PERBANDINGAN METODE ROUGH SET DAN NEURAL NETWORK UNTUK PREDIKSI STOK OBAT DI APOTEK

Abstract

Stock collection of drug is a problems that always pharmacist’s deal because of the difficulty of counting a large amounts of data and lack of checks on the existing data. The amount of data often lead to errors, mistakes and difficulties. The impact of chaos and loss accounts. Based on stock drug problems made predictions using the comparison method of Rough Set and Neural Network. Variables consist of drug names, drug type, drug dose, the drug unit, drug packaging, expiration, the stock of drugs, sold drugs, remaining drug and medication orders. RS make predictions in the form of decision systems, write a class equivalent to a discernibility matrix modulo D to form reduction. Reduct establish rules and knowledge. While the NN using backpropagation to determine the structure of the input nodes, hidden and output. Perform normalization, learning rate, error tolerance and maximum iteration values. Tests using Rosetta V1.4.41 RS and NN using Matlab 6.1. The final result gives the output of the RS prediction rule and knowledge. While the NN in the form of numbers and graphs. The final result of the comparison provides one of the best methods that can help the pharmacist in making decisions.

Keywords: Backpropagation, Neural Network, Prediction, Rough Set, Stock Drugs