The classification of romanian counties from agricultural point of view

Data mining techniques are used recently in more and more fields. Starting from pattern recognition (text, diseases, voice, images) to different prediction applications. Taking into account 33 variables (productions of different cereals, fruits and vegetables and agricultural areas that are cultivated with cereals, fruits and vegetables) are using data mining techniques such as informational synthesizing techniques (principal components analysis) and unsupervised pattern recognition (cluster analysis), the main goal of this article is to classify the Romanian counties into 3 major agricultural performance classes, and to identify hidden patterns of objects. After describing the variables used and analyzing the dataset, the principal components analysis reduced the dimension of the dataset at 6 principal components, while Ward's method confirmed the correct choice of 3 classes and K-means algorithm classified all 41 counties into 3 major clusters. The conclusions of this study reveals that counties with plains have a high performance in cereals production, while counties with hills and mountains produce more vegetables and fruits. From this point of view, each class has its own performance and characteristics.