Research On Recommendation Algorithm For E - Commerce Application

With the rapid development of Internet, e-commerce is booming at an incredible speed. The new commercial environment for enterprises provides new business opportunities, at the same time, also puts forward a new challenge to the enterprises. Customer relationship management (CRM) focusing on consumers, is the electronic commerce environment enterprise attract customers and enhance the cohesive critical. How to attract new clients in a rapidly changing era of e-commerce and improve users’ experience is very important. The solution is website offering excellent products and perfect service to attract customers, e-commerce enterprises. On the other hand, for customers, in the face of so many choices, to pick out what they really need is equivalent to search for a needle in a haystack. In recent years the rise of recommendation system has become one of the important ways to solve these problems.Recommendation system is based on personal preferences and habits to offer information and goods. The initial research motivation comes from the Internet information explosion. People usually use search engines to find the required content, but it is difficult for most of them to use a few short keywords to accurately describe their needs. It leads to the result that they cannot either get any results, or have to get the answer from a large list of result and then check them one by one. So we want to deliver a system which will guess the user’s intention, which can catch goods users like and dislike. It will also select automatically and filter the content out which users don’t like. Nowadays, the mainstream of international e-commerce sites have quite a few successful examples of recommendation system. The recommendation are related not only newsgroup postings but also about DVD, CD, book, etc.Recommendation systems can be applied on ecommerce sites on the consumption of customer preferences, recommending products targeted to each customer, helping users choose from a huge catalogue really suits your need goods. Recommendation system helps the customer at the same time, also improves the customer satisfaction of e-commerce platform, to some extent further supporting of e-commerce sites. Generally speaking, recommend system for e-commerce platform has a positive role as below:1) To help users to get useful information2) To promote sales3) To offer the personalized service 4) To improve customer’s loyaltyThere are many systems, but they all have their own defects or deficiencies, we hope to do the summary about these framework. The common challenge for recommendation system is how to offer suitable products for uses and how to put products in recommendation lists of suitable users. We also find that it’s very difficult to offer effective recommendation for new users. It’s also a challenging work to recommend when we don’t have enough data to use and which is the same to new products. So we would like to improve the classical collaborative filtering. We provide the algorithm based on the user’s interest. Our method will extract the attribute characteristics of products and we will merge the score on different characteristics. Our experiments show that this algorithm works well when users have less number of neighbors. Then we will also offer a new recommendation framework and apply the emerging of the insurance e-commerce platform. We also involve the content based recommendation on insurance e-commerce platform which will solve the cold start problem. Above all, we hope our framework will help more and more e-commerce platforms to offer the recommendation service and improve the users’ satisfaction.