The Recommendation System Based On Core Knowledge Improvement With An Application On Economics Majored Undergraduates In National Taipei University

• Main Authors: WANG, SHENG-YAU, 王聖曜
• Other Authors: LIN, MAO-TING
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/mg45xb

碩士 === 國立臺北大學 === 經濟學系 === 107 === Most recommendation systems follow user’s existing preference or habits to recommend, which might not be a good idea for education efficacy purpose. To recommend courses that can enhance student’s major knowledge development, sometimes it is necessary to break his/her learning patterns. In this research we propose an exogenous intervention recommendation algorithm that can be divided into two stages of algorithm. The first stage is to estimate the effect of an exogenous change of course arrangement. The second stage is to recommend the collection of courses that produces the maximal exogenous impact on a student’s learning outcome. We use the data from the undergraduate students of National Taipei University who enrolled between 2011 and 2014 and majored Economics to test our algorithm for Economics selective courses recommendation. At the first stage, we chose neural network model over nonlinear regression model for estimating the effect of the exogenous change of selective courses on student’s learning efficacy of his/her required courses. At the second stage, the recommendation is formulated to maximize such efficacy.