Twice Set Cover Algorithm Based On Set Cover Greedy Algorithm

With the increasingly disgusted with the performance of the traditional massmarketing,commercial organizations focus to invest in another way-direct marketing，sometimes called viral marketing. Different with the mass-marketing, direct marketing doesnot select all the potential customers as the marketing object, but only targeting at a smallnumber of potential users for marketing at the beginning stage. Then, word of mouth amongusers will play the role to spread the information about products. The development of Web2.0allows the rapid development of online social networks,also provide research data andbusiness operations platform for practice of maximize the influence. Thus, in the case withthe business needs and technical support, people promote a question: If you can only select asmall number of users as sources of information at an early stage, for example, k users,howto select the k potential users to maximize the influence of marketing,so get the besteffectiveness?We introduce social network theory and the classical propagation model, and under thespreading rule defined by the linear threshold model (LM), give a twice set coveringalgorithm (TSC) base on cover greedy algorithm. Our algorithm divide the propagationinto two parts: greedy stage and the second covering the stage, try to focus on strengtheningthe influence of the initial node to other nodes on the basis of the original greedy cover, so asto achieve the goal of providing the effect of the algorithm.Comparing experimental results on four data sets, we believe that under the stagedividing parameter c=0.9, the search scale parameters S=5, the algorithm has the betterperformance. With randomized algorithms, InDegree inspired algorithm, the PageRankranking algorithm, the set covering greedy algorithm, we found that only there’s littledifference between them in a lower activating threshold, but when there’s a higher activatingthreshold, and the set cover greedy algorithm and the twice set cover algorithm issignificantly better than other algorithms, and set cover greedy algorithm effect improved.