FertiForecast: Machine Learning-Based Fertilizer Quality Assessment Using NPK Levels

Effective fertilizer management is crucial for optimizing agricultural productivity and sustainability. This study presents FertiForecast, a novel system for identifying fertilizer quality based on Nitrogen (N), Phosphorus (P), and Potassium (K) levels using a combination of Arduino technology and machine learning algorithms. The system integrates an Arduino microcontroller with an NPK sensor to measure the nutrient levels in fertilizers. The real-time data collected by the sensor is displayed on an LCD screen for immediate visualization and is concurrently transmitted via serial communication to a machine learning model for analysis.The machine learning component employs both Random Forest and Linear Regression algorithms to classify the fertilizer as either "good" or "bad" based on its NPK values. The Random Forest model, a robust ensemble learning technique, is used for its ability to handle complex, non-linear relationships and provide high classification accuracy. In contrast, the Linear Regression model offers a simpler approach, useful for understanding linear correlations between NPK values and fertilizer quality.The system aims to provide a practical tool for farmers and agricultural professionals by automating the assessment of fertilizer quality, thereby facilitating better decision-making and resource management. The integration of real-time data acquisition with advanced machine learning techniques represents a significant step towards smarter and more efficient agricultural practices.

Keywords: Arduino, NPK sensor, machine learning, Random Forest, Linear Regression, fertilizer quality assessment, real-time data analysis.