Research & Decision Science
Where forecasting research becomes operational decision infrastructure
RabbitHawk is built on peer-reviewed work in probabilistic forecasting, hierarchical reconciliation, optimization and applied AI, translated into systems that improve real decisions under real constraints.
- publications
- 1,181publications
- citations
- 53,485citations
- highly cited papers
- Top 1%highly cited papers
Scientific foundations
Methods proven in the literature, applied in production
Our team’s research is recognized at the field’s leading venues and used at scale by major organizations. We bring that rigor to enterprise planning, but judge every method by whether it improves decisions in practical settings.
Research → product
| Research area | Enterprise problem | RabbitHawk capability |
|---|---|---|
| Hierarchical forecasting | Local plans do not add up to executive targets | Reconciled forecasts across every level |
| Probabilistic forecasting | Point forecasts hide risk | Full predictive distributions, intervals and scenario ranges |
| Optimization | Forecasts do not tell teams what to do | Constraint-aware recommendations |
| Agentic AI | Context is trapped in emails and notes | Context ingestion and decision agents |
| Human-in-the-loop AI | Automation without trust creates resistance | Approval, override and audit trails |
| Alignment science | Teams optimize against conflicting goals | 90North alignment and goal clarity |
Hierarchical forecasting
Local plans do not add up to executive targets
Reconciled forecasts across every level
Probabilistic forecasting
Point forecasts hide risk
Full predictive distributions, intervals and scenario ranges
Optimization
Forecasts do not tell teams what to do
Constraint-aware recommendations
Agentic AI
Context is trapped in emails and notes
Context ingestion and decision agents
Human-in-the-loop AI
Automation without trust creates resistance
Approval, override and audit trails
Alignment science
Teams optimize against conflicting goals
90North alignment and goal clarity
Selected publications
The full peer-reviewed database: search by title, author, venue, category or year.
1,181 results · page 76 of 99
2005
Topic-specific scoring of documents with discrete PCA
W Buntine, K Valtonen
ICML 2005 Workshop 4, 34-41, 2005
Research1 citationsStructuring Documents Efficiently
R Marshall, S Bird, P Stuckey
Proceedings of the Australasian Language Technology Workshop 2005, 120-126, 2005
ResearchImproved Inference for Checking Annotations
PJ Stuckey, M Sulzmann, J Wazny
arXiv preprint cs/0507036, 2005
Bayesian MethodsFast Node Overlap Removal in Graph Layout Adjustment
T Dwyer, K Marriott, PJ Stuckey
Monash University Publishing, 2005
Research13 An MDL Framework for Data Clustering
P Myllymäki, W Buntine
Advances in Minimum Description Length: Theory and Applications, 323, 2005
ResearchProceedings of Learning in Web Search (LWS 2005)
S Bloedorn, WL Buntine, A Hotho
OptimizationProceedings of the Workshop on Learning in Web Search (LWS 2005), 7-11 August 2005 in Bonn, Germany
S Bloehdorn, W Buntine, A Hotho
Optimization
2004
The refined operational semantics of Constraint Handling Rules
GJ Duck, PJ Stuckey, MG De La Banda, C Holzbaur
International Conference on Logic Programming, 90-104, 2004
Machine Learning239 citationsSpeeding up constraint propagation
C Schulte, PJ Stuckey
International Conference on Principles and Practice of Constraint …, 2004
Machine Learning94 citationsImproving type error diagnosis
PJ Stuckey, M Sulzmann, J Wazny
Proceedings of the 2004 ACM SIGPLAN workshop on Haskell, 80-91, 2004
Research70 citationsSound and decidable type inference for functional dependencies
GJ Duck, S Peyton-Jones, PJ Stuckey, M Sulzmann
European Symposium on Programming, 49-63, 2004
Bayesian Methods54 citationsA scalable topic-based open source search engine
W Buntine, J Lofstrom, J Perkio, S Perttu, V Poroshin, T Silander, H Tirri, ...
IEEE/WIC/ACM International Conference on Web Intelligence (WI'04), 228-234, 2004
Optimization49 citations
Learn the science behind the engine
“Forecasting for Data Scientists”, a free 30-chapter video course by co-founder Dr Christoph Bergmeir, from fundamentals to advanced deep learning.