Available tags:
- Document mode:
slide
,paper
,book
,article
- Document type:
research
,application
Ideas:
- feed this repo to an LLM to enable search and summarization
- Modeling: from basic to advanced concepts
- Non-Stationarity in Time-Series Analysis: Modeling Stochastic and Deterministic Trends: a didactic paper on how to deal with time series non-stationarity (and on the consequences of ignoring it)
- Principles and Algorithms for Forecasting Groups of Time Series: Locality and Globality
- Cross-validation
- Experiments on cross-validation (Samuele Mazzanti, inspired by Victor Cerqueira et al)
- Feature-based approaches
paper
Characteristic-based clustering for time series. Using feature-based approaches to cluster time series greatly reduces the dimensionality of your problemslide
Meta-learning on how to forecast time series. FFORMS: what type of models are most suitable for a given time serie?
- Metrics
paper
Forecast Evaluation for Data Scientists: Common Pitfalls and Best Practices. Hewamalage, Ackermann, Bergmeir
book
Demand forecasting best practices (N. Vandeput)slide
Retail sales forecasting at Walmart (Seaman). Predict for what? Addresses replenishment and pricing issuespaper
Forecasting with trees: why tree-based methods achieved excellent results in the M5 competitions? What are the directions to improve even further?paper
Learnings from (various) Kaggle forecasting competitions (Bojer and Meldgaard)
paper
Empirically validated probabilistic forecasts of energy technology costs. How to estimate stuff in a regime of technological uncertainty (due to learning effects)?
- Sarem Seitz. Featured: when point forecasts fail, autoregressive random forests, on probabilistic forecasts
- Marco Cerliani. Featured: forecasting with trees (hybrid modelling)
paper
retail
Managing Functional Biases in Organizational Forecasts: A Case Study of Consensus Forecasting in Supply Chain Planning
This section is not strictly related to time series but collects useful tips in Python, Model deployment and specific libraries
- Monash Time Series Forecasting Archive (arxiv)
- Modern strategies to time series regression (arxiv)
- Forecasting for social good (arxiv)
- FRANS automatic feature extraction via CNN (arxiv)
- "Our simplified HINT approach shifts away from multivariate inputs, leading to a 13% accuracy increase over existing solutions (such as MinTrace, PERMBU, HierE2E) across four hierarchical datasets" arxiv, nixtla tutorial
- How to win a forecasting competition? Youtube video
paper
Good and bad judgment in forecasting, lessons from four companiespaper
Can we obtain valid benchmarks from published surveys of forecast accuracy? (Kolassa)book
Supply Chain Strategy and Financial Metrics: The Supply Chain Triangle Of Service, Cost And Cash on amazonbook
The Business Forecasting Deal: Exposing Myths, Eliminating Bad Practices, Providing Practical Solutions on Amazon