ENHANCING THE CLOTHES SEARCHING SYSTEM USING COMBINATION OF K-PROTOTYPE AND K-NEAREST NEIGHBOR SEARCH ALGORITHM

Abstract

In clothing shops, finding user’s desired dress manually is time consuming and needs a large amount of human labor. Although k-Nearest Neighbor (kNN) search algorithm is one of the most popular methods among many classification tools, it has resulted in higher computation time complexity as it used all training data to classify a new test sample. In order to solve the above mentioned problems, it is necessary to combine conventional kNN algorithm and K-Prototype algorithm depend on partitioning methods for this search system. In proposed system, kNN search algorithm is mainly used to find the similar items with user preference and to produce the dress codes of the item. There are two main stages in this system. The first stage is the grouping stage, which is clustering the clothes group depending on their attribute values by using K-Prototype algorithm. The second stage is testing stage, which is searching the most similar group depending on user choices by using kNN search algorithm. The proposed system is then analyzed in terms of processing time and root mean square error (RMSE) values and compared to highlight the processing time by using the combination of K-Prototype and kNN algorithms.