Abstract:
Agriculture is a significant source of human survival and it accounts for the socio-economic growth in many developing countries including Sri Lanka. Paddy Cultivation occupies a remarkable place in Sri Lankan agricultural sector. Unpredictable climatic change has become a critical issue for paddy farmers while unawareness on pest, diseases, new technologies, etc. have also adversely affected Paddy Cultivation productivity. As a solution, the focus on the requirement of accurate weather predictions and timely access to the information for decision-making in Paddy Cultivation is highly progressive. This study introduces eKeth: a mobile platform that provides proper guidance for Sri Lankan paddy farmers through allowing timely access to data enhanced with machine learning. A weather prediction model based on machine learning has been developed to recommend the most suitable days for each farming task in paddy cultivation. The application includes several other features integrated with this machine learning model. Farmers can directly reach help from agriculture experts by posting a query on pest and disease-based issues. Fertilizer management feature allows calculating the amount of fertilizers upon different paddy types and growth stages. Buy and sell feature integrated with this mobile solution guide farmers on newly available machineries and the places where they can make purchases. Farmers can stay updated with the latest agriculture news though the news module while maintaining communications with other farmers and agriculture experts through the community forum empowered by this application. Machine Learning Model used in weather prediction achieved 89% accuracy for Random Forest. Statistical analysis of the user testing results recognizes that the system has been able to achieve a higher user satisfaction.
Citation:
J. S. A. N. W. Premachandra and P. P. N. V. Kumara, "eKeth: A Machine Learning-Based Mobile Platform to Facilitate the Paddy Cultivation Process in Sri Lanka," 2021 6th International Conference on Information Technology Research (ICITR), 2021, pp. 1-6, doi: 10.1109/ICITR54349.2021.9657468.