Hybrid Classification Model with Tuned Weights for Crop Yield Prediction

• Published in: Wireless personal communications 2023-11, Vol.133 (2), p.1325-1347
• Main Authors: Kolipaka, Venkata Rama Rao, Namburu, Anupama
• Format: Article
• Language: English
• Subjects: Adaptive algorithms; Agricultural production; Classifiers; Communications Engineering; Computer Communication Networks; Crop production; Crop yield; Engineering; Feature extraction; Networks; Performance evaluation; Prediction models; Predictions; Preprocessing; Signal,Image and Speech Processing
• ISSN: 0929-6212; 1572-834X
• DOI: 10.1007/s11277-023-10781-x

One of the difficult problems in agriculture is predicting the crop production. At the international, regional, and crop level, it is crucial to make decisions on this. In Most of the cases, agricultural, land, climatic, atmospheric, and other characteristics are used to forecast crop production. ML is a crucial decision-support model for estimating agricultural yields, enabling choices about which crops to cultivate and what to do while they are in the growing season. Numerous ML and DL algorithms have been applied to support studies on agricultural yield prediction. In this paper, a new crop yield prediction model is proposed which includes preprocessing, feature extraction and yield prediction phase. In preprocessing, data cleaning will takes place. Higher order statistical feature, information gain and improved entropy based features are extracted in feature extraction phase. The prediction is done by the hybrid model that combines Bi-GRU model and Maxout classifiers. To enhance the performance of this hybrid classifier, a new Self Adaptive Archimedes Optimization Algorithm (SAAOA) is introduced for training the weight parameters optimally. Finally the overall performance is evaluated and the better result is determined.