On-Device Application Categorization using Application Resources
S. Sangeetha, Ramesh Inturi, Ayush Srikanth, T Hariprasath, P. Abishavarthana, S. Samyuktha · 2023
Recent advancements in mobile devices resulted in a phenomenal growth in the usage of smartphones. Applications for these smartphones can be downloaded from the play store. The Android Play Store presently has over 2.5 million applications, a figure that has been growing significantly over the previous several years. As a result, the number of applications downloaded by customers has increased significantly. With such a surge in application (app) production and usage, it becomes necessary to classify or organize applications according to the services they provide. Classification helps to facilitate navigation and makes it convenient to search for the needed application. Hence, research has been conducted on the navigability, grouping, and search capabilities of these devices’ apps. In particular the current technique to app classification is manual labeling or data extraction from the app store. These strategies are inefficient in terms of time and resources, and they are also inefficient in terms of scalability. In this paper a novel architecture is proposed for classifying mobile applications. Our work solves the classification problem using information contained in the application packages (APKs), eliminating the need for external dependencies and allowing processing to take place entirely on-device. Our proposed system utilizes machine learning and deep learning techniques for text classification using the string.xml files. We have also automated the generation of datasets directly from the APKs. Experimental results confirm that the proposed solution shows improved results than the existing systems.