THE POSSIBILITY OF USING DATA MINING IN THE RESEARCH OF AGRICULTURAL HOLDINGS

THE POSSIBILITY OF USING DATA MINING IN THE RESEARCH OF AGRICULTURAL HOLDINGS

Purpose. The aim of this study was to examine the usefulness and accuracy of Data Mining techniques on the example of testing the presence of impact evaluation of the quality of the land on the level of income of agricultural holdings on the basis of test samples. Methodology. The study was analysi...

Full description

Bibliographic Details
Main Authors: Milan Milunović, Radovan Damnjanović, Nedžad Imamović, Radan Kostić, Mihailo Ćurčić, Vladimir Ristić, Dragan Bojanić
Format: Article
Language:English
Published: Naučno društvo agrarnih ekonomista Balkana, Beograd; Institut za ekonomiku poljoprivrede, Beograd i Akademija ekonomskih nauka, Bukurešt 2018-09-01
Series:Ekonomika Poljoprivrede (1979)
Subjects:
agricultural
data mining
unsupervised discriminant analysis
decision tree
Online Access:https://www.ea.bg.ac.rs/index.php/EA/article/view/888
Description
Summary:Purpose. The aim of this study was to examine the usefulness and accuracy of Data Mining techniques on the example of testing the presence of impact evaluation of the quality of the land on the level of income of agricultural holdings on the basis of test samples. Methodology. The study was analysis conducted on a random sample for identifying key factors in the research of impact evaluation of the quality of the land on the level of income of agricultural holdings, on a data set of 179 examples, where the input consists of various variables: factor of erosivity, the power of the land, reducing the pH value, presence of organic matter, then target discrete variables with two descriptive values: at a expected yield and real yield. Results and Conclusions. The results obtained from the experiments agree confirmed a physical and chemical factors properties largely determines the classification results.
ISSN:0352-3462
2334-8453

Similar Items