Analýza a klasifikace dat
| Title in English | Data Analysis and Classification |
|---|---|
| Authors | |
| Year of publication | 2012 |
| MU Faculty or unit | |
| Citation | |
| Description | The book “Data Analysis and Classification” links to a certain extent to the textbook “Multivariate Statistical Methods” and theoretically evolves topics described there. Both publications are intended mainly for students of the Computational Biology study programme at the Faculty of Science of the Masaryk University. They were both supported by the ESF grant no. CZ.1.07/2.2.00/07.0318 „Multidisciplinary Innovation of Study in Computational Biology“. Chapter one introduces fundamentals and principles of data processing and analysis, defines their aims and describes individual phases of processing of static and dynamic data. The second chapter, the longest in the publication, deals with individual feature based methods of statistical pattern recognition. The chapter begins with methods of classification by means of discrimination functions, which are demonstrated by description of Bayesian classifiers. This is followed by methods of minimum distance classification. For this purpose the terms metric and similarity metric, respectively, are defined and subsequently extended by specific metrics for determination of distance and similarity of two patterns described by quantitative and qualitative features. These metrics are further developed into deterministic and probability methods of determination of distance between two sets. Another part of the chapter is focused on algorithms of classification into classes defined by borders in feature space and, finally, methods of sequential classification. The third chapter analyses methods for selection of attributes; the highest emphasis is put on analysis of principal components and analysis of independent components. Chapter four, the last but not least, is focused on methods of structural analysis and classification. The chapter introduces terms such as primitive, relation, and relation structure, deals with methods of structural description of classification classes, particularly grammars and automats, and is concluded by description of methods of structural pattern recognition of both non-deformed and deformed relation structures. |
| Related projects: |