Based On Data Mining Product Information Delivery Platform For Research And Implementation

The rapid growth of e-commerce websites'information brings a lot of difficulties to the development of e-commerce industry. Because the consumer find it difficult to retrieve their interested product information, and sellers are also facing the difficulty that the product information can't be sent to their target consumer accurately. To solve these problems, experts propose to the personal recommendation technology. Currently, most of the personal recommendation systems only take into account the description of the product itself, but not the news of the product, such as news about discount goods, news about product advertisement, news about product promotion and so on. It is an important product promotion approach for the offline shop to send these news to the consumer. Because the news can effectively attract consumers to purchase product. So, this paper proposes to a delivery platform for product news, which helps the seller send the product news to their target consumer accurately.To be satisfactory for product news delivery platform's requirements, the paper proposes to the LSI-based delivery model for product news. By analyzing the basic process of information delivery,it's obvious that the recommended strategy is the most critical factor of the accurately matching calculation in information delivery, and the information representation model is the most important part of the the recommended strategy. Compared with other information representation models, the Latent Semantic Analysis model takes into account the semantic structure of text, and effectively solves the synonym and polysemy problems, so the Latent Semantic Analysis model takes full advantage of other information representation models in information representation field. The LSI-based delivery model for product news preprocess a large number of training text, and extracts the feature from those training text to construct the original term-document matrix. And the model performs a singular value decomposition of the original term-document matrix, and the high dimension matrix of document in vectore space model is changed to low dimension matrix in latent semantic space,but the new latent semantic space keeps the semantic relationship of the original word-document matrix, and ignores the noise information from that a word always has several different usage, so that the precision of the delivery of product news is improved. In addition, consumer demand model reflects the interest of consumers, and it is closely related to the precision of the delivery of product information. So the analysis of the disadvantage the consumer demand model. This paper analyzes the disadvantage of the consumer demand model, and apply ontology to improve the ability of the expression of consumer demand model, thus futher improving the precision of the delivery of product news significantly.Aiming to improve the precision of the delivery of product news, the paper proposes to the LSI-based delivery model for product news creatively. In addition, the LSI-based delivery model for product news is applied to construct the delivery platform for product news. By the knowledge of object-oriented design, the pager complete the design and implementation for the project. And the paper designes the experiments to validate the precision and recall of the model.