WHADDAPP!?: A social app discovery and recommender system for android

• Main Authors: Agustin, Natalie S., Ang, Sofia Indira Odessa G., Lee, Joan Denise M., Ng, Anthony Ersen S.
• Format: text
• Language: English
• Published: Animo Repository 2012
• Subjects: Application software > Development; Mobile apps; Application program interfaces (Computer software); Computer Sciences
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/11037
• Institution: De La Salle University

In this age where people are always on the go, technology has also gone mobile. With that, developers started designing mobile applications for mobile devices. Now with over 480,000 mobile applications available in the Android platform, there arises the need for effective recommender systems to minimize the information overload experienced by the user and to make the process of app discovery more convenient. The purpose of this research is to combine the important features of existing recommender systems and to improve app recommendations by exploiting the concept of homophily in one social graph which considers the familiarity of friends and the similar interests of friends. WhaddApp?! system was developed to help reduce the information overload experienced by the users when finding interesting applications. To evaluate the created system, both qualitative and quantitative methods were used. The quantitative methods showed that the system has been effective in increasing the visibility of Android applications. However, actual installation of the recommended apps yielded low scores because whether an app will be installed by the user is affected by factors that can not be controlled by the system. The qualitative method shows that the users are satisfied with the system and would use it if ever it was published. There is a need, though, to improve the loading time and the user interface. All in all, the system has achieved its main objective of minimizing the information overload experienced by the user.