An Improved Apriori-Based Personal Recommendation Algorithm for E-commerce

In order to improve the quality of the recommended result, the personalized recommendation system should identify the similarity degree of visitor´s accessing behavior so as to predict customer´s interests. The key technology is to calculate the similar distance among different objects over either all or only a subset of the dimensions. This paper, first of all, analyses the commonly-used methods and points out their shortages, and then proposes an improved apriori-based personal recommendation algorithm for e-commerce. This algorithm considers overall the minable data source, users´ similarity metric and k-support bound to get the data of those access Web pages, construct a matrix model having relatively high purchasing power about customer behavior, get the similar access behavior over the all or partial property space with high efficiency, help the customer find out the merchandise he wishes to buy through the mine of the similar pattern character between latent buyer and high buyer, promote customer satisfaction and truly promote the sale achievements for the enterprise.