An Efficient Application-Device Matching Method for the Mobile Software Ecosystem

Abstract

In the mobile software ecosystem, a method that finds out the applications that are compatible with the device of an end user is called application-device matching. In the current mobile software environment, the device fragmentation causes substantial degree of inaccuracy in matching applications with devices as the traditional platform-centric method handles only the features of platform vendors without considering the unique feature set of a certain device, such as device-manufacturer's features, resulting in a low accuracy in matching applications and devices. This paper proposes a new matching method that is device-centric, which achieves high accuracy in application-device matching by grouping features of existing devices and then using it as criteria of application-device matching. To demonstrate the performance of our method, we conduct a case study with 22 devices and 10 applications in the Google Android mobile software ecosystem. The result of case study shows our proposed method shows a higher accuracy.