A Novel Method for Prediction of Crop Yield Using Deep Neural Networks

Abstract

The study of data mining is essential to obtain the significant information by exploiting the insightful information from the enormous data sources. Being a key industry in any country, the agriculture has direct impact on Gross Domestic Product (GDP). Financial production and market prices are the key factors in generating revenue in agriculture. Higher yields results in more money and profit. Whereas, the lower yields can cause the fall in agricultural GDP. As a result, the yield monitoring becomes crucial in boosting the countrys’ economic resources. Early yield prediction models are based on manual computations, which occasionally had errors because of possible incorrect inputs. To address these issues, this paper provides a novel method by employing representative neural network models to evaluate the precision of yield predictions. This paper specifically investigate the use of hybrid and deep learning models, such as long short term memory (LSTM), convolutional neural network (CNN), and recurrent neural network (RNN) to choose the most precise model and evaluate yield prediction accuracy. Several tests are run by utilizing the farm production datasets collected from kaggle to assess the performance of these algorithms. The proposed method evaluates the predictive power of each model through these tests, providing insightful information for future projections of agricultural productivity. These findings have the potential to significantly increase the precision and dependability of yield projections, which would be advantageous for the agricultural industry and to boost the country's economy.