Ablation Study Results for Adaptive Ensemble Tourist Flow Forecasting in Aosta Valley

An adaptive ensemble AI framework for tourist flow forecasting, validated on over 41 million vehicle passages from 14 sensor portals in Italy's Aosta Valley. The study by Marco Alderighi, last updated in May 2026, demonstrates a 23.7% MAE improvement and 31.2% MSE reduction compared to individual models, with R² scores exceeding 0.94 for short-term predictions.

Use Cases

• Benchmarking adaptive ensemble weighting mechanisms based on the described ATE algorithm.
• Analyzing the impact of meteorological conditions on tourist traffic based on integrated weather data.
• Developing short-term traffic forecasting models based on the reported high R² scores for predictions.
• Studying seasonal pattern robustness in forecasting systems based on the framework's validation across seasons.

Strengths

• Validated on a large-scale dataset of over 41 million vehicle passages.
• Demonstrates concrete performance improvements: 23.7% MAE reduction and 31.2% MSE reduction.
• Achieves high predictive accuracy with reported R² scores exceeding 0.94.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The dataset is very small at 5.5 KB, suggesting it contains summary results rather than raw sensor data.

Provenance

Source

Marco Alderighi via figshare.

Collection Method

Data collected from 14 strategically positioned sensor portals monitoring highway and valley roads.

Time Range

null

Freshness

Last updated 2026-05-06 17:37:51; freshness should be verified.

Geography

Aosta Valley region, Italy.