| Summary: | 博士 === 國立清華大學 === 資訊工程學系 === 100 === To estimate the number of customers in target markets by taking both product competition and customer preference into consideration provides important information for decision making of product plans on product marketing. For achieving this purpose, this dis-sertation defines and solves two query processing problems of product positioning, named k-Most Demanding Products (k-MDP) discovering and k-t Most Favorite Products (k-t MFP) discovering, from different perspectives on estimating the number of customers.
Given a set of customers demanding a certain type of products with multiple features, a set of existing products of the type, and a set of candidate products which can be offered by a company, the k-MDP discovering problem is to choose k products from these candi-date products such that the expected number of the total customers for the k products is maximized. We show the problem is NP-hard when the number of the features of a prod-uct is 3 or more. One greedy algorithm is proposed to find approximate solution for the problem. We also attempt to find the optimal solution of the problem by estimating the upper bound of the expected number of the total customers for a set of k candidate prod-ucts for reducing the search space of the optimal solution. An exact algorithm is then pro-vided to find the optimal solution of the problem efficiently by using this pruning strate-gy.
On the other hand, a reverse top-t query for a product returns a set of customers, named potential customers, who regard the product as one of their top-t favorites. Given a set of products and a set of customers with different preferences on the features of the products, the k-t MFP discovering problem is to select at most k products such that the total number of potential customers is maximized. Two versions of the k-t MFP discovering problem are proposed according to whether existing products are considered in the problem. We exploit several properties of the top-t queries and skyline queries to reduce the solution space. In addition, an upper bound of the potential customers is estimated to reduce the computation cost of performing the reverse top-t query for a product. Because each ver-sion of the problem can be shown as an NP-hard problem, we provide one greedy algo-rithm with an approximation ratio (1-1/e) of the maximum total number of potential cus-tomers by using the designed pruning strategies.
Furthermore, for both topics of studies, a series of experiments on synthetic datasets and real datasets are performed to show the effectiveness and efficiency of our proposed algorithms.
|
|---|
