Sales Forecast Using a Hybrid Learning Method Based on Stable Seasonal Pattern and Support Vector Regression

摘要

An obvious seasonality appears in customer demand of many industries. It can have a repetition period from a month to a year. In this paper, researchers use a hybrid learning method to improve sales forecast and supply chain management. This hybrid method combines Stable Seasonal Pattern (SSP) and Support Vector Regression (SVR) analysis. It provides a flexible approach which gives accurate forecast for budget and manufacture planning of companies.