An Intelligent System based on Back-propagation Neural Network and Particle Swarm optimization Based Neural Network for Diagnosing Anemia in Pregnant Ladies

Volume 6 Issue 2 March - May 2017

Research Paper

An Intelligent System based on Back-propagation Neural Network and Particle Swarm optimization Based Neural Network for Diagnosing Anemia in Pregnant Ladies

Neha Sharma*, Vikas Khullar**, Ashish Luhach***

* PG Scholar, Department of Computer Science & Engineering, CT Institute of Engineering Management and Technology, Jalandhar, India.

** Assistant Professor, Department of Computer Science & Engineering, CT Institute of Engineering Management and Technology, Jalandhar, India.

*** Vice Principal & Associate Professor, Department of Computer Science, and Engineering, CT Institute of Engineering Management and Technology, Jalandhar, India.

Sharma, N., Khullar, V., and Luhach, A. (2017). An Intelligent System based on Back-propagation Neural Network and Particle Swarm optimization Based Neural Network for Diagnosing Anemia in Pregnant Ladies. i-manager’s Journal on Information Technology, 6(2), 27-32. https://doi.org/10.26634/jit.6.2.13574

Abstract

According to WHO According to World Health Organization (WHO) survey, anemia is one of the most commonly encountered medical deficiencies during pregnancy [5]. BP network has been successfully used for anemia diagnosing in pregnant ladies, however BP network's drawbacks, such as long execution time and its easy fall into local optima have restricted its wider applications. Recently proposed stochastic optimization method Particle Swarm Optimization (PSO) is also been discussed. Also the way BP network's initial weights and bias are optimized; Particle swarm optimization is also carefully discussed. In this paper, firstly BP is used to initially train and test the BP network, then the Particle Swarm Optimization Based Back-Propagation (PSO-BP) networks is used to train and diagnose the anemia in pregnant ladies. While concluding the experimental results, it shows variation in the taken parameters, execution time, and accuracy [10, 11], [16].