Boosting Ensembles of Heavy Two-Layer Perceptrons for Increasing Classification Accuracy in Recognizing Shifted-Turned-Scaled Flat Images with Binary Features
Journal Title: Journal of Information and Organizational Sciences - Year 2015, Vol 39, Issue 1
Abstract
A method of constructing boosting ensembles of heavy two-layer perceptrons is stated. The benchmark classification problem is recognition of shifted-turned-scaled flat images of a medium format with binary features. The boosting gain is suggested in two aspects. The earliest one is the ratio of minimal recognition error percentage among the ensemble perceptrons to the recognition error percentage performed by the ensemble. The second gain type is the ratio of minimal variance of perceptrons’ recognition error percentages over 26 classes to variance of the ensemble’s recognition error percentages over 26 classes. Both ratios increase as the number of perceptron classifiers in the ensemble increase. The ensemble of 36 classifiers performs with increased accuracy, where recognition error percentage is decreased for 33 %, and the variance is decreased for more than 50 %. Further increment of classifiers into ensemble cannot increase accuracy much as there is the saturation effect of the boosting gain. And the gain itself depends on the range of noise modeling object’s distortions. Thus, the heavier perceptron classifier the less gain is expected.
Authors and Affiliations
Vadim Romanuke
Keywords
Related Articles
- EP ID EP486349
- DOI -
- Views 83
- Downloads 0
Ontology in Information Security
The past several years we have witnessed that information has become the most precious asset, while protection and security of information is becoming an ever greater challenge due to the large amount of knowledge necess...
Business Model Grounds and Links: Towards Enterprise Architecture Perspective
The business model concept emerged in theory and practice without a consensus on the understanding of the concept, but it has become a well accepted and useful construct in fields such as strategy, organization, informat...
Discretization for Naive Bayes Taking the Specifics of Heart Data into Account
At the present time heart disease is a major cause of death. Factors such as physical inactiveness, obesity, diabetes, social isolation and aging are expected to make the situation worse. It is worsened even further with...
Analysis of Methods and Techniques for Prediction of Natural Gas Consumption: A Literature Review
Due to its many advantages, demand for natural gas has increased considerably and many models for predicting natural gas consumption are developed. The aim of this paper is to present an overview and systematic analysis...
Do You Walk the Talk in Quality Culture?
We present an action research project to foster quality culture in business processes. The client setting is in the food industry, a vital sector for our society and one of the most regulated in the world. Food productio...