基于数据挖掘的商品资讯投送平台研究与实现
发布时间：2018-10-08 15:46
【摘要】：随着电子商务网站信息量的迅速增长,给消费者获取感兴趣的商品信息带来了困难,卖家也面临着很难将商品信息精准发送给目标客户的困境,而这些困难成为制约电子商务网站持续发展的问题。为解决上述问题,行业提出了商品个性化推荐技术。目前个性化推荐技术的推荐对象基本以商品本身为主,很少关注如商品打折、商品广告、商品促销等类似新闻的商品资讯文本信息。而向消费者发布这些资讯信息都是线下商店推广商品很重要的方式,对消费者购买行为有很强的影响。针对该问题,本课题构建一个商品资讯投送平台,帮助卖家将商品的资讯准确投送给目标客户。 针对商品资讯投送平台的需求,本论文提出了一个基于潜在语义索引的商品资讯精准投送模型。通过分析资讯投送的基本过程,很显然影响精准匹配的最关键问题是推荐策略,而推荐策略中最重要的部分是信息表示模型。相比其它信息表示模型,潜在语义索引模型考虑到文本的语义结构,并有效地解决了同义词和多义词的问题,在信息表示的准确方面比其它几种模型都更具优势。基于潜在语义索引的商品资讯投送模型是对大量训练文本进行预处理,将这些文本集表示成词-文本矩阵,利用奇异值分解将词-文本矩阵从高维矩阵降到低维矩阵,形成潜在语义空间。生成的潜在语义空间保留了原词-文本矩阵的潜在语义关系,同时去掉了因具体用词方式不同而带来的噪声信息,从而提高商品资讯投送的精准度。除此之外,消费者需求模型反映消费者的兴趣,对商品资讯投送精准度有很强的影响,本论文通过分析消费者需求模型的不足,利用本体论改进消费者需求模型表达消费者需求的能力,从而进一步提高商品资讯投送的精准度。 本论文围绕实现商品资讯精准投放的目标,创造性地提出了基于潜在语义索引的商品资讯精准投送模型。除此之外,本论文还将该商品资讯精准投送模型应用到商品资讯投送平台系统的构建中,并结合面向对象设计知识完成了对商品资讯投送平台系统的设计与实现。最后,论文设计实验方案,验证商品资讯精准投送模型的关键性能。
[Abstract]:With the rapid growth of information on e-commerce websites, it is difficult for consumers to obtain information of interest, and sellers are also faced with the difficulty of accurately sending commodity information to target customers. And these difficulties become the problem that restricts the sustainable development of e-commerce website. In order to solve the above problems, the industry put forward the personalized recommendation technology. At present, the recommendation object of individualized recommendation technology is basically the commodity itself, and little attention is paid to the text information of commodity information such as commodity discount, commodity advertisement, commodity promotion and so on. Releasing this information to consumers is an important way for offline stores to promote products, which has a strong impact on consumers' buying behavior. To solve this problem, this paper constructs a commodity information delivery platform to help sellers accurately deliver the information to the target customers. In order to meet the demand of commodity information delivery platform, this paper proposes a model of accurate delivery of commodity information based on latent semantic index. By analyzing the basic process of information delivery, it is obvious that the most critical problem affecting accurate matching is the recommendation strategy, and the most important part of the recommendation strategy is the information representation model. Compared with other information representation models, the latent semantic index model takes into account the semantic structure of the text, and effectively solves the problems of synonyms and polysemous words, and is superior to other models in the accuracy of information representation. The commodity information delivery model based on latent semantic index preprocesses a large number of training texts, and represents these text sets as word-text matrices, and reduces the word-text matrix from high-dimensional to low-dimensional by singular value decomposition. Form the latent semantic space. The generated latent semantic space preserves the latent semantic relationship of the original word-text matrix and removes the noise information brought about by the different ways of using the words so as to improve the accuracy of the delivery of commodity information. In addition, the consumer demand model reflects the interests of consumers and has a strong impact on the accuracy of the delivery of commodity information. Using ontology to improve the ability of consumer demand model to express consumer demand, so as to further improve the accuracy of commodity information delivery. In this paper, we creatively propose a model of accurate delivery of commodity information based on latent semantic index. In addition, this paper also applies the model to the construction of commodity information delivery platform system, and completes the design and implementation of commodity information delivery platform system based on object-oriented design knowledge. Finally, the paper designs the experimental scheme to verify the key performance of the model.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TP311.13
本文编号：2257332
[Abstract]:With the rapid growth of information on e-commerce websites, it is difficult for consumers to obtain information of interest, and sellers are also faced with the difficulty of accurately sending commodity information to target customers. And these difficulties become the problem that restricts the sustainable development of e-commerce website. In order to solve the above problems, the industry put forward the personalized recommendation technology. At present, the recommendation object of individualized recommendation technology is basically the commodity itself, and little attention is paid to the text information of commodity information such as commodity discount, commodity advertisement, commodity promotion and so on. Releasing this information to consumers is an important way for offline stores to promote products, which has a strong impact on consumers' buying behavior. To solve this problem, this paper constructs a commodity information delivery platform to help sellers accurately deliver the information to the target customers. In order to meet the demand of commodity information delivery platform, this paper proposes a model of accurate delivery of commodity information based on latent semantic index. By analyzing the basic process of information delivery, it is obvious that the most critical problem affecting accurate matching is the recommendation strategy, and the most important part of the recommendation strategy is the information representation model. Compared with other information representation models, the latent semantic index model takes into account the semantic structure of the text, and effectively solves the problems of synonyms and polysemous words, and is superior to other models in the accuracy of information representation. The commodity information delivery model based on latent semantic index preprocesses a large number of training texts, and represents these text sets as word-text matrices, and reduces the word-text matrix from high-dimensional to low-dimensional by singular value decomposition. Form the latent semantic space. The generated latent semantic space preserves the latent semantic relationship of the original word-text matrix and removes the noise information brought about by the different ways of using the words so as to improve the accuracy of the delivery of commodity information. In addition, the consumer demand model reflects the interests of consumers and has a strong impact on the accuracy of the delivery of commodity information. Using ontology to improve the ability of consumer demand model to express consumer demand, so as to further improve the accuracy of commodity information delivery. In this paper, we creatively propose a model of accurate delivery of commodity information based on latent semantic index. In addition, this paper also applies the model to the construction of commodity information delivery platform system, and completes the design and implementation of commodity information delivery platform system based on object-oriented design knowledge. Finally, the paper designs the experimental scheme to verify the key performance of the model.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TP311.13
【引证文献】
相关硕士学位论文 前1条
1 严水发;基于Agent的个性化服务平台的应用研究[D];中南大学;2012年
，本文编号：2257332
本文链接：https://www.wllwen.com/wenyilunwen/guanggaoshejilunwen/2257332.html
