Statistic Metrics for Evaluation of Binary Classifiers without Ground-Truth - Archive ouverte HAL Access content directly
Conference Papers Year : 2017

Statistic Metrics for Evaluation of Binary Classifiers without Ground-Truth

(1) , (2)
1
2

Abstract

In this paper, are presented a number of statistically grounded performance evaluation metrics capable of evaluating binary classifiers in absence of annotated Ground Truth. These metrics are generic and can be applied to any type of classifier but are experimentally validated on binarization algorithms. The statistically grounded metrics were applied and compared with metrics based on annotated data. This approach has statistically significant better than random results in classifiers selection, and our evaluation metrics requiring no Ground Truth have high correlation with traditional metrics. The experiments were conducted on the images from the DIBCO binarization contests between 2009 and 2013.
Fichier principal
Vignette du fichier
UKRCON-2017_Doc_or_Tex_Paper_288.pdf (556.92 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-01563615 , version 1 (17-07-2017)

Identifiers

  • HAL Id : hal-01563615 , version 1

Cite

Maksym Fedorchuk, Bart Lamiroy. Statistic Metrics for Evaluation of Binary Classifiers without Ground-Truth. IEEE First Ukraine Conference on Electrical and Computer Engineering (UKRCON) , May 2017, Kiev, Ukraine. ⟨hal-01563615⟩
329 View
677 Download

Share

Gmail Facebook Twitter LinkedIn More