Knowledge Discovery and Data Mining in Agricultural Database Using Machine Learning Techniques

Abstract

Government of Gujarat introduced Soil Health Card which contains macro and micronutrients records of soil samples taken from farms and examined in soil laboratory to record the details into database. It is a very large database containing information of all the farms in Gujarat. To find interesting patterns from the database, we applied and analyzed concepts of data mining on soil health card database. The object classification is an important area within the field of data mining and its application extends to various areas. In this research, we have concentrated on machine learning k-Nearest Neighbor classification algorithm to classify soil sample instances into appropriate fertilizers deficiency category. There are various limitation of k-Nearest Neighbor classifier: namely, high computational, storage requirements, etc. To overcome these limitations, we proposed fast k-Nearest Neighbor, training set reduction k-Nearest Neighbor and hybrid k-Nearest Neighbor classification methods. We have applied our new approaches of classification on soil health card agriculture dataset and our evaluation illustrates that these approaches can solve the mentioned problems effectively. newline newline