Central Statistics Office
Machine-learning location classification of RIP.ie death notices
Ireland's CSO replaced manual death-notice location classification with machine learning
Problem
The CSO used online death notices to monitor mortality before official registrations arrived. A team had read each notice to classify whether the death occurred in the Republic of Ireland, Northern Ireland or elsewhere.
Approach
A CSO statistician developed classification models trained on earlier manual labels. The office applied the models to all RIP.ie data in its February 2022 release, revising some historical figures for consistency. The method also classified newly collected notices.
Results
No published figures
Takeaway
Earlier manual classification supplied training data for the replacement process. Applying the model retrospectively changed the historical statistical series.
Categories
- Industry
- Public sector
- Business function
- Internal work and knowledge management
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
The CSO described replacing manual location classification with models trained on previous labels. It applied the method across the release's death-notice dataset but did not publish classifier accuracy or staff-time savings.
Full details are in the original sources.