Based On The Application Of Cluster Analysis Of Water Pollution Monitoring System

Along with database application and Internet application rapid development, the people's demand for information is more and more. In the face of so large and abundant data resources, the people pay attention to the matters specially, which is how to find useful data to oneself and to gains more important information in these data. Therefore, the data mining technology produced along with itClustering analysis is a important technology in data mining. Clustering is that similar data is discovered and gathered different clusre from the data sets.Through Clustering, the people can obtain the distribution of data, observe characteristic of every cluster and make the further analysis to the specific clusters. This paper on cluster analysis methods in practical application of the water pollution monitor system.This paper makes researches cluster analysis method in the water pollution monitor system application, and mainly elaborated the rivers pollution area division , cluster retrieval of the monitor data, the rivers synthesis water quality appraisal and the monitor optimal points selection.This paper apply fuzzy clustering analysis method to the rivers pollution area division. in this method, sample matrix is first composed of primitive monitor data and standardized secondly. Then the standardized data is computed by similar-degree which is a kind of new statistic. This statistic contained two parameters which are distance factor and similarity factor. It can quite good reflect similar degree of samples on "distance" and "shape" aspects. Next, based on similar matrix, hierarchical agglomerative cluster is performed. Each cluster represent a pollution area.This paper makes researchens on how to perform cluster retrieval for monitor data based on fuzzy cluster. Cluster retrieval method can produce classification of similar data through clustering, and data searching is processed based on this. It can effectively enhance the data retrieval efficiency. The process of cluster retrieval is that original data clustered firstly, then a cluster center document is obtained through computing cluster center. Based on this, cluster retrieval carried on by comparing diffirent cluster centers with the data which want to search.Fuzzy cluster evaluation method applies to the rivers water quality appraisal . Themethod determines each pollution target weight according to the actual density value and mainly through the establishment reasonable subordination function.This ensured that the appraisal result is objective, and accurate. The monitor points appraisal matrix is obtained through multiplied the pollution target weight collection by the degree of subordination matrix. Finally, a synthesis appraisal collection is obtained through multiplied the proportion that controls length of the monitor point upstream occupy the whole river length by the appraise matrix.A new fuzzy cluster analysis method based on the relation functions of matter element analysis is formulated through the combination of cluster analysis with matter element analysis theory. The main idea of the new method is that the relation functions of matter element analysis are taken as the similarity statistics between swatches. Based on these relational functions, we can create the fuzzy relation matrix and carry out cluster analysis with these swatches. Experiment on optimal selection of water monitoring points show the feasibility of the proposed method.Finally, the paper discussed some deficiencies and amelioration of cluster retrieval application in the water quality pollution monitor system.