Sentiment analysis and usability model for Mysejahtera’s contact tracing application

Abstract

Infectious disease crisis management relies heavily on contact tracing applications to curb transmission. With advancements in technology, smartphones, and mobile applications have revolutionized the process, enabling rapid identification and notification of individuals exposed to infectious diseases. These contact tracing applications, such as Malaysia's MySejahtera, provide real-time data to health authorities, facilitating targeted interventions and reducing disease transmission. Understanding user sentiments toward these applications is critical to improving their effectiveness and acceptance. However, there remains a significant research gap regarding sentiment analysis, influencing factors, and usability models specific to Malaysia. This research investigates user sentiment polarity toward contact tracing applications, identifies factors driving positive and negative sentiments, and proposes a usability model tailored to these factors. The methodology involves sentiment analysis using supervised binary classification algorithms and thematic analysis of user reviews of MySejahtera. Key factors influencing sentiment are identified and integrated into a usability model, developed deductively from Hoehle and Venkatesh’s (2015) framework and relevant literature. The findings indicate that the K-Nearest Neighbour (K-NN) algorithm achieves the best performance with an F-measure of 83.99%. The majority of users express positive sentiment, attributed to factors such as User Interface Input, Application Utility, and Application Design. Negative sentiment, however, is also linked to Application Utility and Design. The proposed usability model comprises seven higher-order constructs and fourteen lower-order constructs, which were validated statistically to confirm its robustness. This research contributes significantly to the fields of sentiment analysis and usability by providing actionable insights for developers and public health authorities. The usability model offers a comprehensive framework for improving contact tracing applications, addressing user concerns, and optimizing usability to enhance public health outcomes.