Forest Change Detection and Prediction using the Crow Chicken Swarm Optimization based Deep LSTM Classifier

Abstract

The detection of the changes in the earth surface is an important factor, as it enables the understanding regarding the interaction and relationship between the natural and human phenomena for achieving better decisionmaking strategy. Due to the socio-economic factor, there exist frequent changes in the forest area. Various methods are introduced to predict the change detection in the forest area, but to accurately find the change prediction in the forest region is a challenging task in the research community. Therefore, an effective method named Crow-Chicken Swarm Optimization algorithm based Deep Long Short term Memory (C-CSO based Deep LSTM) is proposed to find the forest change detection and prediction. The proposed algorithm uses the intelligent behaviour of crows with the hierarchical order and the mimicking behaviour of chicken groups to enhance the effectiveness of prediction in forest area. The time-series data is generated from the result of change prediction in order to increase the performance of the Deep LSTM classifier. The proposed C-CSO based Deep LSTM attained better performance in terms of the metrics, like Mean Square Error (MSE), and accuracy with the values of 0.0014 and 81.200%, respectively.