Sistem Intelijen Penilaian dan Prediksi Kualitas Susu Pasteurisasi (Studi Kasus pada PT . Industri Susu Alam Murni Bandung)

Abstract

The aim of this research is developing an intelligent system for assessment and prediction of pasteurized milk quality. The system is implemented by integrated expert systems and Artificial Neural Network (ANN). This system is designed and developed by Matlab 6.5 and called SINKUAL-SP. The logical strategy model used is “Forward Chaining” and the tracing method used is “Best First Search” method. The method of Certainty Factor (CF) is used for handling uncertainty. Model Based Management system consist of five model : quality of fresh milk model, quality of process model, quality packaging and storage model, statistical process control model and prediction model by neural network. Multilayer system is used for the neural network architecture with one input layer, one hidden layer dan one output layer. The activation function is Sigmoid Bipolar activation function which gives the best performance network with learning rate 0.005 and momentum 0.9 together with RMSE, MSE and SSE as criterion error. The validation for neural network indicates the conformity between output of neural network and output of goal with value of RMSE 0.009959, MSE 0.00999 and SSE 0.009950. Decision for acceptance fresh milk and pasteurized of milk is implemented by acceptance grade which gives control quality process is easier than previously system. Based on the grade acceptance, the management of industry can take decision and action allow with a significant output of grade. In the research implementation stage, the data, gained in the company of “PT. ISAM” in April 2005, brings a significant output of research regarding quality of fresh milk such as the grade B, grade A for quality of process, and grade B for quality packaging and storage. This system gives a suggestion for user to improve the quality of pasteurized milk. Based on the outputs from the systems used, highly priority strategy quality system reconstruction, with 0.236 as weighting value. The quality improvement system upon the results of research is seemed to be a way of improving the quality of milk in process of pasteurization.