Application of Artificial Neural Networks in Estimating Participation in Elections

TL;DR

This paper demonstrates that a two-layer feedforward neural network can predict election participation rates with 91% accuracy, outperforming traditional methods in analyzing electoral flows.

Contribution

It introduces a neural network model specifically designed for election participation prediction, achieving high accuracy in a real-world regional case study.

Findings

Neural network achieved 91% prediction accuracy.

Assessment methods confirmed the model's effectiveness.

Application to Iranian election data proved practical.

Abstract

It is approved that artificial neural networks can be considerable effective in anticipating and analyzing flows in which traditional methods and statics are not able to solve. in this article, by using two-layer feedforward network with tan-sigmoid transmission function in input and output layers, we can anticipate participation rate of public in kohgiloye and boyerahmad province in future presidential election of islamic republic of iran with 91% accuracy. the assessment standards of participation such as confusion matrix and roc diagrams have been approved our claims.