Smartphone help contents re-organization considering user specification via conditional GAN. Issue 129 (September 2019)

• Record Type: Journal Article
• Authors: Lee, Younghoon; Cho, Sungzoon; Choi, Jinhae
• Is Part Of: International journal of human-computer studies. Issue 129(2019)
• Page Start: 108
• Page End: 115
• Publication Date: 2019-09
• Subjects: Help systems -- Contents re-organization -- User specification -- App usage sequence -- Conditional GAN -- Seq2seq; Human-machine systems -- Periodicals; Systems engineering -- Periodicals; Human engineering -- Periodicals; Human engineering; Human-machine systems; Systems engineering; Periodicals; Electronic journals; 004.019
• DOI: 10.1016/j.ijhcs.2019.04.002
• Languages: English
• ISSNs: 1071-5819

Abstract

Highlights: We proposed contents re-organization method in help systems of smartphone considering user specification. We applied a conditional GAN approach to generate and predict usage of help contents for each user as a new way of contents re-organization. Moreover, we consider the user specification that reflects users interests and characteristics with app usage sequence via seq2seq approaches. Our method outperforms previous approach in help contents usage prediction and re-organization by verifying with real business dataset from LG Electronics. Abstract: There are various help systems embedded in smartphones that are intended to provide assistance to users. These systems should be conveniently accessible in locations where users may need assistance. Moreover, when help contents are provided without regard to the users interest, it makes it difficult for the user to find relevant content. Thus, the present study provides a new method of re-organizing help content by considering each users interests and preferences using their app usage sequence. Based on the user specification derived from the app usage sequence, help contents usage prediction is generated for each of them with conditional generative adversarial network (GAN) architecture in a new way. Further, another method to pre-process data in applying conditional GAN, originally devised to generate image data, is proposed in our problem. The experiment result showed a higher absolute performance level of help contents usage prediction and better performance of effectiveness in re-organization of top- k contents compared to the existing benchmark method. Thus, the proposed method reflects the users interest and provides appropriate help contents for each user effectively.