PREDIKSI HASIL PEMILU LEGISLATIF DKI JAKARTA DENGAN METODE NEURAL NETWORK BERBASIS PARTICLE SWARM OPTIMIZATION

General elections are a means of implementation of the sovereignty of the people in the UnitaryState of Indonesia based on Pancasila and 1945 Constitution. Elections held in Indonesia is tochoose the leadership of both the president and vice president, member of parliament, parliament,and the DPD. The releated research of general election usually using decision tree algorithm orneural network algorithm. Each of methode has strong and weakness, but neural neutwokalgorithm can solve problem in decision tree algorithm. The accuracy using neural networkalgorithm in predicting the election has less accurate. In this study created a model neuralnetwork algorithm and neural network algorithm model based genetic algorithm to get the rule inpredicting the outcome of legislative elections and provide a more accurate value of accuracy.After testing the two models namely neural network algorithm and neural network algorithmbased on particle swarm optimization, the results obtained are the neural network algorithmproduces an accuracy value by 98,50% and the AUC value of 0.982, but after the addition ofneural network algorithm based on particle swarm optimization value of 98,85 % accuracy andAUC value of 0.996. So both methods have accuracy rate of 0.35 % difference and the differencein the AUC of 0.14.