Acta Univ. Agric. Silvic. Mendelianae Brun. 2017, 65, 1007-1014
Classification of Specialized Farms Applying Multivariate Statistical Methods
Classification of specialized farms applying multivariate statistical methods The paper is aimed at application of advanced multivariate statistical methods when classifying cattle breeding farming enterprises by their economic size. Advantage of the model is its ability to use a few selected indicators compared to the complex methodology of current classification model that requires knowledge of detailed structure of the herd turnover and structure of cultivated crops. Output of the paper is intended to be applied within farm structure research focused on future development of Czech agriculture. As data source, the farming enterprises database for 2014 has been used, from the FADN CZ system. The predictive model proposed exploits knowledge of actual size classes of the farms tested. Outcomes of the linear discriminatory analysis multifactor classification method have supported the chance of filing farming enterprises in the group of Small farms (98 % filed correctly), and the Large and Very Large enterprises (100 % filed correctly). The Medium Size farms have been correctly filed at 58.11 % only. Partial shortages of the process presented have been found when discriminating Medium and Small farms.
Keywords
discriminatory analysis, linear discriminant analysis, classification of farms, economic size, canonical analysis, Farm Accountancy Data Network.
Funding
The information and data presented in the paper have been obtained thanks to the Grant No. 20161011 by the Internal Grant Agency, Faculty of Economics and Management, Czech University of Life Sciences Prague, titled “Long Term Structural Changes of Czech Agriculture and the Impact of These on Self-sufficiency in Farm Production ”.