Intelligent User Profiling for Recommender Systems in the Era of Big Data

Pervasive devices and Web 2.0 applications are popularly used and attracts billions of users. With these devices such as smart phones and smart watches and applications such as Facebook and Twitter, human users have created massive amounts of data such as exercises, sleep, geolocation, social media posts, social connections, and other information. To be able to cope with complex, huge, and dynamic data, people need the assistance of intelligent information systems such as recommender systems for finding, sorting, and filtering the available information. Recommender Systems is one effective personalisation tool. They can suggest content and services that tailored to individuals based on knowledge about their preferences and behaviours. User profiling is the foundation of recommender systems, which is to discover knowledge about users such as interests and information needs.

The general aim of this project is to develop scalable and effective user profiling approaches to discover new knowledge about users’ individual interests, preferences, sentiment and emotional status, and information needs from large scale social and pervasive data, for the purpose of facilitating recommender systems to provide better services for users. To better capture the unique distinctive features of social and pervasive data, this project will adopt advanced machine learning techniques such as reinforcement learning, deep learning, and semi-supervised machine learning methods to develop novel user feature learning models. Its outcomes will alleviate the major difficulty--the lacking of accurate user profiles--of many personalisation in many domains such as e-commerce, e-health, and e-learning. This project will contribute to the discovery of new knowledge about human users and the social world, and the new solutions to make better usage and processing of big data.