Maritime forecasting using machine learning and deep learning

The project contributes to the statistical, machine learning and deep learning fields from different aspects. First, we implement, document, and compare statistical, machine learning and deep learning solution for maritime market forecasting. This is the first attempt to utilize machine learning for this purpose. Moreover, we will present an enhanced explainable method that describes the contribution of the different features in forecasting and also provides explanations for different results. Furthermore, we will design an interpretable and explainable platform that can justify the forecasting results and convince clients to choose Soshianest for their consultation. This goal can approach through calculation and the global set of selected features that will be determined by ranking the features that are generally more important and also using post-hoc models such as LIME and SHAP.

Faculty Supervisor:

Fayez Gebali

Student:

Partner:

Soshianest Enterprise Miner Inc

Discipline:

Computer science

Sector:

Professional, scientific and technical services

University:

University of Victoria

Program: