Deep Evolving Dual Attention Network for Event Prediction to evolving participants in Event based Social Networks

As flexible venues for interested individuals to exchange knowledge on a wide range of social and technical subjects, event-based social networks have grown in popularity. various event planners build several events on these platforms because of the various advantages of event-based social networks. Because of the proliferation of identical events, users find it difficult to forecast which of the competing possibilities will triumph. Instead of focusing on the user’s shifting factors, deep learning models have focused on various components of the event. However dual attention on event and user has not be focused in predicting the successive events to the selective group of the participants. This work uses a solitary value to execute the imputation approach and irrelevant attribute reduction. It is suggested to use an online deep developing dual attention network for subsequent event participants. First, the evolving decomposition method’s prediction was used to fill in the missing values in the dataset. The evolving user and event features are then extracted using linear discriminant analysis. In order to calculate the subsequent event for the user upon processing the user attention layer and event attention layers, those features are projected onto the online deep dual attention network. Layer concatenation offers a representation learning method for user mapping to the event to the drift in the feature vectors of the user and the event. At last, it is possible to recommend the event to evolving users and participants. The model’s performance is calculated using precision with 99.2%, Recall with 99% and f measure with 99%.