Hybrid Feature Selection with Genetic Algorithms and other Methods
Abstract
Feature Selection (FS) processes are becoming more important for image-base cell profiling as the field begins to rely more heavily on computational means of analysis. A single image analysis can be scanned for more than a thousand different features each with varying metrics. Current researchers have realized the holistic approach to measure every possible feature lead to the issue of determining the features that provide an adequate experimental analysis. Past FS methods were to preselect features from a smaller feature set, however, increases in computational speed and advances in machine learning methods have allowed quick expansive analysis to become more prevalent.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 16, 2020
- Accession Number
- AD1123160
Entities
People
- Daniel W. Cowan
- Heather A. Pangburn
- Patrick M. Mclendon
- Timothy Ho
Organizations
- Rensselaer Polytechnic Institute