ABSTRACT
Nowadays, the continuous generation of data from multiple sources has created numerous opportunities in different domains such as intelligent transport systems, smart environments, smart health, and smart agriculture. However, the state of the art in smart agriculture should be redefined and revisited since new technologies and tools are coming into the agriculture domain bringing novel and innovative paths for improving the resilience and the efficiency of agriculture. The objective is intended to increase the productivity of crops and to reduce the number of workers in the farm field by analyzing the sensed value and to give accurate decisions automatically to farmers using the Internet of Things and Big Data. Also, to make profitable decisions, farmers need guidance throughout the entire farming cycle at the right time. The information required by the farmers is scattered in various places including real-time information. Various sensors are deployed at different locations on the farm to measure real-time data. Values generated continuously from various sensors at multiple locations over a period will lead to Big Data. These need to be analyzed using Machine Learning techniques and compared with a knowledge base. The agricultural data collected from various heterogeneous resources over several years including agricultural researchers, email, web crawling, and satellite data are discovered and stored in a knowledge base for predictive analysis. In order to synthesize Big Data and to communicate among devices using IoT Machine Learning techniques are employed. Various Machine Learning techniques include regression analysis, clustering, bayesian methods, decision trees and random forests, support vector machines, reinforcement learning, and deep learning. Machine Learning techniques serve as a basis for predictive analysis. IoT has the capability to provide information about crop yields and rainfall to farmers.
