A Mobile Application Classification Method with Enhanced Topic Attention Mechanism.

Faced with the explosive growth of mobile applications, how to classify mobile applications correctly and efficiently is more helpful for users to choose their own mobile applications, which has become a challenging issue. For this reason, we propose a mobile application classification method with enhanced topic attention mechanism. Firstly, our approach uses LSA to obtain the global topic of mobile application description text. Then, the local hidden representations of mobile application are trained by BiLSTM model. Secondly, for mobile application content representation text rich in global topic information and local semantic information, attention mechanism is introduced to distinguish the contribution degree of different words and calculate their weight values. Thirdly, the classification and prediction of mobile application can be completed by using the softmax activation function through a full connection layer. Finally, we evaluate our method on a real and open dataset Mobile App Store. On the whole, the experimental results illustrate that the performance of our approach is better than other comparison methods, and the classification accuracy of mobile applications is indeed improved. Particularly, compared with the standard LSTM model, the method proposed in this paper increased more than 12.7% in F1 score.