Research On Fashion Product Sales Forecasting Method Considering Demand Censored

In recent years,the people's living consumption level has gradually improved and consumption has gradually upgraded.There are more and more products with short life cycles,and fast fashion products are an important category.The short life cycle makes sales forecasting more difficult,but it is necessary and important to meet consumer demand and improve corporate efficiency.High-precision forecasting also helps companies improve management efficiency and customer satisfaction.Fast fashion products have a short life cycle and large demand fluctuations,and no replenishment in a single production.And manufacturers often adopt low inventory strategies.Each physical store maintains a low inventory level due to limited inventory space and a wide range of product categories.Such production and inventory strategies will lead to serious demand censored problems,that is,some demand cannot be met.Makes sales forecasting difficult.For this reason,this article proposes a fast fashion product sales forecast research considering demand censored,which is divided into total sales forecast and each period sales forecast research.In the total sales forecast,this article explores the relationship model between demand,inventory,and sales.The demand is estimated through linear regression,and the demand censored caused by inventory constraints is measured by establishing a demand conversion level coefficient.Establish a total sales forecast model considering demand censored.Due to the strong uncertainty in the market,the forecast of futures sales is equally important for enterprises and consumers to support the logistics and distribution of enterprises and the formulation of marketing strategies to avoid large inventory backlogs in the later stages of market changes.On the basis of total sales forecast,we build a demand diffusion model based on BASS to determine the proportion of demand in each period.Taking into account the heterogeneity of products,we use the k-means algorithm for product clustering to make the demand within each category.The diffusion trend is similar enough and there are obvious differences between the categories,which helps to make more effective use of historical data.Considering the similarity between fast fashion products,we apply the fuzzy classification method to product classification,then get an estimate of the period sales based on the total sales.In the empirical analysis part,we tested the model with real data from a fast fashion company in Singapore.The experimental results show that in most cases,our model does not consider the effect of demand censored.In addition,we obtain the demand conversion level coefficients to measure the reasonable level of the company's current inventory allocation and the rationality of the production volume.This provides useful guidance for the production decisions of fashion industry managers.