An intelligent fuzzy-based recommendation system for consumer electronic products

https://doi.org/10.1016/j.eswa.2006.04.012Get rights and content

Abstract

Developing an intelligent recommendation system is a good way to overcome the problem of overloaded products information provided by the e-commerce enterprises. As there are a great number of products on the Internet, it is impossible to recommend all kinds of products in one system. We believe that the personalized recommendation system should be built up according to the special features of a certain sort of product, and forming professional recommendation systems for different products. In this paper, based on the consumer’s current needs obtained from the system-user interactions, we propose a fuzzy-based system for consumer electronics to retrieve optimal products. Experimental results show the system is feasible and effective.

Introduction

In recent years, the on-going advances of Internet and Web technologies have promoted the development of electronic commerce, which has caused both companies and customers to face a new situation. In order to expand their markets and create more business opportunities, enterprises have been developing new business portals and providing large amounts of product information, as a result of which customers have more opportunities to choose various products that meet their needs.

Information from Internet marketing and electronic commerce is of uncertainty, but it really has potential and impact. It expands the opportunities for branding, innovating, pricing, and selling. However, the exponentially increasing information along with the rapid expansion of the business Web sites causes overload of information. So consumers have to spend more and more time browsing the net in order to find the information needed. One way to overcome the above problem is to develop intelligent recommendation system to provide personalized information services (Schafer, Konstan, & Riedi, 2001): retrieving the information a consumer desires and helping him determine which one to buy. The purpose of personalized information services is to adjust strategies of promotion and advertisement to fit customer interests.

New enterprises can deal with a particular customer’s Web experiences by providing customer personalization service and communicating or interacting with customers. Such understandings of customers can be transformed customer information into quality services or products (Weng & Liu, 2004). To improve customer satisfaction, feedback rate, loyalty, Web sales, and reputation, one-to-one marketing is seen as the most effective approach for customer relationship management. However, with the great number of customers, how do enterprises identify their interests? The answer to this question is to build personalized Internet services. The purpose of personalization is to adjust strategies of promotion and advertisement to fit customer interests. First, it is necessary to understand customer interests and preferences and then provide suitable products or services at a right time. The mechanism of this research aims not only to promote visiting rate of Web stores, increase opportunities of selling, and advertisement revenue, but also to increase a website’s profits.

Depending on the sort of product, various personalized recommender systems can be built up to guide the consumers in a large product feature space. For such frequently-purchased products as books, CDs, and DVD films, recommendation systems can be developed to reason his preferences by analyzing his personal information, browsing history, and the products he purchased. By contrast, a common consumer less frequently-purchased such commodities as notebook computers and digital cameras, what is more, enterprises lack enough information about the customer’s past purchases and his specific requirements for a particular product, so it is difficult and impossible to reason a customer’s previous preferences. In this situation, advises from domain experts are strongly demanded. Recommendation systems are thus expected to have specific domain knowledge and capability to interact with consumers. Consequently the systems can acquire and analyze a customer’s current needs on some kinds of products he identified, and then evaluate the relevant products to help him choose the optimal ones.

With the personalized recommendation systems, consumers can immediately access the information they are interested in, and save their time for reading the electronic documents. On the other hand, enterprises can get to know customers’ buying behaviors and then develop most appropriate marketing strategies to attract different customers and efficiently deliver the information they are interested in. The customer’s satisfaction and loyalty can thus be increased, and the increase in the visiting frequency of the customers can further create more transaction opportunities and benefit the Internet enterprises.

In this approach, we present a fuzzy-based recommendation system for those less frequently-purchased products, especially for consumer electronics. When a common buyer is going shopping, some distinctive features of consumer electronics may get him into some trouble. First, compared with other products, the life of a new model consumer electronic is short, namely about 2 years. As the new models of a product come out at all times in the market, it makes a common consumer difficult to know all models of a product. Second, accompanied by newly-produced models, a great number of new techniques come forth to improve product functions. Eventhough knows the details of the techniques in a new model, he still does not know whether it is worth spending more money on it. Last, because of the widening price gap among the different models of the same product and the continuously dropping price of a new model within its life cycle, it is hard for a common consumer to know the prices of all product models. For new models, the price falls by half within several months since it comes into the market.

The proposed system aims to assist a consumer to navigate the product feature space in an interactive way in which the consumer has his own need in each feature dimension so that the customer can find the optimal products according to his personal preferences. We have also built up a system of this kind for laptop computer recommendations. The experimental results show that both systems can give sensible recommendations, and adapt to customers’ up-to-date preferences. The remainder of the paper is organized as follows. In Section 2, research background is expatiated, including personalization and recommendation systems. Section 3 gives an overview of the recommendation system. The elementary theoretical background is provided in Section 4, followed by Section 5 explaining the implementation issues of the proposed method. Section 6 reports the experimental process and the results of the study. Finally, the conclusion is given in Section 7.

Section snippets

Research background

The purpose of this research is to build up a personalized recommendation system based on product features.

System architecture

In this section, an on-line intelligent recommendation system for personalized Internet shopping is proposed, which uses data mining techniques and fuzzy logic in accordance with the proposed marketing strategies to help the business prepare the highly potential and suitable promotion products for each individual customer. The recommendation system is demanded when a customer is going to buy such less frequently-purchased products as laptops. As analyzed in Section 1, the experiences of buying

Linguistic definition and fuzzy numbers

Based on the proposed system, the consumer needs and the candidate product features can be expressed in an appropriate way. In the approach, we use triangular fuzzy numbers to characterize consumer needs and product features.

A triangular fuzzy number is a particular case of fuzzy sets. It has a triangle-shaped membership function, which can be viewed as possibility distribution. It is supposed that q˜ is a triangular fuzzy number with membership function μp˜(x), and is denoted as q˜=(q1,q2,q3),

Implementation methods

Since collecting and analyzing a consumer’s personal needs are the basis of the system, our aim is to establish a transformation model for translating customer needs into optimal combination suggestions of applicable alternatives. To establish this model, the relationship between customer needs and product features needs to be constructed. Utilizing fuzzy operation, optimal alternative searching is performed based on the consumer’s subjective needs. The procedure for establishing this system is

Experiment and results

The proposed system could also be applied to recommend the products that a consumer generally does not often buy in a short period of time and has his specific needs in each single purchase. And because of the reasons mentioned in Section 1, it is especially suitable to recommend consumer electronic products, such as cell telephones, digital cameras, PDA and so on. Therefore the experiments concentrate on evaluating the system behaviors; that is, we shall observe whether the overall system can

Conclusions

In this paper, we first state that, in addition to developing or improving the software and hardware equipment directly related to the Internet infrastructure, Internet enterprises need to provide personalized information services to make a successful Internet business. Then we suggested that developing personalized recommendation system is a promising way to achieve this goal. Finally, we build up a personalized recommendation system for the consumer electronic products.

Because the consumer

References (10)

There are more references available in the full text version of this article.

Cited by (202)

View all citing articles on Scopus
View full text