Experimental Dataset Statistics: CSI 300 High-Frequency Pricing Data 2019-2024

Qi Fu's experimental dataset supports a study on a Transformer-based Efficiently-Fused Optimized Bayesian Network (Trans-EFOBN) for financial asset pricing. The dataset contains minute-level high-frequency data for CSI 300 constituent stocks from 2019 to 2024, sourced from the Wind database. It is used to demonstrate model performance gains, including a mean absolute error of 0.037 and an R² of 0.86.

Use Cases

• Training transformer-based models for sequential feature extraction in financial time series.
• Evaluating dynamic Bayesian network performance for hierarchical structural dependencies between macro and micro market variables.
• Benchmarking asset pricing models using metrics like MAE and R².
• Simulating trading scenarios to assess annualized returns and Sharpe ratios.

Strengths

• Model performance metrics are provided, including a mean absolute error of 0.037 and an R² of 0.86.
• The data covers a specific 5-year time range (2019-2024) for CSI 300 stocks.
• The study reports simulated trading results, including a 14.2% annualized return and a Sharpe ratio of 0.95.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is very small at 5.5 KB, indicating limited scope or summary statistics rather than raw data.

Provenance

Source

Wind database.

Collection Method

Empirical analysis utilizing minute-level high-frequency data.

Time Range

2019 to 2024.

Freshness

Last updated 2026-05 08 17:28:07; freshness should be verified.

Geography

China (CSI 300 index).