The Research Of Clustering Data Mining Based On Swarm Intelligence Algorithm

In the internet era,in order to avoid falling into the distress situation of "rich data,lack of information," Data mining undertakes an important mission of Extracting valuable and potential informations from huge amounts of data and implementing the data value.Data mining has become the focus of research of many scholars in the information age.Clustering algorithm is an important research field of data mining algorithm,and as a tool of data mining has applied to many fields.Swarm intelligence algorithm is a new heuristic optimization algorithm,According to simulating the biological survival,feeding and mating behavior in the ecosystem.It has the characteristics of self learning,distribution,self-organization and parallelism andit can well deal with the complex problems what the traditional calculation method is difficult to solve,especially in the aspect of data analysis.Swarm intelligence algorithm has the huge development potential in dealing with some complex optimization problems.This paper introduces the basic knowledge of data mining and several common groups of intelligent algorithms,and analyzes the problems of the clustering algorithm.The paper researchs and improves the firefly algorithm,using the improved algorithm to solve the problem of clustering.The main work is as follows.(1)In order to overcome the problems of traditional fuzzy C-means clustering algorithm,which has the problems of randomly selected from the initial clustering center,easily to fall into local optimum and the low efficiency,This paper introduces chaos theory and puts forward to a method of chaos initialization.By using Logistic mapping to change firefly location updating formula,it gets a better clustering effect.The experimental results show that the algorithm accuracy is high and less number of iterations.(2)To overcome the traditional fuzzy C-means clustering algorithm has the problems of poor global search ability,sensitive to the initial clustering center selection,the poor clustering result,the paper presents a new niche fireflies fuzzy clustering algorithm on the basis of the previous algorithm.Firstly,the algorithm uses cubic map initialization population which has better randomness and ergodicity,introducing random inertia weight to modify position updating formula in order to balance the performance of the exploration and development.The experiment result shows that the algorithm reduces the sensitivity to the initial clustering center,has high accuracy and strong stability.(3)In order to overcome the shortcomings of k–means which has poor clustering algorithm effect,excessive dependence on the initial cluster center selection and the poorglobal search ability.This paper presents a clustering algorithm which is combined with levy flight mechanism of firefly partition.The algorithm is used to initialize the population based on density and maximum-minimum distance method.And the firefly individual position update formula is introduced in levy flight mechanism,to avoid falling into local optimum and at the same time make faster convergence speed and have good global search ability,and then balance evaluation function of variance was used to optimize the objective function.The experimental results showed that the algorithm not only avoids into local optimum,improves the quality of the clustering results k-means algorithm but also weakens the degree of dependence on initial value.