Communication data based user activity recommendations

The amount of information exchanged in the communication increases rapidly as the user plays a more role in her organization and social networks. It has become overwhelming to many people in this communication and information centric era and the need for assistance to manage communicated information and the communication relations has grown. We present a communication data based user activity recommender system that aims to help the user use the communication services more effectively and easily. The communication data of a user provides abundant information about the topics the user is working on, the people the user communicates with, and the communication and information needs of the user. Our system analyzes the communication data and extracts such information, generates recommendations for user's communication services, and provides the information the user needs according to the user's communication context. Machine learning and natural language processing methods are utilized for communication data analysis and the performance of an example of our recommenders -- predictive meeting assistant -- is presented in details. Our recommender system has been implemented for large-scale deployments and its core architecture is presented.