Predicting crop diseases using data mining approaches: Classification

Umair Ayub, Syed Atif Moqurrab

2018 · 30 citations · 16 references

Concepts

Abstract

Agriculture research is rapidly growing, due to advancement of technologies and upcoming challenges. It has been proven to be leading role in improving the overall growth rate of any country. Especially in Pakistan, there is a dire need to do extensive research for better productivity in agriculture. To improve the growth rate of agriculture, researchers of this domain used different data mining techniques to solve agriculture related problems. Data mining approaches such as classification helps to predict the crops diseases, production and loss. It supports farmer while taking right decisions. This paper focuses on prediction of loss due to grass grub insect. We analyze the damages by using well-known classifiers such as Decision Tree, Random Forest, Neural Networks, Naïve Bayes, Support Vector Machines and K-Nearest Neighbor and design Ensemble Models of above mentioned classifiers which gave better results as compared to classifiers. Neural Networks and Random Forest produced slightly better results than other classifiers. Ensemble model improve the results of weak classifiers and proven as fruitful technique for our agriculture related problem. To improve the results further, hybrid of evolutionary algorithms and data mining techniques will be used which is our future research direction.

References

16