Research Of Day-by-Day Behavior Mining

’With the development of information technology, more and more human behaviors are recorded by computers. Because the computer records human life day by day, we call this information Day-by-Day behavior dataset. By mining the Day-by-Day behavior dataset, we can identify the behavior patterns of human beings.In this paper, we research the data mining technology in Day-by-Day behavior dataset. The paper focus on the problems of similarity computing, peculiarity groups mining and clustering. Based on dynamic programming, we propose a Common Behavior Pattern(CBP) mining method to solve the problem of finding the behaviors that acted on different date. Meanwhile, we design an exponential function to compute the similarity between objects. After solving the similarity computing problem, we propose a graph-based method to mining the peculiarity groups in Day-by-Day behavior dataset and an algorithm for graph sparsification. From the experiments, we validate the effectiveness and efficiency of G-PGM algorithm. Lastly, the clustering problem has also been researched in this paper. We give the definition of cluster in Day-by-Day behavior dataset. Based on this definition, we design the mining and pruning technology in this paper.