Sales Forecasting based on SVR with Keywords Filtering of Web Search Queries
Recently, development cycle of high tech new products becomes shorter and sales forecasting of such new products
is important but difficult. According to a survey from CNNIC, 77% of Internet users search on the Internet for information of
products before purchasing them and historical big data of search record is useful for new product sales forecasting. To
improve accuracy of forecasting new products, forecasting method by using historical sales data and multi variables time series
data of Web search queries data might be promising approach.
Keywords - Sales forecasting, New product, Web search trend, Google trend, SVR