AN INVESTIGATION OF MACHINE LEARNING TECHNIQUES FOR THE CLASSIFICA-TION OF IOT-ENABLED SMART IRRIGATION DATA

ABSTRACT

Recently, different smart IOT irrigation devices have been fabricated to aid farming all-round the year. For the fact that agriculture depends on adequate supply of water to farms, it is important to predict the soil moisture of agricultural farms in order to achieve high yields. Anytime water is needed for smart irrigation, the smart pump turns ON to avoid dryness of the farms, which may lead to the death of crops on such soils. Furthermore, to avoid over flooding of farms, the smart pump turns OFF anytime the farms have optimum soil moisture. At any point the smart pump is ON or OFF, data is generated. Hence, it is very important to investigate and classify data generated by smart IOT devices when these devices are ON or OFF. In this paper, the soil moisture, temperature, humidity, and time are used as inputs into machine learning techniques such as logistic regression, random forest, and support vector machine for the classification of data generated when the smart irrigation device is either ON or OFF. Experimental results showed that the logistic regression achieved an accuracy of 71.76%, random forest achieved an accuracy of 99.98%, and support vector machine achieved an accuracy of 90.21%. We observe that the random forest has more potential to assist in determining smart irrigation conditions (moist or dry) in an optimized manner based on the high accuracy it achieved

Index Terms— IOT, Agriculture, Smart irrigation, Machine learning techniques.