An Adaptive Graph Based Feature Selection And Deep Learning Classification Framework For Rice Disease Prediction

Document Type : Primary Research paper


1 Assistant Professor of Computer Science, Gobi Arts & Science College, Gobichettipalayam.

2 Assistant Professor of Computer Science, Gobi Arts & Science College,Gobichettipalayam.


The volume of information in every application necessitates the use of deep learning approaches for data analysis. Designing a deep learning-based classifier for decision-making and, as a result, upgrading the whole system's knowledge base is one such data investigation task in a dynamic context. The building of a classifier represents the retrieval of interesting patterns from a huge database of data and the prediction of future trends based on those patterns. The classification system's time consumption rises with time, and the system becomes inefficient as it is continually learnt for adding new groups of data to the current ones. If the knowledge of previous data collected by the classifier is used with the new group of data to construct the updated classifier, it may be done without learning the same classifier for all of the data. In the paper, for the selection of significant features, the principles of graph based adaptive feature selection approach are used. Because the features of rice illnesses fluctuate over time owing to changes in climatic, biological, and geographical variables, the deep learning-based classifier is ideal for use on the rice disease dataset for forecasting of disease. The suggested technique has been tested on synthetic rice disease datasets as well as benchmark datasets, and the classification accuracy has been assessed and compared to other state-of-the-art classification methods. The approach is also assessed based on the algorithm's performance in order to determine its importance and efficacy.


Volume 12, Issue 3 - Serial Number 3
ICMMNT-2021 International Virtual Conference on Materials, Manufacturing and Nanotechnology, 30th June, 2021.
June 2021
Pages 3058-3067