A Novel Approach to Recommend Products for Mobile-Commerce site Using Weighted Product Taxonomy Prof. Shanti Verma Assistant Professor L.J. Institute of Computer Applications Ahmadabad, India verma.shanti@gmail.com Dr. Kalyani Patel Assistant Professor Gujarat University Ahmadabad, India patelkalyani05@gmail.com Abstract— The importance of the World Wide Web has increased enormously in recent years which leads to a large amount of information accessible through the web. The increasing importance of the web provides huge benefits to business, so to improve the business over web recommender systems has been proposed. Recommender systems works based on effective prediction which helps people for convenient access to their products that they might be interested in the real world. In this paper authors proposed a novel approach for product recommendation based on weighted product taxonomy. Product taxonomy is a hierarchical organization of products with different levels of hierarchy. Customer behavior and navigational factors are used for calculation of weight for product taxonomy. Authors also proposed a heuristic algorithm to search product “watch” in weighted product taxonomy. To prove that results of proposed heuristic algorithm are fast, authors apply independent sample ‘t’ test at 5% level of significance. Keywords—Mobile-shopping, Mobile commerce, Heuristic search, Recommendation engine, product Taxonomy, Independent sample‘t’ test I. INTRODUCTION Recommendation systems are used to predict ratings and opinions which a user might have feeling to express about the products. For example, On the basis of content provided on Netfilx customer might purchase product from Amazon. Now a days mostly all major business industries used recommendation systems to increase sales and profit. customers came across recommendation systems as they see ‘Suggested Friends’ on Facebook, ‘Suggested Videos’ on YouTube, or ‘Other Jobs for you’ on Naukari etc.. Recommender systems are based on customer behavioral and navigational patterns [10]. Behavioral patterns included the ratio of click for a specific type of product, time used in reading the profile of product and the frequency of visits to a product exist in specific category whereas Navigational patterns includes browsing status , searching status, product click status , basket placement status , and actual purchase status of a product. There are 3 types of techniques used in recommendation system. Fig.1 describes the types of techniques used in recommendation systems. Fig 1: Recommendation Systems Techniques For precise and accurate results collaborative filtering technique is used. Collaborative filtering (CF) is a technique basically used to provide personalized recommendation. Various popular websites use the CF technology include Amazon, Netflix, iTunes. In collaborative filtering, algorithms are used to make automatic predictions about a user's interests with use of several users’ preferences of same cluster [3]. In collaborative filtering techniques two techniques are used; 1) Model based and 2) Memory based. In the proposed approach of authors Model based collaborative techniques are used. Fig.2. Elaborate the architecture of model based collaborative filtering technique used in recommendation of products for Mobile-commerce sites. In the figure you can see that recommendations are based on clustering techniques. Clustering is a technique which is used to group the person having homogenous properties. To find homogeneity or similarity factor correlation technique is used. IEEE - 45670 10th ICCCNT 2019 July 6-8, 2019, IIT - Kanpur, Kanpur, India