運用Petri-Net建構基本電學知識構圖及初學者認知途徑以強化適性學習之研究

No Thumbnail Available

Date

2021

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

電機與電子領域專業知識概念間存在錯綜複雜的關係,具有相同學習成效的學生可能存在不同的迷思概念與結構,而多數既有學習導引機制缺乏推薦學習者適性學習內容與途徑的規劃,因此,技職專業人才的培育面臨極大的障礙與困境。由於科技的進步,知識構圖被廣為應用於推薦系統,找出符合使用者需求的新訊息。本研究由5 位技術型高中電機與電子群專業科目知識概念專家、8 位試題發展專家與5 位電機與電子群業界專家及資深教師,運用三回合修正式德菲法剖析電機電子領域核心基礎:「基本電學」12個主題的58個專業概念(其中,篩選出4個奠基概念、4個核心概念及11個綜整性概念)與95個對應相關性。利用具有圖形特性的派翠西網路 (Petri-Net) 技術,進而建構出「基本電學」Petri-Net知識構圖,並發現對後續學習影響最鉅的概念依序為電路型態及特性、電的單位、向量運算及電壓。此外,本研究以初探性導入建立不同學習類型個案學生的Petri-Net知識構圖,並剖析他們個別的學習歷程與狀態,結果顯示Petri-Net知識構圖的應用得以:1.提供視覺化學習鷹架增強初學者的認知結構;2.適性診斷不同類型學習者的迷思概念;3. 推估後續概念的學習效果;4. 推薦個人化學習內容與路徑助益自主學習與補救教學。此外,運用Petri-Net知識構圖所視覺化呈現的學習認知途徑,能有效分析並導引初學者適性化的學習;763份評量紀錄的回歸分析結果顯示,除「4-2迴路電流法」、「5-1電容器」及「6-2電感器」三個綱要概念外,初學者在基本電學各概念的知識概念構圖模式均能顯著預測其效標概念的學習成效。據此,本研究提出相關的討論與建議,作為發展測評系統及擬定教學策略參考。
There exists an intricate relationship between professional knowledge concepts in the electrical and electronic engineering fields. Students with the same learning performance of professions might have an extremely different understanding from each other, and so do individual's concept structure. However, most of existing learning guidance mechanisms could not recommend adaptive and personalized learning contents and pathways to the learner. Thus, it was faced with serious barriers and difficulties to cultivate professional talents. With the advantages of information technologies, the Knowledge Graph (KG) has been widely applied in the recommender systems to facilitate the representation of knowledge structure and mining new messages or knowledge that meets user's needs. This study applied a three-round modified Delphi approach conducted by 18 domain experts to identifying 58 concepts (four cornerstone conceptions, four keystone conceptions, and 11 capstone conceptions were highlighted), and the corresponding interdependence relationship of the core course:"fundamental-electricity" in the electrical and electronic engineering domain, and then the Petri-Net technology with graphic features was used to construct its KG, so-called Expert Petri-Net KG. The preliminary exploration case studies were conducted to create personalized Petri-Net KG for three different learning types of students and to analyze their learning progress and status. Finally, 736 assessment records were used for regression analysis. The major results and findings of this study would be depicted below: 1. The"Circuit pattern and characteristics" is the most important concept, which affects the learning of the subsequent 12 concepts, and the total impact reaches to 6. Followed by concepts of "units", "vector operations" and "voltage" in order. 2. The proposed Petri-Net KG provides students with a visualized learning scaffolding for discovering experts' cognitive structure. It also clarifies those prior concepts for each conception. 3. By utilization of weights of inter-relationships between conceptsand their prior concepts, the reasoning engine would adaptively diagnose their misconceptions, and further predict student's learning effectiveness of subsequent concepts. 4. Different learning types of students have different types of cognitive structures. By integrating student's learning portfolio data into the proposed Petri-Net KG, the reasoning engine would recommend an adaptive and personalized learning pathway. 5. On the other side, novices' knowledge concept map models of fundamental-electricity can be used to predict learning performance of "criterion concept" significantly, exception of criterion concept"4-2 cyclic current method", "5-1capacitor" and "6-2 inductance". Reasons for the findings and implications for future research are discussed.

Description

Keywords

初學者, 知識構圖, 派翠西網路, 認知途徑, 適性學習, Adaptive Learning, Cognition Pathway, Novices, Knowledge Graph, Petri-Net

Citation

Collections