Events & News
CS Colloquium
| Category | Lecture |
| Date | Friday, January 23, 2015 |
| Time | 11:00 am |
| Concludes | 12:15 pm |
| Location | Gould-Simpson 906 |
| Details | Please join us for cofee and light refreshments at 11am in Gould-Simpson 906. Faculty Host: Saumya Debray |
| Speaker | Marina Barsky |
| Title | PhD |
| Affiliation | Ontario Institute of Cancer Research |
Memory-Based Reasoning, Neighbours and Recommendations
Data mining lecture (3-rd, 4-th years)
Memory-based reasoning (MBR) is the process of solving new problems based on the solutions of similar past problems. MBR is not only a common classification method in machine learning but also underlies everyday human problem solving.
In this lecture we develop a method that uses the memory-based reasoning to classify previously unseen data records into known groups – the K-Nearest Neighbour classifier. We discuss different aspects of the classifier design, and explore its applications to real problems, the most common being an automated recommender system.
Biography
Marina Barsky got her Ph.D in Computer Science from the University of Victoria, Canada, and her M.Sc. in Biology from the Moscow State University. Her research has been on the intersection of database systems, data mining and algorithms, and her main contributions are scalable indexes for DNA sequence databases, reference-free bioinformatics and extraction of correlations from transactional databases.
She has taught multiple graduate and undergraduate courses, specializing in data science, algorithms, and software design.