNN-MF算法的RMSE为0.605 5,对比于PMF算法和ConvMF算法,分别提升了9.93%和2.97%,且在召回率指标中,本算法的召回率均高于另外两种算法;在系统测试方面,PU系统可实现功能可视化,满足系统需求。综合分析可知,PU系统在创业项目内容推荐中可取得较高的精度,推荐效果显著提升,可在高校第二课堂中进行应用开展。

In view of the problem of poor recommendation effect of entrepreneurial projects in the process of carrying out the second classroom activities in higher vocational colleges, the visual interface design of PU system function in colleges and universities based on JSP web page technology is proposed. First of all, a PU system in colleges and universities based on JSP web page technology is designed, and then taking the text information of the entrepreneurial projects as the basic data, the entrepreneurial project recommendation algorithm based on deep neural network and matrix factorization(DNN-MF) is propose to realize the accurate recommendation of college students’ entrepreneurial projects through the PU system. The results show that in terms of recommendation algorithm and different number of features, the RMSE of the proposed DNN-MF algorithm is 0.605 5, which is improved by 9.93% and 2.97% compared with the PMF algorithm and ConvMF algorithm, respectively, and in the recall indicator, the recall of the proposed algorithm is higher than that of the other two algorithms. Moreover, in terms of system testing, PU systems can visualize functions to meet system requirements. Comprehensive analysis shows that the PU system can achieve high accuracy in the content recommendation of entrepreneurial projects, and the recommendation effect is significantly improved, which can be applied in the second classroom of colleges and universities.