An unsupervised regime-gated framework is proposed for intra-day trading under non-stationary conditions. The dataset, 181.7 MB in CSV format, was authored by Nasser Minaei and last updated on June 4, 2026. It likely contains results from a walk-forward K-means clustering layer used to identify market states and filter execution signals.
Use Cases
- Developing unsupervised market regime detection models based on the described walk-forward K-means clustering layer.
- Testing adaptive trading strategies that gate execution signals based on identified market states.
- Analyzing downside risk-adjusted performance and maximum drawdown across six asset classes as mentioned in the description.
- Benchmarking new intra-day trading frameworks against the described gated strategy and baseline.
Strengths
- Dataset size is 181.7 MB, indicating a medium-scale data volume.
- Released under a permissive CC-BY-4.0 license.
- Methodology is described, including the use of a walk-forward K-means clustering layer and testing across six asset classes.
Limitations
- Row count is unknown, which may limit suitability assessment.
- Column-level documentation is absent; field semantics must be inferred after download.
- Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
- Source
- Nasser Minaei via figshare
- Collection Method
- Likely generated from financial market data as part of research on an unsupervised trading framework.
- Freshness
- Last updated 2026-06-04 15:41:02; freshness should be verified.