Classifying rubber breed based on rough set feature selection

Abstract

Rubber is the economic crop that is planted widely in almost all regions of Thailand and makes a lot of income for the export of this country. Selecting a rubber breed for a particular region is one of the principal factors for the achievement of the rubber plantation. If the agriculturists get the rubber breeds unsuitable to be plant in their rubber garden, once the time to slit, the rubber water may have low quality and quantity. The objective of this work is to generate the rubber breed classifier by using the k-nearest neighbor technique based on selected set of features of rubbers. Rough set feature selection is proposed in this research to select a subset of relevant features of rubber optimally while retaining semantics. The data samples of 10 well-known breeds of rubber, 30 samples per breed, cultivated in the northeast of Thailand were used to generate the breed classifier. The accuracy rate of classifying the breed of rubber is rather good. Therefore, this generated breed classifier can assist the agriculturists classify and select the correct breed of rubber from the features of rubber in hand before cultivate in the rubber garden.