Forecast chart and methods
View the forecast, baseline, and evaluation approach.
Explore reportWhat are we trying to understand?
How well does an 11-lag autoregressive model forecast the final 20 years of the sunspot dataset?
What the analysis found
Replace this text with the model and baseline mean absolute errors printed by experiment.py, and explain which performed better.
Interactive reports
Open a report to explore its charts and tables. Reports open as full pages for the best reading experience.
How the work was done
The script loads the sunspots dataset from statsmodels, fits an autoregressive model with 11 lags, and evaluates it on the final 20 years against a baseline that predicts the training-data mean.
Limitations
This evaluates one historical holdout period. It does not establish how well the model predicts future activity; describe any other important limitations of the data or approach.
Sources behind the work
- Sunspots datasetView source
statsmodels · Historical annual sunspot observations included in the statsmodels dataset
The experiment loads this dataset through the statsmodels API.
Accessed 2026-10-31 · License: Verify the dataset's reuse terms before publishing.