A formal representation of the WHO and UNICEF estimates of national immunization coverage: a computational logic approach

Abstract

Production of official statistics frequently requires expert judgement to evaluate and reconcile data of unknown and varying quality from multiple and potentially conflicting sources. Moreover, exceptional events may be difficult to incorporate in modelled estimates. Computational logic provides a methodology and tools for incorporating analyst's judgement, integrating multiple data sources and modelling methods, ensuring transparency and replicability, and making documentation computationally accessible. Representations using computational logic can be implemented in a variety of computer-based languages for automated production. Computational logic complements standard mathematical and statistical techniques and extends the flexibility of mathematical and statistical modelling. A basic overview of computational logic is presented and its application to official statistics is illustrated with the WHO & UNICEF estimates of national immunization coverage.

Figure 1. Inference triangle.

Figure 2. Processing levels.

Figure 3. Sample output.

Figure 4. Sample output: Lithuania.

Figure 5. Sample output: Tanzania.

Figure 6. Sample output: India.