Machine Learning based Mobile Applications for Autonomous Fertilizer Suggestion

To ensure food security, advanced agriculture and the Green Movement haveincreased overall output, but with a yield gap. However, the use of agrochemicals such as fertilizers and pesticides, as well as modern cultivars, irrigation, and other inputs, continues to suffer agriculture in some locations. The agriculture-based on traditional agrochemicals may result in soil and environmental degradation, acidification, and a reduction in soil fertility. Only by utilizing appropriate fertilizers and limiting the number of soil resources that are exploited, can help to maintain soil quality and crop yield. As a result, fertilizer recommendations in agriculture are crucial to meeting productivity, and financial needs, and minimizing nutrient losses to the environment. In this research, the smartphone app will be used to construct an autonomous fertilizer suggestion system for farmers. The dataset was gathered from the website Kaggle. The encoding technique is done to make the data entries numeric. Three Machine Learning (ML) models like K-Nearest Neighbour (KNN), Random Forest (RF), and Decision Tree (DT) are employed to predict fertilizer based on environmental and soil characteristics. The best model is identified using performance metrics and deployed on the mobile app. From the performance analysis, the RF gives high accuracy of 90
^%
.