Abstract:
The growth of agriculture industry is depending on the available resources and the way of using
those resources. It is very important to use available resources effectively to meet the maximum
productivity and fulfil the demand for agricultural products that increasing with the growth of
population. Farm land conditions such as soil parameters and environmental conditions are the
most important parameters for growth of the crops. Using of traditional agriculture practices
are not the best way to make decisions on farm land conditions. It is not easy and effective
to use fertilizes, Water and chemicals without knowing real conditions. This can be a reason
to waste fertilizer, money, water, time and to reduce the harvest. It is better to select the most
suitable crop according to the current farm land conditions or change current farm conditions to
reach optimum conditions of selected crop. The main objective of this research was to design
and implement a device to measure the farm land conditions in highest accuracy using IOT device and present those data to farmers through a mobile application. Further, propose the most
accurate classification algorithm through Naive Bayes, Bayes Net, Random Forest and Random
Trees algorithms to predict the most suitable crop for current farmland conditions. This device
facilities farmers to stay touched with their farm land from anywhere in the world and this is a
step to improve smart agriculture concepts using IOT technology. And also, this device facilitates to understand patterns of environmental conditions using past data through web interface.
This system is able to send notifications to farmers whether any parameter is overdue or not its
optimum conditions. As the result of this research it is found that the random forest classification algorithm is the most accurate algorithm from selected classification algorithms in crop
prediction
