{"title":"Identifying Cognitive Attributes Using Deep Learning Classification Techniques","authors":"Shuai Zhao, Xiaoting Huang","doi":"10.1145/3397453.3397458","DOIUrl":null,"url":null,"abstract":"Cognitive diagnosis is very useful to teachers and students, but its application is limited at present. This is largely because identifying the cognitive attributes of items currently is labor intensive and time-consuming. In this study, we used text classification techniques to automatically identify cognitive attributes. Specifically, two popular deep learning classification models, long-short term memory and bi-directional long-short term memory, were employed in conjunction with word embeddings. As the baseline, support vector machine with feature selection using information gain was also adopted. Experiments based on a sample of 805 third grade math items showed that both the deep learning models performed better than support vector machine, and bi-directional long-short term memory achieved the best performance, yielding the accuracy of 82% and the F1 measure of 80%. Our result indicated that text classification methods, especially deep learning models, have great potential in identifying cognitive attributes efficiently, and in turn, make cognitive diagnostic more feasible to practitioners.","PeriodicalId":129569,"journal":{"name":"Proceedings of the International Workshop on Artificial Intelligence and Education","volume":"20 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2019-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the International Workshop on Artificial Intelligence and Education","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/3397453.3397458","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 2
Abstract
Cognitive diagnosis is very useful to teachers and students, but its application is limited at present. This is largely because identifying the cognitive attributes of items currently is labor intensive and time-consuming. In this study, we used text classification techniques to automatically identify cognitive attributes. Specifically, two popular deep learning classification models, long-short term memory and bi-directional long-short term memory, were employed in conjunction with word embeddings. As the baseline, support vector machine with feature selection using information gain was also adopted. Experiments based on a sample of 805 third grade math items showed that both the deep learning models performed better than support vector machine, and bi-directional long-short term memory achieved the best performance, yielding the accuracy of 82% and the F1 measure of 80%. Our result indicated that text classification methods, especially deep learning models, have great potential in identifying cognitive attributes efficiently, and in turn, make cognitive diagnostic more feasible to practitioners.