Strategies for land use and ecological restoration around highways under improved neural networks

The unchecked development of highway is causing extensive ecological disturbance such as habitat fragmentation, landscapes degradation and biodiversity loss. To meet the above challenges, this study puts forward an improved artificial neural Network (Artificial Neural Network (ANN)) model developed and calibrated by using Krill Herd Algorithm (KHA) to predict and control land use and ecological rehabilitation along highway corridors. The analysis utilizes co-incident spatial, environmental and infrastructure data to assess the interplay of road transportation infrastructure with local ecosystem. The Artificial Neural Network (ANN)-KHA model was built and compared with four models from literature, namely baseline Artificial Neural Network (ANN), Random Forest (RF), Support Vector Machine (SVM) and Artificial Neural Network Cellular Automata based ANN (ANN-CA). Analysis of the data and evaluation of the model was conducted with accuracy, precision, recall, f1-score, AUC measures. The experimental results show that the proposed Artificial Neural Network (ANN)-KHA model is more predictive and robust compared to both conventional and hybrid models. The model is good for identification of ecologically sensitive zones and restoration priority areas that can be incorporated in data driven based strategy design towards sustainable highway construction and ecological conservation.