A Deep Learning and Machine Learning Approach to Crop Recommendation for Sustainable Farming in Bangladesh

The agricultural sector in Bangladesh contributes approximately 11 to 12% of the country’s GDP, with the total B2C market size for agriculture-related products estimated at ${\$}$47.54 billion. This study focuses on the use of machine learning algorithms to forecast accurate crop recommendations based on environmental and soil characteristics, such as nitrogen, phosphorus, potassium, temperature, humidity, pH value, and rainfall. By utilizing standard machine learning tools, particularly deep learning, we have achieved an impressive $99.55 \%$ accuracy in determining the most suitable crops to grow under specific conditions. This high accuracy significantly aids farmers in making informed decisions regarding the optimal crops to cultivate, both under current conditions and in the anticipated conditions of the near future. Moreover, this technique enhances crop yield and production, contributing to improved farming practices tailored to weather patterns and soil types.