Design And Implementation Of Location Sensing Impact Maximization Algorithm And Propagation Model

Influence maximization problem was first introduced into the social network field by Domingos and Richardson,and became a hot research issue.After the problem was introduced,the scholars in the field have came up various algorithms to solve the influence maximization problem.In this thesis we do our research on location-aware-business influence maximization,and the major research contents are as follows:First of all,we do research based on preference and location factors,and put forward a greedy algorithm.Existing works using meaningless edge similarity choosing methods,while in this thesis we use preference similarity and diatance factors to define edge probability,select seed set according to distance.By using this way,we eliminate the invalid nodes at the beginning and make it practical as well.Secondly,we put forward a influence-success-based regional divisionalgorithm. Existing influence maximization algorithms use the whole graphto obtain seed set, which spend a great deal of overhead. Our algorithmby first using regional division and then to select seed nodes in every partition and finally merge all the seed nodes to obtain the final resultseed set. We also put forward a influence-success-based propagation model.In this model, every activated node is assigned a influence success rate whenit trys to activate its unactivate neighbor, and this rate directly decide whether its neighbor will be activated.Finally, we use real data and simulated data to conduct experiment.From time and influence aspects to verifiy the preference and location factors based greedy algorithm as well as the influence-success-based regionaldivision algorithm.