Design And Implementation Of Contextual Marketing Based On Distributed Computing Hive And Data Mining

Due to rapid growth of the population of online shopping, Internet companies collect Terabytes-Level web logs and data every day. These data records every single consumer's activities, which have very high value to mine and research. Through processing these data, companies can get considerable potential and valuable information. Then analysts can learn consumers' behavior patterns and provide Product Marketing staff with truly reliable basis which helps them make effective marketing strategies.Traditional relational database has appeared to be inadequate when facing massive T-level data. The appearance of the distributed computing platforms (hive, hadoop) is right able to overcome the bottleneck of massive data computing, which makes data mining on massive data possible.Under the circumstance of huge number of people with steady growth who do shopping on Taobao.com, online-shops owners begin to transform their business mode from extensive type to fine type. Contextual Marketing is one of their strategies. This paper is based on massive activity logs and data of all consumers on Taobao.com, to calculate the correlation of products, using one classical Association Rule Algorithm apriori algorithm, along with ETL process. It also uses distributed computing to extract, clean, transform and load. Finally, online-shop owners are provided with the result, which they use to make Contextual Marketing strategies. This helps them improve largely on business, in order that they gain high level of user stickiness and average customer spend.