Autonomous Fast Classifiers for Pharmaceutical Data Sets
Talk Information
Muncie, IN, 5/24/04
Tags
High Throughput Screening, Logistic Regression
Abstract
This was an invited talk about my LR work. LR is relevant to some pharmaceutical problems, such as high throughput screening (HTS). This talk also highlights some of Ting Liu's fast k-nearest neighbor work. This talk was presented at the Midwest Biopharmaceutical Statistics Workshop (MBSW) 2004, and was later presented at Applied Biosystems, Inc.
Attachment Data
Download (application/pdf, 1.4 MB)
Links
- Paper: Making Logistic Regression A Core Data Mining Tool With TR-IRLS
- Paper: Logistic Regression for Data Mining and High-Dimensional Classification
- Paper: Making Logistic Regression A Core Data Mining Tool: A Practical Investigation of Accuracy, Speed, and Simplicity
- Efficient Exact k-NN and Nonparametric Classification in High Dimensions
- Software: LR using TR-IRLS, source and binaries available under the GNU GPL
- Software: Sparse Logistic Regression
- Software: Dense Logistic Regression
- Software: Sparse K Nearest Neighbor
- Software: Dense K Nearest Neighbor
- Paul Komarek's logistic regression webpage