【作者】 韩希先；

【导师】 王义和； 姜守旭；

【作者基本信息】 哈尔滨工业大学 ， 计算机科学与技术， 2006， 硕士

【摘要】 随着Internet技术的飞速发展和电子支付手段的日益完善,电子商务已蓬勃发展起来。而基于C/S结构的B2C等电子商务模式的缺点也越来越明显,并已逐步成为电子商务快速发展的制约。网络中海量的商品信息的无组织化使得商品的搜索变的复杂,而且同一商品的不同名称也给商品搜索带来了很多的不便。P2P从中心向边缘转移的特性决定了它对于电子商务应用的固有适应性,可以解决当前电子商务中的固有问题。将P2P技术应用于电子商务领域也面临着很多挑战,其中首先要解决的问题是如何组织、存储和查询数据。基于这个原因,本文提出了一种方法,在电子商务平台中利用分类树建立索引,来组织商品信息,对商品信息的管理加入名字空间的概念,将海量的数据划分成有意义的、可管理的空间,也就是社区。用户在发布商品的时候,利用分类树来进行发布,不同商品被主动划分到不同的空间中,使得商品信息主动有序,有效地组织了系统中的商品信息,提高了查询效率。而且,利用分类树来组织商品信息,整个系统中的用户使用同样的概念集注册、分类、搜索商品,对于同一个商品也就有了一个全局唯一的商品名称。本文提出一种基于分类树的搜索算法,基于分类树所划分的社区,用户要查询相应的商品的时候,只要到特定的社区中进行搜索就可以,减少了搜索的范围,提高了搜索的命中率,缩短了搜索时间。同时,为了解决社区内搜索的问题,本文提出一种基于支配集的搜索算法,结合分类树组织海量的商品信息,将商品信息划分到不同的社区中的方法,利用支配集形成的骨干网,结合基于提醒的k-walker搜索策略,在社区内进行商品信息搜索,获得了很好的效果。更多还原

【Abstract】 With the fast development of Internet and increasing perfection of e-pay, the E-commerce is in full flourish. But current mode of E-commerce which is based on C/S such as B2B has already become the obstacle of development of E-commerce. It’s complicated to search proper commodities because of the inorganization of large number of commodities in the Internet. And the problem that one commodity held by different people can has different titles makes it harder to search. The characteristic of P2P“Center-To-Edge”fits for application of E-commerce very well, which can solve the problem in current E-commerce.But there are many problems while P2P is applied to E-commerce. The first problem is how to organize、store and search commodities. For this, this paper proposes a method that Category tree is used for organising commodity information in the network. Using category tree, the commodity information in the whole system are divided into different namespace, which is convenient to organize and manage merchandise information, that is community. People use Category Tree to publish commodity information, so different commodities are divided into different communities initiatively, which make the commodity ordered actively. By this, search space is reduced and effect of search is improving. Moreover, because all users publish、organize and search commodity using unique category tree, there is only one title for every commodity.This paper proposes a search algorithm based on category tree. Because of communities divided by category tree, when users wants to search some commodity, he only needs to send his search requests to special community, which reduces the search range, improves search effect and decreases the search time. For the sake of solving the search problem in community, we propose a search algorithm based on dominating set, combining category to organize lots of commodities, dividing commodities into different communities. The algorithm uses backbone network created by dominating set, combining informed k-walker search algorithm, to search commodity in community, which is proved to be more efficient.更多还原