Sales Prediction : Analysis of Time Series Data Using K-Means Based Smooth Subspace Clustering

Abstract

The large database to mine information that is a Data mining process and convert it into a reasonable structure for further use. Launched and order to support the organization is decision making, business planning and Data mining techniques. Data analysis, increase profitability, innovation, efficiency in resource utilization is based on important management tool in data mining. Today companies gains competitive advantage from collecting past data and using for future forecasting. Past data and information based on future estimates. In this paper, the research subject is selected as the data of a consumer electronics store company. Two year Time series sales amount data of consumer electronics was used and grouped as four quarters in a year. Next year’s regression equations and naive bayes classifier methods and comprised by real sales amounts using the first quarter sales are forecasted. The real amounts and seasonal factors are really important to some product ranges that are near the sales forecasts results. In this context, various campaigns and marketing approaches have been proposed for the sales of company products by evaluating the forecast results.