Condition Analysis and Forecasting in the Fashion Industry

The work considers the problem of demand for the clothing industry's goods. It shows how this problem is connected with the mathematical problem of the partition of the set. Investment decisions depend on a diagnosis based on forecasting demand in individual product groups. These groups are characterized by a number of features and even in the simplest situations (3 attributes) lead to computationally complex situations. In this situation, the recursive partitioning method can be used. This is a method related to the construction of classification trees (regression). These methods are widely used in natural, technical and economic sciences. The main direction of their applications is to support decision-making processes. The article shows how to support the construction of classification trees. The paper proposes a practical solution to the problem using the method of random partitions. The proposed method can be a complement to the recursive partitions method, or used in some situations instead. The submitted method is a practical proposal to avoid the problem of computational complexity. The numerical example shows how to replace a population of about 52 trillion by a sample of only 100. The applied method was justified by an example of a less numerous population, where the result could be verified empirically by reviewing all possibilities. Such verification is not practically possible in the case of 20 product profiles. Such a number generates a number of partitions amounting to almost 52 trillion. The article also presents the estimation of the calculation time. These results are useful from a practical point of view, although they are not optimal.