Machine Learning on Simulated and Real Farm Data Based on an Ontology-Controlled Data Infrastructure
Conference: International Symposium on "New-Dimensions in Agrometeorology for Sustainable Agriculture" At: Pnatnagar, Uttrakhand, India, 2014
Abstract: Rice is one of the most important food crops of India in term of area, production and consumer preference. Weather variability has a significant impact on crop growth and development. Timely and accurate crop yield forecasts are essential for decision making and managing the risk associated with these activities. Weather variable includes maximum and minimum air temperature, solar radiation, relative humidity and rainfall. In this study two different approach CERES-Rice model and Multiple Linear Regressions analysis were used for predicting rice yield for West Medinipur district of West Bengal, India.
Conference: International Symposium on "New-Dimensions in Agrometeorology for Sustainable Agricultural" At: Pantnagar, Uttarakhand, India
Summary: Low soil moisture during the growing period of peanut crop can reduce crop growth and decrease pod yield and dry matter production. Deficiency in soil moisture reduces the relative turgidity and eventually inhibits the leaf expansion and stem elongation (Jain et al., 1997). Irrigation based on crop water requirement i.e. reference evapotranspiration (ET 0) is more appropriate to increase yield and water use efficiency (Utset et al., 2004).
Summary: The increasing population and food demand have generated a great challenge for food security in developing nations like India. Therefore, the knowledge of yield distribution and forecasting are important for governmental planning and policy decisions. Rice is one of the most important food crops of India and is feeding more than 3 billion people. Presently due to the uncertainty of the major weather parameters, farmers are facing economic losses in rice cultivation. Therefore, yield forecasting under weather variability condition can be a useful to minimize the uncertainties associated with various decision makings. The objective of this study was to predict the yield for rainfed rice for West Medinipur district of West Bengal state using the DSSAT CERES-Rice v4.5 model. The 30 years (1983-2013) crop yield and weather data for of this region were collected from the India Meteorological Department, Pune. Out of the 30 years simulated values, yield data of 26 year (1983-2009) was used for calibration of the linear regreesing model and 4 years (2010-2013) data was used for validation.