Geo-spatial Data Mining Applications in Marketing

The project will develop methods for constructing profiles and predictive models for geo-spatial data. The socio-demographic profile will be developed for the geographical areas of interest. In addition, the models will be developed to identify areas with high potential; for acquiring new customers. The targeting of these areas may be conducted through unaddressed direct mail. Methods of testing effectiveness of marketing campaign and predictive models will be developed.

Classification of Acoustic Signals using Deep Learning Architectures

The classification of acoustic signals whose factors of variation are due to different atmospheric and sound propagation effect is a challenging problem. The internship will explore new learning algorithms for this application, which have the potential to capture some of the complex structure in the data.

A Generative Model of Impulsive Sound Production and Propagation

The overall objective is to design a learning system that takes a training set of acoustic signals and produces a classifier that can identify the category of the acoustic signal, out of a small number of categories. We have already found that it is important to consider the factors of variation that can influence the signal, and the proposed project aims at exploiting physical modeling knowledge to structure a generative model of the acoustic production of the observed signals.

Diabetes Tele-Monitoring for Older People via the Use of BlackBerries and Tablet PCs

In this project, 10 older patients who have uncontrolled diabetes will use wireless technology including a blood glucose monitor and a Blackberry or a Tablet PC to transmit their blood glucose readings from home to two nurses working at a private homecare company. These patients will transmit their blood glucose readings approximately three times a day during two 3]month phases, the total duration of the study period being 6 months. Throughout, patientsf blood glucose levels will be monitored by the two nurses.

Adapting Sigma° software for the mining of transportation information using aerial photos as part of updates to topographical maps

Sigma° is a software application for updating topographical maps developed by Synetix for the Quebec government department for natural resources and wildlife (MNRF). The algorithm can be used to update communication channels with remote-detection images and the principle of using topographical maps to guide detection procedures. The initial version of Sigma° was developed and validated for Radarsat-1 and SPOT Panchromatic satellite images with a resolution in the order of 5 m.

Designing a Multimodal Transportation Network

This project will develop a mathematical model and computer tools that will minimize transportation and infrastructure costs related to the forest road network, in order to ensure the transport of wood products to receiving plants. The model will consider several decision levels, including the location of transfer yards and the choice of mode of transportation. Constraints will include the storage capacity of terminals and transportation units, and the model will take into account travel distances as well as transportation and handling costs.

Information Extraction from Unstructured Data

Information extraction from unstructured data is a wide and relatively recent domain. For this research project, the focus will be on the information extraction from finance reports and news, more precisely related to the commodities market. This includes Natural Language Processing (NLP), expert systems (such as ontology-based systems) and information fusion as tools for analysing qualitative information in finance and producing investment decisions. NLP is a science studying the automated understanding of natural human languages.