AN APPROACH TO CONSTRUCT DECISION TREE USING SLIQ AND KNN FOR LAND GRADING SYSTEM

Abstract

India is a nation of farmers where most population of country is dependent on the crops and agriculture. But poor land quality and their composition affect the performance of crops production. Therefore according to the soil chemical composition, their categorization required. In order to design an appropriate, efficient and accurate classifier the decision tree algorithm is selected for data modeling. Therefore first using the soil composition a dataset is prepared and classification algorithm is optimized for accurate classification. The classification algorithm first utilized SLIQ decision tree for preparing the decision tree than KNN algorithm is applied on decision tree to extract the rules. These rules are optimized further for reducing the number of comparisons during classification. Additionally for justifying the proposed rule based solution the presented data model is compared with the traditional ID3 and SLIQ algorithm. The proposed is implemented using visual studio development technology and the performances of the algorithms are compared with traditional algorithms based on the various qualities of service parameters such as memory consumption, training time, search time and accuracy. According to the obtained results the previous algorithm shows 80% accuracy rate compared to which implemented algorithm shows 90% of accuracy.