Application research of cluster analysis and association analysis

For applications of data mining techniques in geosciences, through mining spatial databases which are constructed with geophysical and geochemical data measured in fields, the knowledge, such as the spatial distribution of geological targets, the geophysical and geochemical characteristics of geological targets, the differentiation among the geological targets, and the relationship among geophysical and geochemical data, can be discovered. Due to the complexity of geophysical and geochemical data, traditional mining methods of cluster analysis and association analysis have limitations in processing complex data. In this paper, a clustering algorithm based on density and adaptive density-reachable is presented which has the ability to handle clusters of arbitrary shapes, sizes and densities. For association analysis, mining the continuous attributes may reveal useful and interesting insights about the data objects in geoscientific applications. Quantitative association rules aims to deal with the relationships among continuous attributes of geoscientific data objects. An association analysis algorithm based on the distances among clusters projected on attributes is presented in this paper. Experiments and applications indicate that the algorithms are effective in real world applications.