Clustering Generating Mechanisms in a Mixture of Additive Noise Models

Pattern analysis is the study of patterns in observed data. One important question in pattern analysis is to find and group similar objects together, e.g., those people with cardiovascular diseases among a set of volunteers willing to donate an organ. This frequently-encountered problem in the domain of data analysis is referred to as classification and clustering with some differences between the two. Despite being common, the problem still carries many aspects that are not well established and need to be dug further. In the current project, we study one of the relatively less discussed features of the clustering problem. That is, instead of grouping the data points based only on the observable values of some features, we allow for the possible associations among the features of a single object as well. TO BE CONT’D

Faculty Supervisor:

Francois Soumis

Student:

Partner:

Huawei Technologies Canada Co Ltd (Montreal, QC)

Discipline:

Mathematics

Sector:

Technology; Information and Communications Technology; Other

University:

École Polytechnique de Montréal

Program: