Performance Evaluation Of Unsupervised Learning Algorithm In Biometric Based Fraud Prevention System
Journal Title: International Journal of engineering Research and Applications - Year 2018, Vol 8, Issue 10
Abstract
Recently biometrics is the best alternative for the token based and knowledge based security systems. Unlike commonly used traditional identification technology based on passwords and keys, biometrics is more reliable, more convenient and more secure. Several algorithms have been employed, especially supervised learning algorithms, as data classifications. This paper implement unsupervised learning algorithm in multi-modal biometric system for its suitability. The system architecture consists of morphological pre-processing, feature selections, feature level fusion by concatenation, and matching stages. The performance of the Self Organizing Feature Map is compared with back-propagation neural network. The processed data were matched for recognition using self organizing feature map and back-propagation neural network algorithms for performance. The back-propagation neural network produced recognition accuracy rate of 93.7, genuine acceptance rate of 98.4, and false acceptance rate of 7.7 while self organizing feature map yielded recognition accuracy rate of 93.5, genuine acceptance rate of 93.7, and false acceptance rate of 7.8. And it was deduced from the results that self organizing feature map relatively well as back-propagation neural network.
Authors and Affiliations
Ismaila W. Oladimeji, Shittu Jaleel K. , Ismaila Folasade M, Ajayi Ayomide O.
Keywords
Related Articles
- EP ID EP401969
- DOI 10.9790/9622-0810016267.
- Views 152
- Downloads 0
A Comparative Study Onpreprocessing,Segmentation And Classification Techniques For Cancer Detection
Cancer is one of the diseases which is the reason of death all over the world. Among them lung cancer is most common type of cancer. Early detection of cancer is very important task for the survival of human being. Image...
Health Gamification Research – a bibliometric analysis
In the last five years, the researchers embarked on original research in Gamification for health, however the Gamification professionals and health researchers still do not know where the references of the area are. Ther...
Li-Fi:Ultra Fast light based Wireless Data Transmission
In the present time, Wi-Fi is the most drifting area. As web customers are increasing consistently, there is an enormous load on radio range that leads to congestion. To show signs of improvement in data transfer, capaci...
Blockchain: An Inside Review
Blockchain is a database which can be directly shared between non-trusting parties, without a mediator. Blockchains are more secure than regular databases in some ways, and less secure in others. In this article we will...
Field Effects on Multiple Particle Contamination And Breakdown Voltage Inside Gas Insulated Bus Duct.
The metallic particle contamination in the Gas-Insulated Systems (GIS), plays a key role for the occurrence of Partial Discharge (PD), which causes the failure of insulation in the GIS. This paper presents the study of f...