@article{609, author = {Alfred Holl}, title = {Inflectional Morphology, Reverse Similarity and Data Mining – Finding and Applying Compact and Transparent Descriptions of Verb Systems of Natural Languages}, journal = {International Journal of Computational Linguistics Research}, year = {2011}, volume = {2}, number = {2}, doi = {}, url = {http://www.dline.info/jcl/fulltext/v2n2/3.pdf}, abstract = {Under the term “data mining”, the field of computer science includes many different techniques for data analysis, among them methods of cluster analysis. In the approach presented, a special method is designed for the analysis of inflectional systems. The algorithm is independent of individual natural languages and parts of speech. It finds two types of clusters: morphologically homogeneous ones, which contain reversely similar (which possess the same trailing letters), morphologically analogous lexemes of the examined language-part-of-speech combination, and morphologically inhomogeneous ones, in which the largest part of the lexemes is morphologically homogeneous. The resulting registers are compact and transparent as well as easily extensible and correctible. In condensed form, they provide linguistically and didactically usable, structural results on inflectional systems. For instance, it is possible to assign arbitrary lexemes to clusters with a detailed explanation based upon the structure of an inflectional system. The approach is most often applied to verb systems in inflecting and agglutinating language.}, }