首页|South Indian agricultural crop yield prediction using deep learning and transfer learning models
South Indian agricultural crop yield prediction using deep learning and transfer learning models
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
Springer Nature
Abstract Agriculture is an essential part of the Indian economy, so crop yield (CY) prediction is vital to help farmers and their businesses understand when to plant a crop and when to harvest based on seasons for better CY. This study proposes a deep learning model called Mish Activation-based Bidirectional Gated Recurrent Unit (MABGRU) to forecast CYs primarily grown throughout India. The system comprises the following steps: preprocessing, deep feature extraction, feature selection, and CY prediction. Initially, the system performs preprocessing steps such as imputation and normalization to handle the missing values and standardize the dataset. Then, deep features are learned using the Deformable Attention-based Residual Network-50 (DARN50). Then, the best features are chosen from the extracted feature set using the Tent chaotic map and phasor operator, including the Sparrow search algorithm (TPSSA). Finally, MABGRU is employed for CY prediction. The result of the system is compared with the existing systems, and our model outperforms the previous techniques by achieving a maximum accuracy of 98.88% with less computational time (15.9 ms).
NETL
NSTL
Springer Nature
