Pesticide Toxicity Classification using Data Mining Techniques

Abstract

Agriculture is the main occupation of our working population in India. The techniques and strategies used in agriculture since ancient times varied with many different civilizations. As time varied, the attitude towards production of crops changed according to the necessities. Pesticide is considered as an integral input for crop production during the green revolution regime. The application of pesticides was justified due to social and economic consideration, when food security was the major concern. However, these estimates were made without any regard for the environmental and human health effects of pesticide use. Chemical exposure may contribute to the increasing occurrence of the health disorders and various serious unrecoverable diseases. Many are the victims who are prone to these affects due to lack of awareness. To bring such awareness, some study on agricultural data, some of the data mining techniques can be used. We can classify the pesticides based on their usage and nature using some of the classification algorithms which are available. Probabilistic methods like NaiveBayes and TAN are considered for the classification. Based on these probabilities, data classification helps to indicate the less usage of highly toxic pesticides and adopt alternative solutions to avoid human health hazards. A comparative analysis is made between these two data classification algorithms for the given data, so that to determine which algorithm is performing well with the classification of pesticide toxicity.