Algorithm for Producing Compact Decision Trees for Enhancing Classification Accuracy in Fertilizer Recommendation of Soil

Abstract

mining is the process of automatic classification of cases, based on data patterns obtained from a data set. Number of algorithms has been developed and implemented to extract information and discover knowledge patterns that may be useful for decision support. This paper proposes a technique that compact the decision tree increase the classification accuracy. The algorithm is developed by cascading the clustering and decision tree classification algorithm. The algorithm completes the process of two steps. In the first step, clustering is performed on training instances and in second step then the classification occurs on the clusters. A Schwartz criterion is used to get the optimal number of clusters. The algorithm is tested with the soil data set and various other online available datasets using WEKA. The simulation result shows that compact tree is formed, and the classification accuracy of the proposed algorithm is better than the classification accuracy of existing algorithms. The paper also presents the real-world application of proposed work in recommendation of fertilizers for soil dataset. describes the material and methods used in the paper with the existing technique and its drawback. The section 3 describes the modification of the existing technique and its implementation results on various datasets. The final section gives the application of the proposed algorithm in real world for fertilizer recommendation.