Graduate Student Seminar
Wednesday, April 30, 2014 · 11 AM - 12 PM
| First Speaker | Iris Gauran |
| Second Speaker | |
| Session Chair | Nilabja Guha |
| Discussant | Dr. Malinovsky |
| Place | BS 120 |
- Title
- Classification of Congenital Hypothyroidism in Newborn Screening using Self-Organizing Maps
- Abstract
- Detecting the Congenital Hypothyroidism among neonates is a major concern of medical practitioners because it provides richer information than the healthy cases. However, because of the rarity of this metabolic condition, existing classification algorithms oftentimes misclassify a neonate as “normal.” This paper aims to investigate the efficiency of Self-Organizing Kohonen Maps (SOM)—a type of artificial neural network—as a tool for classification, particularly in detecting outliers. The proposed methodology was able to address the problem of finding a statistical threshold for TSH1 as a reasonable and sound cut-off value. The bootstrap estimate of the standard errors is roughly 5%.
