The data generated from real-world applications are now exhibiting multifaceted nature with multiple contexts. For example, in addition to the location and time of a smartphone user, the in-built sensors can also gather heartbeats, activities, etc., of the user. Similarly, any web-based applications can record user feedbacks in the form of text and ratings. Extracting meaningful insights from this type of data is constrained due to the inability of existing machine learning techniques to adapt to the multifaceted multi-context data. Moreover, handling these complex data requires computationally powerful techniques.
While factorisation techniques have been found suitable for multiple tasks like text clustering, prediction, pattern mining, etc., it suffers from efficiency and quality issues when dealing with complex multifaceted multi-context data. The objective of this project is to develop a more sophisticated factorisation technique that can handle complex data efficiently and capable to adapt for multifaceted multi-context data, providing solutions for many real-world applications.
The proposed project enables the process of insights discoveries and predictions for various domains like healthcare, transportation, web 3.0, smart city, etc., that generates multifaceted multi-context data.