• Homa Rizvi
• Shamim Ahmad
• Dr. Yusuf Perwej
• Farheen Siddiqui
• Dr. Nikhat Akhtar

Agriculture is a vital profession globally, reliant on climatic conditions and precipitation. The goal of this article is to use climate, soil, and temperature data to estimate crop output early. This study proposes a classification-based approach for agricultural production prediction using Long Short-Term Memory (LSTM) with an Attention Mechanism. The Economics and Statistics department of the Government of Karnataka collects the manual data. This technique used data from the Department of Economics and Statistics on three crops: jowar, rice, and ragi. To fill in the missing and null values in the dataset, the linear interpolation approach is used. The feature selection procedure is useful for the Correlation based Feature Selection Algorithm (CBFA) and the Variance Inflation Factor Algorithm (VIF) because it helps them choose and remove groups of features that are linked. We use Accuracy, R2, Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Square Error (RMSE), and Mean Absolute Percentage Error (MAPE) to see how well the model works. The proposed LSTM model gives results with assessment metrics including accuracy, R2, MAE, MSE, and RMSE values of roughly 99.10%, 0.44, 0.132, and 0.233, respectively.

Agriculture Data, Feature Selection, Long Short-Term Memory, Crop Selection, Pre-Process, Crop Yield Prediction.

"Crop Yield Forecasting Through an Attention-Driven LSTM Deep Learning Mode", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 11, page no.e891-e898, November-2025, Available :http://www.jetir.org/papers/JETIR2511510.pdf

"Crop Yield Forecasting Through an Attention-Driven LSTM Deep Learning Mode", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 11, page no. ppe891-e898, November-2025, Available at : http://www.jetir.org/papers/JETIR2511510.pdf