Friday, September 24, 2021 - 11:00 to 12:30
SCHOOL OF COMPUTER SCIENCE – Colloquium Series
The School of Computer Science at the University of Windsor is pleased to present…
Dr. Javad Rahimipour Anaraki | Postdoctoral Fellow
Institute of Biomedical Engineering
University of Toronto
Rosebrugh Building, 164 College Street, Room 407
Toronto, Ontario M5S 3G9 Canada
Meeting URL: https://us06web.zoom.us/j/87823071058?from=addon
Passcode: If interested in attending this event, contact the Graduate Secretary at email@example.com with suffient notice before the event to obtain the passcode.
Consider a supervised dataset D = [A|b], where b is the outcome column, rows of D correspond to observations, and columns of A are the features of the dataset. A central problem in machine learning and pattern recognition is to select the most important features from D to be able to predict the outcome. In this paper, we provide a new feature selection method where we use perturbation theory to detect correlations between features. We solve AX = b using the method of least squares and singular value decomposition of A. In practical applications, such as in bioinformatics, the number of rows of A (observations) are much less than the number of columns of A (features). So we are dealing with singular matrices with big condition numbers. Although it is known that the solutions of least square problems in singular case are very sensitive to perturbations in A, our novel approach in this paper is to prove that the correlations between features can be detected by applying perturbations to A. The effectiveness of our method is verified by performing a series of comparisons with conventional and novel feature selection methods in the literature. It is demonstrated that in most situations, our method chooses considerably less number of features while attaining or exceeding the accuracy of the other methods
Link to Article: A Feature Selection based on perturbation theory
Javad received his Ph.D. from the Memorial University of Newfoundland in Computer Science, investigating feature selection methods based on fuzzy-rough and classic sets theories. Later, he joined Prism Lab at the University of Toronto as a postdoctoral fellow to work on machine learning pipelines targeting brain signals to provide kids with severe movement disabilities with the tools to communicate and play games with friends and family. Currently, he is a senior researcher at Huawei while continuing to support Prism Lab.
5113 Lambton Tower 401 Sunset Ave. Windsor ON, N9B 3P4 (519) 253-3000 Ext. 3716 firstname.lastname@example.org (working remotely)