Effect of Difference in Hidden Layer Amount in Artificial Neural Networks Against Prediction of Needs of Captopril and Paracetamol in Hospitals

Main Article Content

Priyo Wibowo

Abstract

Research to predict drug needs at the hospital in Semarang. By using the data needs drug for 6 years from January 2007 to December 2012 made the prediction that the drug needs to avoid shortages or overstocks of drugs. The system is required to obtain the necessary amount of drug in a given month in order to avoid shortages of drugs and overstocks of drugs. This study uses matlab software to design the look and design a Neural Network architecture and using Microsoft Office Excel software to provide input data. Spreadsheet data model helps the program as input to the training data. Artificial Neural Networks provide interpretations of the drug needs. any incoming data given initial weight, is processed in the hidden layer using the specified parameters are like max epoch, learning rate, activation, duration, and the pace of learning. Parameters of success using the Mean Square Error and the value of the correlation coefficient is a point of reference. The results obtained using Neural Network architecture with a number of hidden layer neurons 50 and 75 managed to get the value of an accuracy of 74.89% for the type drug Captopril, then architecture using the number of hidden layer neurons to 20 and 15 types of drug Paracetamol resulted in 86.21%. The results obtained are very varied topography which influenced the shape or characteristics of the data and the amount of data being tested as a data input. More and more data is being tested affect the accuracy percentage.

Article Metrics

PDF Downloads:

Download data is not yet available.

Article Details

Abstract Reads:

13 times on this page.