The 6th International Conference on Information Technology and Applications (ICITA 2009)
An Automated University Admission Recommender System for Secondary School Students Simon Fong and Robert P. Biuk-Aghai
neural network approach, an improved recommendation output can be achieved. The remainder of this paper is organized as follows. Section II discusses the problem of student recommendation in more detail. Section III then presents the design of our recommender system, RSAU (Recommender System of Admission to University), and Section IV evaluates the performance of our system. Section V discusses the decision rules used by RSAU. Finally, Section VI makes conclusions. II. RECOMMENDATION PROBLEM The choice of a university that is suitable for a given secondary school graduate can be a difficult decision to make. Reputation of the university, perceived difficulty of the degree program, distance from home, tuition and living costs, student’s areas of academic strength as well as actual scores achieved are just some of the factors that may be considered by a student graduating from secondary school. Likewise, the university has its own set of admission criteria, mainly based on academic standard of the student to be admitted, but possibly also including others, such as minority and gender representation, local vs. domestic vs. overseas student proportion, and others. Choosing the most suitable among the many thousands of candidates that apply to a university every year is not a trivial matter. Some universities avoid many of these issues through a simple unified admission process, such as pre-defined secondary school completion scores required for admission. This approach, however, does not always result in the most suitable candidates reaching the university “best” for them. Moreover, many accepted candidates end up not taking up the offer extended to them, resulting in wasted administrative effort on the part of the university. Most importantly, however, it assumes that secondary schools have comparable education standards and curricula. When this assumption is not valid then student selection for university admission becomes problematic. In this paper we focus on this particular problem, and show in the sections that follow our design of a recommender system suitable to countries and territories in which country-/territory-wide education standards are not defined. For illustration purposes and as a case study we examine this problem in detail for the case of Macau. Macau is a small territory, formerly a colony of Portugal and since 1999 a Special Administrative Region of China. It has a population of about 0.5 million, mainly of Chinese background but also including about 5% of non-Chinese speaking nationals, including those for whom Portuguese, English, and others are the languages primarily spoken. Moreover, the Chinese
Abstract—University or college admission is a complex decision process that goes beyond simply matching test scores and admission requirements. Past research has suggested that students’ backgrounds and other factors correlate to the performance of their tertiary education. However, almost all admission and enrollment studies are based on the perspective of universities or colleges, and only few studies are based on the perspective of secondary schools. This paper presents a hybrid model of neural network and decision tree classifier that serves as the core design for a university admission recommender system. The system was tested with live data from sources of Macau secondary school students. In addition to the high prediction accuracy rate, flexibility is an advantage such that the system can predict suitable universities that match the students’ profiles and the suitable approaches through which the students should enter. The recommender can be generalized into making different kinds of predictions based on the students’ histories.
Index Terms—university admission, recommender system,...
Please join StudyMode to read the full document