Facilitating electronic commerce: Auction, negotiation, and filtering

Electronic Commerce is on-line businesses, which have areas such as: searching for product information, ordering products, paying for goods, services, and customer service. Electronic commerce over the Internet has gained popularity in recent years. As electronic commerce grows, many problems emerge. The need for a new market mechanism, the need for a good supporting system, and the need of a good Web information filtering system are among those problems. A semi-continuos electronic auction market mechanism, a negotiation support system, and a Web filtering system is proposed in this dissertation for those problems. The proposed auction market mechanism takes advantages of both prompt sales in continuous trading and the fair price in periodic trading. Also, this system provides fast matching because it employs parallel computational architecture. The proposed negotiation support system is an improvement of existing system (TELCOT). The proposed system provides potential seller selection and negotiation price estimation. Those two functions can improve buyer's information processing capability but they are not considered in the original system. The Web information filtering system improves the classification capability from the commercial Web information classification systems by employing isotonic separation method. The isotonic separation method has superior accuracy to point separation methods or linear separation methods. The proposed Web filtering system has machine learning capability that helps the system responses to individual users according to individual's preference. Also the isotonic separation method is a network model which is computationally efficient. The performance of the proposed Web filtering system is demonstrated with artificially generated data sets.