eCSCDA: An efficient system for analyzing contents of Computer Science Courses
Peerapon Kamlangpuech, Komate Amphawan · 2021
This paper aims to introduce a new efficient system, called eCSCDA, to efficient analyze Computer Science (CS) course description. The primary task of the system is to identify similar (and dissimilar) contents amount two (or a group of) CS course descriptions which can help to know similar and different focuses on important contents to teach to students. Moreover, it can help to check for integrity and quality and to set up a standard of teaching contents of the course. In eCSCDA, text processing procedure is newly rearranged and developed. Besides, the linguistic rules and their derivation are newly updated and applied to extracted important keywords. Moreover, two synonym corpuses, terminology and word synonyms, are recently designed and collected to consider synonyms of the keywords hidden in the course descriptions. Last, to efficiently identify similar contents, two new matching techniques, sub-keyword and semantic matching techniques, are designed and applied together with exact and subset (superset) matching methods. Experiments were conducted on CS course contents gathered from nine Thai Universities to examine the effectiveness of our proposed system in comparison with previous system and related methodologies. From the results, it shows that eCSCDA is efficient to analyze the course contents and outperforms other related systems in various terms e.g. percentage of similar contents, precision, recall and F-measure, respectively.