An Efficient Filter Based Extreme Learning Machine for Agriculture Yield Prediction

Abstract

The agriculture sector holds immense significance in India's economy, acting as the primary income source for a significant portion of the population. To improve agricultural practices, increase productivity, and enhance efficiency, farmers can integrate scientific tools such as remote sensing and predictive analytics. An emerging trend in this domain is the adoption of the Internet of Things (IoT). Leveraging IoT, the Agriculture Monitoring Model (AMM) is proposed as a solution to continuously monitor and comprehend the soil and environmental conditions in agricultural fields.By deploying remote sensors, AMM captures crucial environmental parameters including water levels, rainfall, temperature, humidity, soil nutrients, and soil pH. These parameters exhibit variations across diverse locations and significantly impact crop yield. Through continuous monitoring, AMM effectively mitigates potential environmental risks, provides real-time updates on plant health, and aids in optimizing fertilizer management and pesticide application. This technological advancement is vital for ensuring the prosperity of the national agricultural sector.In India, farmers often lack awareness and knowledge regarding technological approaches to optimize sowing patterns, irrigation schedules, and fertilizer application. Consequently, achieving desired crop yields becomes a challenge. In this work, a novel filter based classification framework is developed on real-time agriculture data for better decision making. Experimental results show that the proposed model has better improvement over the conventional models.