Enhancing Electoral Surveys with Artificial Neural Networks

Abstract

The objective of this study is to search for the main factors that can influence to predict the results of voting surveys. A system is developed that allows the optimization of Artificial Neural Networks to identify the factors that affect the electoral result, through a computational method that allows the evaluation of the characteristics that influence a successful electoral vote. An Artificial Neural Network with three layers and a back propagation learning algorithm is used. The first phase loads the system by developing a random synthetic database. This will contain the data that will serve as input to the Artificial Neural Network to optimize the most outstanding attributes that affect a vote. The system identifies the inputs to the Artificial Neural Network, and the iterations that can be carried out to optimize its outputs.

Keywords

• Artificial Neural Network
• Conservative
• Algorithm