Data Mining Technology And Its Application In Ecological Evaluation Of Greenland Landscape

The explosive growth of data in information society and the increasingly severe problems of abundant data & scarce knowledge "pull the demands of powerful data analysis tools. Decision makers need to convert large volume of data into valuable information and knowledge. The emergence of data mining provides strong technical support for the urgent need. Cluster analysis is one of the key technologies Data Mining employs. It belongs to unsupervised learning in machine learning, and is used to discover groups (clusters) in the underlying data so that the intro-cluster similarity is maximized and the inter-cluster similarity is minimized. Cluster analysis has long been studied and used in a wide variety of fields.Firstly, data mining technology and analyzes principles and general methods of clustering-based Data mining is studied. Then a new clustering algorithm is presented, i.e. the Shared Nearest Neighbor based Clustering algorithm, SNNC algorithm for short. This approach firstly find k nearest neighbors for every point, then defines a new similarity measure which is based on the number of nearest neighbors shared by two points, and calculates the density of every point. Next, this algorithm builds clusters by eliminating noise and associating non-noise points with representative or core points.Secondly, an evaluation indexes system on suitability is studied. Suitability Analysis is a special way used to study the relationship between a certain land unit and its ecological environment, which is also a core of ecological planning. This research, according to principle of the landscape ecology and its method, studied the relationship between the landscape type of Greenland, landscape environment andlandscape function. Probed into the connotation and the theoretical system of landscape suitability, find a sery of suitable landscape indexes with stability, dominant, effectiveness, and construct an evaluation indexes system which includes the greenland suitability to the environment, circumstantial harmony, ecological function and aesthetic function.At last, the Shared Nearest Neighbor based Clustering algorithm, is used in ecological evaluation on suitability of Greenland landscape.