Recommender system to ensure academic success

• 主要作者: M Heriz Putra Yusoff
• 其他作者: Andy Khong W H
• 格式: Final Year Project
• 語言: English
• 出版: Nanyang Technological University 2023
• 主題: Engineering::Computer science and engineering::Software::Software engineering
• 在線閱讀: https://hdl.handle.net/10356/166943
• 機構: Nanyang Technological University

In recent years, the application of recommender systems has been widely implemented among web services such as e-commerce, e-government [1], e-learning [2], [3], [4], among many others [5]. Recommender system has been utilised as a tool to predict and suggest relevant items to users. In the area of education, previous studies have researched the Intervention recommendation system which plays a significant role in advising students for their future, such as forecasting student academic performance based on past data, suggesting student’s curricula [2], student career forecasting [7], and graduation forecast on time [8]. At Nanyang Technological University (NTU), student care managers (SCM) ensure students have access to the assistance they require when they face difficulties in achieving academic success. To decrease assessment time for timely intervention while alleviating shortage of SCMs, a recommender system is implemented. This paper implements the use of Knowledge Graph Attention Network (KGAT) as a solution for intervention suggestion strategy and framework to enhance academic success in NTU. The KGAT presents a powerful deep learning model that has been proposed to enhance the recommendation system by incorporating the graph structure of the consumer-article interactions. The KGAT model is based on message propagation and the attention mechanism, allowing the model to learn the significance of the different parts of the knowledge graph and achieve distant-dimensional relationships to make precise, varied, and explainable suggestions. The KGAT model has shown to outperform the state-of-the-art recommendation models on several datasets. It has also shown to be effective in capturing the complex relationships between the consumer and article nodes in the knowledge graph. This model can be easily map and applied in the area of intervention recommender strategy to support student’s academic success in NTU, where students are the users and intervention recommendation are the items.