| Summary: | 碩士 === 國立屏東大學 === 資訊管理學系碩士班 === 108 === The Educational Objective is to educate and shape the talents of the students, and the core competence is the abilities that students need to poss to achieve these Educational Objectives. If the correspondence between Educational Objectives and core competencies is not perfect, it may cause certain Educational Objectives to be difficult to achieve. Colleges and universities usually develop their Educational Objectives and core competencies by recruiting scholars and experts inside and outside the school to determine the corresponding relationship manually. The advantage of this approach is that its correspondence is usually in line with human intuition, but the pitfall is that when the correspondences increase, it is often impossible to clearly explain the rationale behind each correspondence. To tackle this problem, this study employed word2vec (a deep learning tool) to obtain the semantic vectors of core competence and educational objective, and by calculating the similarity (using cosine measure) between each pair of semantic vectors (core competence and educational objective) to decide which core competence has correspondence with educational objective. The word2vec is tuned by using the dump of Chinese Wikipedia with different model structures (CBOW or Skip-gram) and different dimensions to find the best combination. The data used in this thesis includes 15 colleges and departments, where the undergraduate divsions of National Pingtung University Department of Information Management and Providence University Department of Computer Science and Information Management, Providence University Department of Computer Science and Information Engineering, and National University of Kaohsiung Department of Chemical and Materials Engineering are used as tuning data to decide the best word2vec model. The rest of the analyses include 9 comparisons between the department's and college’s Educational Objectives, 8 department’s and college’s core competencies, and 18 department’s Educational Objectives and department’s core competencies. The empirical results show that the consistency in the correspondences between word2vec and experts is as high as 70%. It shows that word2vec can really help determine the correspondences between Educational Objectives and core competencies, and find out the inappropriate relationship in them. The scores of the analyzed items for each department are relatively close. Because the correspondence provided by each department are mainly determined by the same group of experts and scholars. Word2Vec can provide more consistent correspondence analysis thus can provide a tool for evaluating and evaluated institutions to examine the consistency of the correspondences.
|
|---|
