| Start Date: | 6/2/2014 | Start Time: | 10:00 AM |
| End Date: | 6/2/2014 | End Time: | 11:30 AM |
|
Event Description
Title: Distance/Similarity Measures in Bioinformatics
Advisors: Drs. Moshe Kam and Leonid Hrebien
Date: Monday, June 2, 2014
Time: 10:00 a.m.
Location: ECE Conference Room 302, 3rd Floor, Bossone Research Enterprise Center
Abstract
Many Bioinformatics applications rely on the computation of similarities between objects. Distance/similarity measures applied to vectors of characteristics are essential to problems such as classification, clustering and information retrieval.
The proposed work explores the usefulness of distance/similarity measures in several bioinformatics applications. These applications are of two categories:
(1) Estimation of the adverse reaction severity of unknown treatments, based on the severity of known treatments, in order to provide guidance for use of the unknown treatments in future clinical trials.
(2) Estimation of the similarity of several time-series data sets resulting from biological measurements and deciding on the best distance measures for similarity assessment.
To address the first category, we studied several clustering and classification approaches, and developed a ranking with distance metric learning approach. The methods were tested on a Cytokine Release Syndrome (CRS) data set, a Cardiotocography (CTG) data set and two Qualitative Structure Activity Relationship (QSAR) data sets. Estimation of the similarity of time-series data also used distance measures and classification algorithms, but employed data fusion methodology as well. The proposed methods were tested on data from Continuous Glucose Monitoring (CGM) system. In the future, more data will be used to evaluate the proposed methods. |
|
Location:
ECE Conference Room 302, 3rd Floor, Bossone Research Enterprise Center |
Audience:
Current StudentsFacultyStaffGraduate Students |
|