The Research Of Apparel Sales Forecast Based On Non-parametric Kernel Estimation Method

The apparel market is the typical oversupplying buyer’s market, and the market competition is fierce. As a short life cycle products, apparel has the characteristic of long lead-times, short sales period, low salvage value and volatile markets. Apparel retailer will be difficult to accurately predict market demand and sales trend. The traditional order place meeting of garment is unable to quick response to fashion tread and consumer preference. Lots of garment companies lack a scientific and thorough plan before market prediction. They make qualitative judgments with personal experience and subjective idea, so the forecasts is low accuracy. The blindly orders and replenishment based on the inaccurate predicted value that cannot match the output with the future sales situation result in the shortage risk. However more garment companies will face the inventory backlog problem, which increase the operating costs, reduce the profit, and affect their long-term development.The research object of this paper is the casual garment company of a famous domestic fashion brand. The characteristics of short life cycle is more prominent in fashion. Combined with the garment company’s product features and sales data, the study intends to use mathematical method to solve the problem of sales forecast, and search reliable forecast scheme to improve the forecast precision of fashion apparel sales.In this thesis, the main research results are as follows: By using MATLAB as a programming tool, to divide the life cycle for the garment company’s apparel products, and design the non-parametric kernel estimation model to forecast sales. The nonparametric kernel estimation function is used to predict the daily sales and total sales of the company’s various products, and the accuracy of the forecast results is tested. On the basis of the precise results, offering the proposal for company’s order decision according to the forecast value. This study can help garment company to establish and improve prediction system, and also provide the reference for company to make rational arrangement of the production and replenishment plan.