Research Of Recommended Method In O2O E-commerce Model

With the rapid development of e-commerce, O2O e-commerce, as a new and efficientbusiness model, rapidly spreads. Along with the appearing of online group-buying websites,more and more companies focus on the O2O mode of e-commerce. Nevertheless, as thesituation of data exploded, a huge amount of data makes the data overload and redundant,which usually makes the users get lost in the sea of data and be difficult in finding theproduct successfully they need. Against this background, personalization technology, as a keytechnology of e-commerce, is becoming one of the research focuses in the area of O2Oelectronic commerce. However, there are relatively few in-depth studies on personalizationtechnology in the area of O2O electronic commerce, and the accuracy and efficiency of theserecommendation algorithms still needs to be improved.In this thesis, the author chooses the recommendation method in the O2O e-businessmodel as the research object. Based on the analysis of the existing e-commerce personalizedrecommendation technology, the thesis proposed a suitable recommendation method for O2Omode of e-commerce, and improved the collaborative filtering algorithm. Firstly, the authorintroduced the theory of personalized recommendation, compared the differentrecommendation technologies, and analyzed the applicability of several recommendationtechnologies to the O2O mode of e-commerce. Secondly, the author made a detailed researchon the personalized recommendation of O2O e-business model. Based on the summary of thedata character in O2O e-business, the author proposed recommendation model in O2Oe-business. Considering of the deficiency of the Big Data-recommendation technique in O2Oe-business website, the author proposed an advanced recommendation function based oncollaborative filtering algorithm which introduced advanced initial clustering center andfuzzy cluster of distant function into recommendation function in order to improve thecoverage rate and match scores of the recommendation model. Finally, the author made a testby experiments and applied it to the website of cinema. The experiment results show that therecommendation method put forward in the thesis is able to meet the demand of the O2Oe-commerce mode. Especially, it improves the recommendation efficiencies in big dataenvironment and possesses better real-time performance and accuracy.