Android is used by most of the population of the users. It is an attractive target for malicious application developers due to its open source nature. These malicious writers are developing new trends to steal sensitive information from the devices. A new trend is represented as collision attack in this manner. During this attack different apps communicate via Inter-Process Communication (IPC) for variety of purposes. In this paper, a dynamic approach is proposed for automatic collision detection between communication applications. The focus of the study is on the sharing of multiple type data. Moreover, to select application for analyzing is difficult task to perform and two predictive functions has been used in this manner. The evaluation was performed on a dataset of 800 android applications for analyzing the colluding couples. The developed methodology produces an accuracy of 97.2% during the experiments by the developed system.