SOIL SALINITY DEGRADATION ESTIMATION BY REGRESSION ALGORITHM USING AGRICULTURAL INTERNET OF THINGS

SOIL SALINITY DEGRADATION ESTIMATION BY REGRESSION ALGORITHM USING AGRICULTURAL INTERNET OF THINGS Salinization of the soil has an effect on agricultural development and nutrition protection. Soil salinity is a tendency of soil depletion which has a significant effect on agricultural manufacturing. Agricultural Internet of Things (AloT) endorsed approach is recommended to assess the degree of soil salinity and environmental factors to prescribe water resources, with the objective of draining salts from the base region of grains in soil salinity. The posterior processing enables measurement of several soil and crop characteristics in a depth manner. These comprise compact X-ray, spectroscopy, digital camera, tablet, checking for multistripe laser triangulation, ground-penetrating radar and scanner for electromagnetism. The Agricultural Internet of Things (AloT) and Machine Learning (ML) are focused on the approximation of the desorption water criteria for saline soils leveraging in-situ analysis of the salt concentration and the temperature in the agricultural land. Sensor devices for water, soil and crop leverage modern technology to promote agricultural production, allow agricultural customers, reduce and conserve input production costs, manage natural resource smartly, and enhance income and competitiveness. This proposal approaches the regression algorithm called Random forest regression (RFR) algorithms which is implemented to the consolidated datasets. RFR is developed by an aggregate of increasing decision trees based on random vectors and needs to start with various bootstrap datasets which are derived arbitrarily from the initial training sample. A primary mechanism in RFR is to employ Bagging in tandem with random extraction of characteristics, since Bagging can significantly mitigate the variance of unpredictable processes including tree formation, resulting in enhanced forecasting and increased accuracy. The proposal demonstrates that the random forest regression model is reliable for estimating degradation of the soil salinity and predicts the salinity in soil with high accuracy. 1 P a g e