Loading...

Personalized Assortment Optimization Using a Non-Parametric Choice Model

Seyed Ghafouri, Mohammad Mahdi | 2020

248 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 53158 (01)
  4. University: Sharif University of Technology
  5. Department: Industrial Engineering
  6. Advisor(s): Modarres Yazdi, Mohammad
  7. Abstract:
  8. In this study, the problem of assortment optimization considering a non-parametric choice model will be introduced. The present study is generally divided into two main parts. The first part deals with modeling customer behavior using the definition of tree models of customer preference behavior. The second part is related to the main problem solving framework in order to get the best product combination in the product classification basket. This research is related to the online seller who intends to sell a certain number of types of products without inventory limit. Customers entering the system, according to their specifications, are provided with the appropriate product classification assortment. In order to personalize the presentation of the product classification assortment to customers, in this research, clustering techniques are used in such a way that products are divided into different categories according to the historical sales data set, each of which is called a cluster. Corresponding to each cluster, a tree model for customer selection behavior is obtained using the innovative algorithm proposed in this research, and finally, the set of all these models of customer selection behavior tree is called the forest selection model. Thus, the main purpose of this research is to provide the best product classification basket to incoming customers according to their specifications so that we can have the maximum possible revenue from the sale of products. In this research, the main framework used for optimization is the dynamic planning model and in order to evaluate the performance of the proposed model and algorithms, the historical sales data set of one of the online vendors is used. Experiments show that the performance of the proposed algorithm and model has the ability to provide solutions with performance accuracy of about 80 to 95%, which will be examined in detail below
  9. Keywords:
  10. Dynamic Programming ; Assortment Planning ; Customer Choice Behavior ; Consumer Choice Model ; Assortment Optimization

 Digital Object List

 Bookmark

No TOC