Abstract:
Pollution carbon dioxide also known as climate change is a menace to the global en vironment, and the transport sector plays a leading role in the growth of a developing country such as Pakistan. The conventional statistical approaches usually cannot fulfill the non linear interactions between environmental outcomes and socio economic drivers, which are complex. In this project, the forecasting system is developed based on data, which uses hybrid ensemble architecture. The model, using a weighted average ensemble strategy to train Feed Forward Neural Networks (FFNN), Adaptive Neuro Fuzzy Inference Systems (ANFIS), and Long Short Term Memory (LSTM) networks, to model past data (1970–2022), demonstrates high predictive performance (R 2 = 0.961, RMSE = 3.67 Mt CO2). The system in its implemented form is a modular web platform that allows policy makers to visualize their trends, as well as conducting comparative analysis of scenarios between a business as usual and an intervention approach giving a decision support tool that is sustainable in terms of transport planning
