Personalized Instant Messaging UI Design for Elderly

Abstract

There is a paucity of studies on the consumption of digital content by elderly people utilising smart devices as well as strategies to get elderly people acquainted with smart gadgets. Usability and familiarity with smart devices for senior persons to utilise and get the most out of smart devices and digital content must be prioritised. A cognitive reaction-based intelligent UI suited for senior persons is proposed in this paper, which is based on the user's cognitive performance and demographics. A cognitive response feedback and demography dataset was built by interviewing a group of elderly in Malaysia. The context of the interview is associated with the unique cognitive keywords that may be anticipated by contextual semantic search. In this paper, two classifiers are used, Support Vector Machine (SVM), and Naïve Bayes (NB), and they are compared in terms of classification performances. The classifiers are validated using k-fold cross-validation (10-folds) using unigram TF-IDF and bigram TF-IDF, and the results were presented using accuracy, precision, recall, and F1 scores. Thus, user interface (UI) pre-sets lists will be matched to the user model based on the dataset classification result and be visualized using a mobile app interface builder.