ABSTRACT
This paper presents a clustering analysis of Minimum Comparable Areas(MCAs) to draw a map of homogeneous grouping from a combination of climatic variables, soil characteristics and agricultural production. The methodology allows the visualization of interactions among the many different variables used, indentifying, for example, coexistence patterns, at the municipal level, of different crops. The discussion presents the traditional algorithms with no contiguity (hierarchical algorithm and k-means) and the agglomerative hierarchical algorithm with contiguity. Therefore, this paper seeks to explore differences among the typologies built with different approaches, as well as, provide alternative configurations of grouping. Also, the methodologies discussed allow the incorporation of traditional criteria for choosing the number of clusters, such as the CCC, pseudo-F and pseudo- t 2 statistics.
Keywords:
Spatial clustering; hierarchical algorithms; k-means
