Stok tükenmesi kaynaklı ürün ikame oranlarını satış noktası verisiyle tahmin eden bir yöntem

Empirical studies in retailing suggest that stock-out rates are considerably large formany product categories. Stock-outs result in demand spill-over, or substitution, amongitems within a product category. Product assortment and inventory managementdecisions can be improved when the substitution rates are known. In this thesis,we study the problem of estimating product substitution rates using Point-of-Sale(POS) data only. We propose a practical approach that clusters POS intervals into stateswhere each state corresponds to a specific substitution scenario. We then consolidateavailable POS data for each state, and estimate the substitution rates usingthe consolidated information. We first analyze the properties of the estimation methodunder a deterministic setting. We then provide an extensive computational analysis ofthe proposed substitution rate estimation method. A comparison with another estimationmethod from the literature and a simulation-based case study show that the estimationmethod we propose performs quite well with limited information.