Expansive growth of online information and services has caused the problem of information overload. Recommender systems are a class of intelligent tools that help individuals to deal with the problem via filtering irrelevant information and providing personalized recommendations. A well-known application of recommender systems is to recommend products to users (e.g. Amazon, Netflix, etc.), usually by tracking the behavior Despite several efforts in academic and industrial community, recommender systems still face several key challenges such as improving accuracy and serendipity of Top-k recommendation, detection shilling attacks, preference elicitation, etc.
In Kdd lab, we are working on addressing some of these issues by suggesting novel approaches for modeling and analyzing users’ preferences and behaviors.
There is an interdisciplinary area of research that applies computer science and mathematical techniques to uncover patterns and build new models that are useful for diagnosing and treating diseases. In KDD lab we mine many kinds of data, including healthcare systems’ information, genomic and proteomic data and biomedical images and signals to do that.
Stock market analysis and prediction along with portfolio selection has attracted a lot of interest due to never-ending existence of new investors looking for an appropriate investment basket. However, stock market prediction is not a straightforward task as it is affected by many factors including the world stock market indices, commodity trading prices, currency prices, etc.
In Knowledge Discovery and Data Mining lab, we seek to improve stock market prediction algorithms through modeling the complicated relationships among stock markets across the world and other influential factors. For this purpose, we are working on designing and exploiting variety of supervised, semi-supervised and unsupervised data mining algorithms such as neural networks, frequent pattern discovery, graph mining, etc.