Features Selection in the Proposed Draft Sheet C1 for General Elections in Indonesia

Abstract

Elections in Indonesia are held every five (5) years. In 2019, the elections were held simultaneously so that general elections could be carried out efficiently because it reduces time wastage. In the implementation of the General Election, using sheet C1 as a sheet to fill in the calculation results, the obstacles that often occur in the implementation of the election are often technical problems such as filling out the C1 sheet, which is still manual, and error input values. This study aims to create a new approach using the features in the C1 draft proposal for general elections to reduce mistakes and manual voting considerations. This study uses an artificial neural network method as a feature of the numerical value prediction process from the proposed new C1 sheet. The neural network method used is the Backpropagation technique, where the machine will recognize each number so that the calculation process will be faster. With the election feature, the C1 design for this gets the results that this feature can detect the writing of numbers accurately
with an accuracy rate of 98%.