DSpace

Future University Hakodate Academic Archive >
Faculty and Students >
Dept. of Media Architecture >
Ito, Kei >

Please use this identifier to cite or link to this item: http://hdl.handle.net/10445/8818

Title: LMSを用いたプログラミング授業における機械学習による得点率予測
Prediction of Score by Machine Learning in Programming Class using LMS
Authors: 松尾, 龍磨
伊藤, 恵
Abstract: 近年, 様々な教育機関でプログラミング教育が進められており, プログラミング学習者が増加している. しかし, 個別での指導体制は難しく, 多人数授業になりがちで, 単位不認定者はますます増加してしまう. さらに, 習熟度に応じた授業が期待されているが, 多人数授業のため, 教員が個々の学生の理解度を把握することは難しい. また, 学生の理解度が低いままで次の演習に進んでしまい, 定期試験の点数が振るわない結果になってしまうという現状にある. そこで, 授業内に収集される学習データを分析することで, 教員が学生のつまずきをいち早く発見し, 単位不認定の可能性のある学生に適切な対応ができるようにすることを目指す. 本研究では, 機械学習を用い試験の得点率の予測を試みる. プログラミング授業におけるLMS(Learning Management System) 内の過去6 年分×約180 人分の学習データを用い, データ間の関係性やパターンを特定し, 目的変数に有益な特徴量を設定することで, 精度の高い予測を目指す. これにより, 教員が理解度の低い学生を早期に発見でき, 単位不認定を防ぎ, 期末試験で点数を取ってもらうように適切な対応ができる. Recent years, programming education is being promoted at various educational institutions, and programming learners are increasing. However, individual instruction system is difficult and the number of non-credit students increases due to the large number of classes. Furthermore, lessons are expected depending on the degree of understanding. But it is difficult for teachers to grasp the degree of understanding of each student. Also, students proceed to the next exercise with a poor understanding the content, and they can not get the good score for the test. Therefore, by analyzing the training data in the class, we aim that teachers can detect student failures quickly and can provide appropriate support to students who may not be able to earn credits. In this research, we try to predict the score of the test using machine learning. We use the data of the past 6 years × about 180 students in the LMS (Learning Management System) in programming class as the training data. By identifying relationships and patterns between the data, and by setting useful features in the target variables, we aim to attempt highly accurate prediction. Therefore, teachers can detect students with a low level of understanding early and can take appropriate measures for them.
Research Achievement Classification: 国内学会/Domestic Conference
Type: Conference Paper
Peer Review: なし/no
Solo/Joint Author(s): 共著/joint
Published journal or presented
academic conference: 
日本ソフトウェア科学会第37回大会
Date: 8-Sep-2020
Publisher: 日本ソフトウェア科学会
Appears in Collections:Ito, Kei

Files in This Item:

File Description SizeFormat
repit1-2.pdf533.75 kBAdobe PDFView/Open

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

 

Copyright © 2010-2012 FUTURE UNIVERSITY HAKODATE.
DSpace Software Copyright © 2002-2010  Duraspace