Agriculture is the main occupation of India and more than 50% of people are dependent on agriculture. Research on agriculture will strengthen the economic growth of the country. Technologies play a vital role to bolster the agriculture. Since soil is the main fount of agriculture , there is a need for significant approach to help the farmer to test and monitor the soil and its properties ,which will boost the fertility of the soil thereby intensifying the crop growth, also if crop recommendations are imparted to farmers in a proper way, crop yield can be enhanced to meet the growing demand for the food. Proper awareness on soil will benefit the farmers to grow the right and healthy crop. To overcome the disadvantages of traditional soil testing practices we are proposing an approach which has Deep learning, an artificial intelligence(AI) technique and IOT features . This helps in getting fast and accurate result. Soil fertility can be calculated by parameters like pH level, temperature, Moisture content of the soil,temperature, humidity and NPK(nitrogen, phosphorus, and potassium) ,organic matter, carbon level. Weather and Climatic conditions along with the soil parameters will help to evaluate the soil fertility. The lacking nutrients in the soil and needed nutrients/fertilizers to boost the soil fertility can be suggested to the farmers and also the crops which can be suitably grown from the given soil sample and nutrients required for all the recommended crops to enhance the yield can be suggested to the farmers.

Agriculture is the major occupation in India. The development of India is in the hands of farmers. Farmers are said to be our nation’s backbone, so there is a need to support our farmers technologically so that the difficulties of traditional agricultural practices would be overcome and also there will be positive impact on the yield, harvest, healthy crop output and the income of the farmers. Farmer needs awareness about his soil and the methods to improve his soil to grow the healthy crops. We propose an approach which involves deep learning and some IOT features to help our farmers. Soil parameters such as nitrogen, phosphorous, potassium (NPK), pH, organic carbon, moisture content and few more things are considered for predicting the fertility of the soil and also to predict the right crops to be grown and nutrition required for it. We have developed a deep neural network model to predict the crop which can be suitably grown in the soil. We have also implemented the other machine learning classifiers on the same collected dataset to test the accuracies of each classifier and our deep neural network model.

Agriculture is the <span>largest workforce of India and biggest contributor to the Indian economy. Improving agricultural practices with the help of modern computer science technologies have great scope. Helping the farmers to know about their soil fertility, crops which can be grown and fertilizers or nutrients required for their land will be valuable inputs for them. Too much or too little fertilizers may harm the soil, so right amount of fertilization is also important. In this paper we have discussed about the bootstrap aggregation regression method, which is an ensemble machine learning technique to recommend the optimum level of nutrients and fertilizers. Hence customized nutrients recommendation reports could be generated to suggest the fertilizers and nutrients with their adequate quantities. This will be really beneficial for farmers to maintain the soil health and helpful for better crop growth and yield. We consider the features and levels of soil parameters such as nitrogen, phosphorus, potassium (NPK), pH level, organic carbon, electric conductivity, humidity, rainfall and other micro nutrients for predicting the right amount of fertilizers and nutrients. We have also checked other regression methods to compare the results based on the previous work done in the same field.</span>