论文分享：马来西亚基于语义的推荐系统

。> Are you good Malaysia?

A Proposed Semantic Recommendation System for e-learning

A Rule and Ontology Based E-learning Recommendation System

http://www.academia.edu/download/49610177/A_proposed_semantic_recommendation_syste20161014-6780-ys9p6u.pdf

这篇文章是马来西亚国民大学在2010年所发表的论文，google scholar上显示引用数量为59。

我发现这些写推荐系统这方面的论文的大家伙有一个共同点，就是喜欢在introduction里面扯一大堆没有用的东西。当然该扯还是要扯的，但是能不能不要每篇文章扯的东西都那么相似啊？？

这篇文章的特殊之处，在于他提到了semantic这个概念。文章注意到，在线学习的使用者通常遇到如下几个问题：

• 找合适的学习资源比较花时间（因为各式各类的学习资源越来越多）；
• 搜索引擎对搜索学习资源的支持不是很好；
• 目前的在线教育系统普遍不够personalize，简单的CF（甚至是HF）推荐系统很多情况下不能满足用户需要。

因此，论文希望能够用 Rule-based 或者是 Ontology-based 类型的推荐系统，进行推荐。他们还做了一个对比图：

与其他论文不太相同的是，这一篇文章并没有详细的介绍他们的系统是如何实现的，只说到了要使用上述的算法。这可能跟这篇文章是一篇会议论文有关，字数限制比较死。

---

Abstract

As web-based e-learning systems becoming increasingly popular, many problems began to arise. Current e-learning systems do not have well support for searching learning materials, thus makes it difficult for students to have a personalized e-learning environment. Also, recommendation systems used in this system is not intelligent enough, which makes things worse. In this paper, we made a proposal for an e-learning system which is rule-based and ontology-based. This system aims to make use of current W3C-standard OWL rules, and try to analyze what content should be recommended by rule filtering.

Conclusion

e-learning platforms plays an important role in today’s education. As the amount of online course content becomes very large, providing personalized content recommendation is a significant functionality for e-learning systems. The recommender system can provide from a technical point of view, useful resources and tools, making it easier to obtain all learning resources. Thus, it assists learners to choose the most suitable method. Also, it offers the possibility of accurate user monitoring and evaluation during the learning process. This paper proposes to use a semantic web based recommendation approach using ontology and rule. The proposed rule based system consists of predefined rules and rule engine. In the next stage the system will evaluate learner’s knowledge and learner’s performances. The system will then present recommendation list according to the results of learner’s evaluations and profile.