PREDIKSI DAYA SERAP PERUSAHAAN TERHADAP ALUMNI TEKNIK INFORMATIKA IBI DARMAJAYA BERBASIS JARINGAN SYARAF TIRUAN

Abstract

Informatics And Business Institute (IBI) Darmajaya was one of the institutions in the Lampung area that was involved in computer education and economics. One of the available routes in IBI Darmajaya in the computer field was Computer Science (CS). CS IBI Darmajaya produced the numbering graduate’s ±1000 students who were spread in the territory. Up to now the level of the success from the CS graduate in the matter fast or not this graduate was absorbed by the world of the work still could not be predicted. So as to be needed one sitem the prediction of the absorbency of the company against the alumnus CS was based on Neural Network. The neural network used the algorithm backpropagatioan and the function activation sigmoid binary that consist of tansig, logsig, and purelin. The network that was formed consisted of 25 of the number of layer cells of input, 52 of the number of layer cells were hidden first (hidden the layer), 26 of the number of layer cells hidden the two, and 5 of the number of layer cells of output. The training data as well as the testing were carried out with the spreading questioner against 93 alumni TI that was taken in a random manner, where questioner this consisted of 24 skills criteria that must be owned by the graduate CS that was received from the spreading of the poll of the 39 companies in the Lampung territory. This data will be put into the prediction system used Matlab v6.0.
The key word: Neural Network, the Backpropagation algorithm, the Aktivasi Function