Designing Of A Precision Marketing Advertising Network System Based On The Parallel To The K-means Values Algorithm

The directional delivery is one major characteristic of network ad. The new techniquesin Digital Age make this delivery possible. The network ad can be delivered to some certainspecific audiences, even the way of one to one. As a kind of advanced data analysis method,the data mining is an effective tool to realize the thorough analysis of the customer data. Thecluster analysis of data mining is an important method which can divide the samples intocertain kinds without training. The network advertising precise marketing system is anapplication of cluster analysis in the network ad delivery. This system helps the users lock thetarget audiences to carry on network ad directional delivery appropriately.This paper introduces the characteristic and the development of the network ad, and thenbased on the elementary knowledge of data mining, this paper proposes the idea of theSegmentation audiences network ad precise marketing, and proposes in this foundation themodel of network ad precise marketing based on the parallel k-average value algorithm. Alsothis paper specifies the design principle of the data base management, front processingsystem, data acquisition system, backstage data analysis system and the effect cost system,and described with emphasis the design of the audiences cluster system based on the parallelk-average value algorithm. According to the audiences' click rate, the click number, theconversion rate, recent visit number as well as the exposure time sum total which together arefive attributes, it determines the audiences appraisal function. Building the parallel virtualmachine using PVM and the Master/Slave pattern, it realizes the program design of thesegmentation audiences module .Finally it proves this system superiority through the actualeffects. At last, this paper makes the summary of how to improve further this system, and alsoindicates the direction for the next step work.Segrnentating the audiences through the parallel algorithm, locking onto target audiencesquickly, actualizing the strategy of network ad marking appropriately, and all of these areattempt to put the network ad into market, which full of great practical and theoreticsignificance.