No abstract

The choice of particular variables for construction of a set of characteristic features relevant to classification can be executed in a kind of external process with respect to a classification system employed in pattern recognition, it can depend on the performance of such system, or it can involve some inherent mechanism, build-in in the system. The three types of approaches correspond to three categories of methodologies typically exploited in feature selection and reduction: filters, wrappers, and embedded solutions, respectively. They are used when domain knowledge is unavailable or insufficient for an informed choice, or in order to support this expert knowledge to achieve higher efficiency, enhanced classification, or reduced sizes of classifiers. The chapter illustrates the combinations of the three approaches with the aim of feature evaluation, for binary classification with balanced, for the task of authorship attribution that belongs with stylometric analysis of texts.

Typically discretisation procedures are implemented as a part of initial pre-processing of data, before knowledge mining is employed. It means that conclusions and observations are based on reduced data, as usually by discretisation some information is discarded. The paper presents a different approach, with taking advantage of discretisation executed after data mining. In the described study firstly decision rules were induced from real-valued features. Secondly, data sets were discretised. Using categories found for attributes, in the third step conditions included in inferred rules were translated into discrete domain. The properties and performance of rule classifiers were tested in the domain of stylometric analysis of texts, where writing styles were defined through quantitative attributes of continuous nature. The performed experiments show that the proposed processing leads to sets of rules with significantly reduced sizes while maintaining quality of predictions, and allows to test many data discretisation methods at the acceptable computational costs.

No abstract

No abstract