iOS application user rating prediction using usability evaluation and machine learning.

摘要

Mobile applications are earning popularity because of the significant benefits of smartphones, such as: portability, location awareness, electronic identity, and an integrated camera. Nevertheless, these devices have a number of disadvantages concerning usability, such as limited resources and small screen size.;A number of studies have investigated usability challenges in a mobile context and proposed definitions and measurement of the usability of mobile applications. Evaluating the usability of applications for mobile operating systems is a crucial step in addressing these difficulties and achieving success in mobile application markets, such as Apple's App Store. Usability evaluation must be tailored to the various mobile operating systems in use, each with its particular characteristics.;Apple App Store is the only source for buying or installing iOS applications. Users rate applications in the App Store and take into the account other users' evaluations when buying an application. Users tend to give higher ratings to the applications that satisfy their functional and non-functional requirements. Usability is one of the non-functional requirements that users consider when they rate applications. Hence, it is important to develop applications with higher usability and better user experience to be successful in the App Store.;Apple publishes a guideline named "HIG (Human Interface Guidelines)" and recommends following these guidelines during the design and development of iOS applications. There are also other guidelines that suggest other design principles and heuristics but the relationship between these guidelines and App Store success is unknown.;In addition, there is not any explicit method to predict the success of an application in the App Store. Developers and development companies spend much time to develop an application with little clue about the success of their application in the App Store.;This research project combines the guidelines from the literature and proposes an iOS application usability evaluation model to evaluate iOS application usability. It presents next an analysis of the relationship between criteria and application's App Store user rating by evaluating 99 applications. This research project also proposes a machine learning model to predict the success of an iOS application in the App Store, based on the evaluation method proposed in the first part of this research.