PERAMALAN PENJUALAN PRODUK PAKAIAN DENGAN KOMBINASI METODE FUZZY TIME SERIES DAN ALGORITMA PARTICLE SWARM OPTIMIZATION
Daftar Isi:
- One of the sales forecasting is to prevent over production and under production which causes the company to lose the opportunity to sell its production and can help the company to conduct sales strategies in the future. In this study, forecasting is done using the Fuzzy Time Series method because it can project future data based on data in a timely manner. To improve forecasting accuracy, optimization using the Particle Swarm Optimization (PSO) algorithm is used to change the interval pattern of linguistic values in the Fuzzy Sets process. The advantages of the PSO algorithm have high decentralization with a simple implementation so that it can solve optimization problems efficiently. Error rates are calculated using MSE and AFER. Based on testing predicting clothing products with Fuzzy Time Series using PSO Optimization obtained a decrease in the results of the average error rate (error) of MSE = 7451450.145 and AFER = 0.234% for Kaos data types, MSE = 2115209.254 and AFER = 0.090% for Shirt data types MSE = 1398279.2 and AFER = 0.087% for Jacket data types with 24 test data. K