The climate, topography, soil and biology are highly dependent on agriculture. In which land plays a significant role. Despite considerable progress in the service sector, agriculture remains India's primary source of employment and income, and price fluctuations in agricultural commodities have a significant impact on people's daily lives as well as agricultural inputs and outputs. In order to overcome all this, soil test values are used in the current study to classify several important soil characteristics, such as the Available Phosphorus (P), Potassium (K), Organic Carbon (OC) and Boron (B) village-wise soil fertility indices, as well as the Soil Reaction parameter (pH). We have addressed various algorithms related to data mining and machine learning (ML) classification techniques in this paper along with IOT hardware that used in agriculture. These algorithms are developed on a data set for yield prediction of crops that have been collected over the years. In addition, a comparative study is carried out to show which classification algorithm is better suited for classification techniques success prediction.

—machine learning classification, artificial neural network, soil analysis