Developing an IoT-Based Data Analytics System for Predicting Soil Nutrient Degradation Level

Abstract

Globally, agriculture seems to be the economic field, which occupies a major part in India’s socio-economic structure. The parameters such as soil and rainfall play a major role in agriculture dependency. Farmers will usually have the mindset of planting the same crop by using more fertilizers and following the public choice. In agriculture, crop productivity will be increased with the incorporation of new technologies. The most commonly used smart farming technologies such as Internet of Things (IoT) has the tendency to process the generous quantities of data from these devices. In the recent past, there have been major developments on the utilization of machine learning (ML) in various industries and research. For this reason, machine learning (ML) techniques are considered as the best choice for agriculture, which is then evaluated to predict crop production for the future year. In this paper, the proposed system uses IoT devices to gather information such as soil nutrient level, temperature of atmosphere, season of the atmosphere, soil type, fertilizer used and water pH level periodically. Further, the data gathered from the sensor will be passed to a principal component analysis (PCA), which are used to reduce features in order to obtain a better prediction level. Also, ML algorithms such as linear regression (LR), decision trees (DT) and random forest (RF) are implemented to forecast and classify the crop yield from the previous data based on soil nutrient degradation level and recommend suitable fertilizer for every particular crop.