Development and Evaluation of Different Matching Algorithms for Online Shop Products

Online shops typically contain thousands of products. In addition, these products are described with different texts and images depending on the different stores. The challenge of a product matching algorithm is to identify offers from different online shops that represent the same real product. Although, the varying information like different data structures, unequal texts and differing images raise the difficulties of finding the same products. Therefore, in this thesis three different product matching algorithms were developed, implemented and evaluated. The challenges, that were already known before the implementation, but also the additional once appeared during the development of the algorithms will be explained. Also, the different structures of the data form the online shops will be shown and a detailed explanation of the different matching algorithms provided. More, the evaluation results of the three matching approaches will be shown and explained.