Using AI to predict SDG goals and targets
Retshidisitswe Lehata , Carel Nchachi
Partner: sdg-hub
Year: 2020
Abstract:
In 2015, the Sustainable Development Goals (SGDs) were adopted by the all members of the United Nations (UN).This was a universal call to end all poverty, protect the planet and bring peace and prosperity for all people by 2030. South Africa’s National Development Plan (NDP 2030),which precedes the SDGs, shares similar aspirations as the UN SDGs (The NDP has a 74% convergence to the SDG). By utilizing the SDGs this gives South Africa an opportunity to accelerate sustainable development in the country, by creating and implementing efficient and effective policies.The South African SDG Hub seeks to address some of these barriers by collecting relevant open access research and making it available in accessible formats. Even though a great deal of progress has been made, the Hub’s SDG classification tool is in need of further refinement.The main objective of the project is to Investigate and Implement Supervised Learning methods to classify articles in terms of the 17 SDGs, create a model that can classify the 169 targets by using supervised learning methods and visualize the results (information) in a way that is easy for policymakers. The SVM and CNN produced the best results for the SDG goals and targets respectively.