With rapidly growing interest in the use of machine learning for official statistics but with limited experience with concrete applications, there was a great need for a common platform where experts in national statistics offices to test their ideas, exchange experiences and collaborate on developments. National statistics offices work on similar type of problems and operate with similar business constraints, hence can benefit from developing shared understanding. To address this need, the High-Level Group on Modernisation of Official Statistics (HLG-MOS) launched the Machine Learing project in early 2019 with aims to:
- Investigate and demonstrate the value added of ML in the production of official statistics, where "value added" is increase in relevance, better overall quality or reduction in costs;
- Advance the capability of ML to add value to the production of official statistics;
- Advance the capability of national statistical organisations to use ML in the production of official statistics;
- Enhance collaboration between statistical organisations in the development and application of ML.
Following these objectives, the project team identified three main areas to advance the use of ML in statistical organisations:
- Work package 1 – Pilot Studies (demonstration of value added)
- Coding and Classification
- Edit and Imputation
- Work package 2 – Quality
- Work package 3 – Integration of ML into organisation
This webinar is the first public event where the outputs of the project will be communicated. This includes study reports, shared code and data, analysis of value added, recommended ML practices, quality framework elements and examples of organisational practices to address integration challenges.
The project will officially close at the end of the year (2020). Since it was launched in March 2019, the number of participants and other collaborators has grown from 20 to over 120. Given this strong interest, the project will evolve into a group to continue the advancement of ML in the production of official statistics.