This project focuses on optimizing operations within a business by assessing the demand, planning capacity accordingly, and using forecasting models to predict future needs. The problem statement focussed on precise capacity planning and accurately forecasting demand.
The objective of the project is to develop a python-based demand forecasting model that integrates harmonic analysis, linear regression and time series feature engineering to highlight complex demand patterns. With a strong focus on forecast accuracy, the project is working towards reducing operational inefficiencies and supporting strategic air network planning.
The key deliverables for this project included developing a python-based forecasting model and developing comprehensive documentation. The progress of this includes a comprehensive review of state-of-the-art time series forecasting methods, Toffee-tree, which is an automatic feature engineering with 286+ trend cycles and BOFA (Bi-objective optimization with fourier-based linear regression for multi-seasonal data).
The project has completed the forecasting of international shipping, engineered features capturing temporal trends and seasonality. A Harmonic Regression Model has been built, the students used MAE, RMSE and MAPE for model performance evaluation.
Contributing Researchers:
Jaswanth (BTech)
Vettri Velavan J (PhD)
