Identifying Smartphone Based Features for Automatic Learning Style Identification Approaches Using Correlation Tests

Learning style identification is attracting considerable interest due to its potential to solve problems in human learning activities and has been widely utilized in learning management systems. Despite this interest, no one has studied learning style identification in an informal learning environment. Moreover, informal learning even plays the biggest role in human lifelong education. This study thus proposes smartphones as a new plausible observation environment, for their uses match the characteristics of informal learning. However, it is not yet known which smartphone parameters are reliable to be used as features in learning style identification models. Therefore, this study aims to find reliable features from smartphones by conducting the Spearman's rank correlation tests between several potential smartphone parameters and the Visual, Aural, Read/Write, Kinesthetic (VARK) learning style scores. The parameters were collected using a dedicated application and tested in two scenarios. We treated the applications individually without clustering in the first scenario, while we clustered the applications by their functions in the second one. The result revealed that there were eight potential features in total. This study has opened more possibilities to improve the development of the learning style identification system.