類神經網路在學生模組實際製作上之應用

Abstract

學生模組在智慧型教學系統中扮演著相當重要的角色。有了學生模組，教學系統在教學與學生學習之過程中就可滿足學生之個別需求。本文將利用類神經網路以建立學生模組，首先老師必須將課程結構與相關之教學步驟設計出，此兩資訊將被量化並放入類神經網路之定義項中，然後此類神經網路會取得學生在學習過程中的反應．並調整網路中之量化值。在量化值之調整中會記錄學生之學習歷程以推論出目前學生之學習狀態。依據此推論之學習狀態，類神經網路會提出學生最弱之學習子題以利教學系統能夠加強學生對該子題之學習。整個研究之目標是利用類神經網路建立學生模組以評估學生之學習狀態，並提供適當之意見給教學系統以決定下一個子題的教學。

The student model plays a critical role in intelligent tutoring systems. With a student model, the tutoring system can adapt the learning process in order to satisfy the individual needs of every student. Here we use neural network methods to implement the student model. In the proposed approach, first a teacher has to apply the curriculum structure and crucial pedagogical steps for the learning process. These two data will be transmitted as proper weights and layers in the network. Then our neural network takes the student's responses as input data and performs learning. During learning, the network records the student's history and changes the corresponding weights in order to record his or her learning state. Finally, the system infers the student's misunderstandings and provides suggestions to the tutor. Our goal is to make a student model embedded in a neural network, and evaluate misunderstandings in the learning processes of learners. Based on the experiment, the proposed neural network can infer the topic where the student's under-standing is weakest as a reference for the next lesson.