Research And Application Of Data Mining Technology In E-commerce

Since the technology of Internet, communication and information is developingand flourishing in recent years，E-commerce is becoming a new kind model ofbusiness and concept due to the high efficiency and low cost．E-commerce hasbrought many good products and convenient online transactions to thepeople．However，considering the security issues that existing in E-commerce，andnew potential safety hazard which is incurred in the open mobile networkenvironment，the privacy problem of E-commerce is emerging gradually，and it hasthe direct bearing on the front view of its business model．In addition, many largedatabases are stored in different distributed sites. The problem of privacy protectionfrom data mining can’t be ignored in distributed environment. The technology of datamining based on privacy protection is one of the hotspots of research in the field ofdata mining in recent years. How to ensure the quality of data mining results andachieve the purpose of privacy protection while is the focus in this paper.Based on the above issues, firstly, the privacy issues arising in the distributeddata mining are analyzed on the basis of the results of previous studies in the paper.How to establish an accurate model to complete the task of data mining, and not tobetray personal privacy, are the mainly discussed issues of privacy protectionalgorithms in data mining.The paper analyzes the previous results of data mining algorithms includereconstructing the original distribution of data and dispersing the attribute of data tohide personal information privacy. Some classic association rule mining algorithms,such as Apriori algorithm and MWFI algorithm, are modified to implement privacyprotection of data mining. In addition, the paper also analyzes the four kinds of securemultiparty computation algorithm to ensure personal information privacy. The papercontinues to study the privacy issues of Web site on the basis of analysis of existingresearch results. Web server logs save the information of the customer access to page,there will be leaking users’ privacy data if not to protect. This paper discusses theprivacy protection of customer act in the Web data mining, and puts forward a methodto convert the information of Web server log into relational data tables，and throughrandomized response methods interfere with the data information, then presents thefrequent item sets and strong association rules discovery algorithm and derives the association rules of online shopping basket. The experimental results validate thealgorithms by applying it on real datasets. In order to improve the privacypreservation of the original visitor’s shopping information and mining result, aneffective method for privacy preserving association rule mining was presented. First,a new data preprocessing approach, Fake Column’s Randomized Response withColumn Replacement (FCRRCR) is proposed to transform and hide the original data.Then, an effective privacy preserving association rule mining algorithm based on bitAND operation was presented. As shown in the experimental results, the algorithmcan achieve significant improvements in terms of privacy, accuracy, efficiency andapplicability.Finally, the paper analyzes and evaluates the research works which have beendone, and put forward the future direction for improvement research.