How ANFIS AI is Transforming Agriculture; AI Predicts Crop Yields with Precision

India’s agricultural sector is a key part of its economy. It supports millions of livelihoods and makes the country the second-largest agricultural producer in the world. A new study from Sardar Vallabhbhai National Institute of Technology (SVNIT) in Surat uses Artificial Intelligence (AI) to predict crop yields with impressive accuracy. This offers farmers a valuable tool. The researchers employed the Adaptive Neuro-Fuzzy Inference System (ANFIS) to forecast yields for five major crops: rice, maize, corn, groundnut, and sugarcane. This research took place in Nashik, Maharashtra, a semi-arid region known as India’s wine capital. (Representative Image Credit - rattanakun/Canva)

How ANFIS Works; Why This Matters

ANFIS combines two effective AI techniques: neural networks and fuzzy logic. Neural networks are good at finding complex patterns in data. Fuzzy logic reflects human reasoning using "if-then" rules, such as "if rainfall is low and temperature is high, then crop yield may be reduced." This blended method helps ANFIS explore the complex relationships between different factors. In the Nashik study, the model concentrated on five climate factors that are essential for crop growth: average rainfall, minimum and maximum temperatures, relative humidity, and evaporation rates.

To develop the ANFIS model, researchers collected climate and yield data from Nashik from 1987 to 2020. They found that 33 years of data wasn’t enough for effective AI training. To fix this, they used mathematical techniques to create synthetic datasets, which helped improve the model’s accuracy. They divided the data, setting aside 70 % for training and 30 % for testing the model’s performance. The results were impressive, especially for sugarcane. However, the predictions for maize were less accurate, likely because it depends less on the climate factors they studied.

Traditionally, farmers and agronomists estimate yields based on experience, manually assessing soil, seeds, water, and weather conditions and is a process that is time-intensive and prone to errors. AI-driven predictions like ANFIS offer a faster, more reliable alternative. By forecasting yields before planting or early in the growth cycle, farmers can optimize water, fertilizer, and other resources, select the most suitable crops, and boost profitability. This is particularly vital in semi-arid regions like Nashik, where climate variability poses significant challenges.

Challenges and Future Potential
Despite its promise, the study has limitations. The model relied only on climate data and left out important factors like soil type, fertilizer use, irrigation methods, and pest impacts. Including these could improve accuracy but might make the calculations more complex, which could slow down the training process. Overfitting, where the model becomes too tailored to the training data, also needs careful data validation to prevent issues. Future research could add more variables and expand the model to other regions. It might also integrate into user-friendly tools like mobile apps for farmers. This study shows how AI can help tackle agricultural issues, especially in regions vulnerable to climate change. By using data synthesis to address limited historical data, the researchers have opened the door for new solutions to improve food security. As climate change becomes more severe, these technologies could help farmers adapt and maintain sustainable production and economic stability.

Source Credit: This article is based on findings from the study published in Scientific Reports by Nature: “Crop yield prediction using adaptive neuro-fuzzy inference system with climate data”

Keywords: AI in agriculture, crop yield prediction, ANFIS, Indian agriculture, Nashik farming, precision farming, climate-based forecasting, artificial intelligence agriculture, sustainable farming, data synthesis, neural networks, fuzzy logic, crop productivity, Indian economy, agriculture technology