Data Mining in the VISTA Clinical Trials Database

The project aims to apply data mining in the VISTA clinical trails database at the Geriatric Medicine Research Unit. The data contains information about Alzheimer’s disease (AD) patients, some of whom received cholinergic treatment and some of whom did not receive this treatment. Because of the variability of responses to treatment as well as the variability in the course of disease progression, it is difficult to identify the patterns robustly associated with improvement, which is crucial in justifying the choice of treatment.

Change Detection in Network Data Streams for Alcatel

Alcatel manufactures telecommunications equipment, ranging from telephone handsets to internet routers. The Research & Innovation group within Alcatel is mandated with evaluating new technologies and new ideas that may benefit Alcatel products, thereby benefiting Alcatel’s customers, and ultimately the end users. One current area of research is the ability to detect when the behavior of a stream of traffic has changed, and to then reallocate resources suitable for the changed stream.

LCD Colour Correction of LED Back Light Arrays

A High Dynamic Range display is a display that can reproduce a wider range of image intensities than typical displays - that is, a range more closely matching the capabilities of the human visual system. Specifically, BrightSide Technologies’ DR37-P display can produce an eye-squinting 4000 cd/m2 (as bright as a ceiling fluorescent light) or as dim as 0 cd/m2.

Intensity Adjustments for Authentic Image Display and Perception in High Dynamic Range Imaging

New High dynamic range (HDR) display devices are being developed which have contract ratios 300 times greater than existing LCD displays. These HDR displays are the first that are able to fully represent the range of intensities that are produced by illuminated objects on a sunny day, and are joining HDR cameras and file formats in creating a pipeline from image capture/construction to display that is able to maintain much higher image fidelity than was previously possible.

Applying Dynamic Constraint Scheduling Technology to the Efficient Management of Emergency Responders

In this project, the intern research team is proposing to research and develop a Decision Support System (DSS) based on current research in constraint-based optimization, dynamic scheduling and probabilistic modelling. The Ottawa Paramedic Services (OPS) can use this system in their command and control centre to improve its operation and as a result improve the lives of the residents of Ottawa. The system is supposed to handle the burden of the combinatorial complexity while the human operator makes high level, strategic and tactical decisions.

Streaming Workload Characterization and Multi-level Client Clustering

This project is part of a larger research project which investigates the applicability of the peer-to-peer (P2P) computing paradigm in designing large-scale content distribution systems. To develop an efficient content distribution system, it is essential to understand the workload that will be distributed, the behavior of content consumers and the environment in which the system will operate.

Performance Modelling of Professional Cyclists Using Self-organizing Maps

This project aims at performance modelling of athletes and involves the collection of detailed data that affects rider performance in professional cycling. This data is utilized for assessment of training and performance and for supporting individual training schedules through modelling and profiling of individual athletes. The methodology is based on pattern discovery and recognition using Self Organizing Maps, an exploratory data analysis model of demonstrated success in automated monitoring tasks involving multiple parameters.

Multi-class Problem Decomposition Using Genetic Programming

Behavioural detectors for intrusion detection require training in order to correctly characterize the operation of a service – protocol combination. Implicit in this is the assumption that the learning algorithm will scale to large datasets and provide simple solutions. This work will address both requirements under a Genetic Programming context through the use of a combined multi-objective, host-parasite model. It has already been demonstrated that both schemes are appropriate independently.

Modelling Game Strategies to Improve Performance

The objective of the project is to determine the impact of various game features on VLT game play. These features are based on the current strategies adopted by SPIELO for its North American market. This data will be collected from players in a real life setting, then analyzed, and used for player profiling and as an input for improving existing strategies and measure their effectiveness. The analyzed data, the derived player profiles and the inference from the study will be used in developing a software simulator of original gaming environment.

Mobile IP Infrastructure for Medical Data Transmission

Currently, EMS paramedics record patient information such as history, medical assessment, and treatments rendered onsite with pen and paper. They then convert this information to a paper-based call report and hand it over to the hospital along with the patient. This method is time-consuming and error prone. It causes delay for data analysis and lengthens paramedic turn-around time. This project will produce a data collection tool based on the mobile IP infrastructure which will reduce clerical errors, improve data analysis and medical care.