Abstract—In Taiwan, textbooks for elementary and junior high school education are edited in accordance with the Grade 1-9 Curriculum Guidelines, then submitted to the national committee for review and approval. Various educational experts from multiple disciplines bring different perspectives into textbooks to show the advantage of content diversity based on guidelines of the curriculum. However, different versions of textbooks result in the problem of “incoherent curriculum articulation.” That is the learning map of an individual student who studies textbooks from various publishers. The circumstance is worse for transfer students. Educators usually deal with this problem by focusing on the gap between different versions of textbooks and playing the tutor role in aiding transfer students. In this paper, we develop a system to automatically collect Curriculum Guidelines from three major textbook publishers (named by K, H, and N) during 2012-14 years. Using the domain “Nature and Technology” as a case study, the system extracts textbook and curriculum metadata of four grades (Grade 3-6) and merges them into the complete data set (24 textbooks: 4 grades × 2 semesters × 3 publishers). Applying text processing and mining methods, semantically same lessons from 3 publishers are firstly fused into a single class hierarchy. Then, we analyze lessons from different year-semesters to explore the life cycle of the lesson and the variation by time. Based on the result, educators can easily review the whole learning map of the Curriculum and observe the problem of incoherent curriculum articulation from the learning map as a consideration to assist students with some learning disabilities.

Index Terms—Text mining, curriculum articulation, catalog fusion, curriculum convergence.

Hao-Min Yan and Shian-Hua Lin are with Department of Computer Science and Information Science, National Chi Nan University, Puli, Taiwan (e-mail: shlin@ncnu.edu.tw).
Ming-Hsun Yang is with Department of Computer Science and Information Science, National Chi Nan University, Puli, Taiwan. He is also with Education Network Center of Chiayi County, Taiwan.

Cite: Hao-Min Yan, Ming-Hsun Yang, Shian-Hua Lin, "Lesson Fusion System (LFS) — Applying Text Analysis to Seal the Learning Gap from One-Guideline-Multiple-Text," International Journal of e-Education, e-Business, e-Management and e-Learning vol. 7, no. 1, pp. 1-12, 2017.