Korean Corporate Performance Data with Financial and CEO Message Features, 2016-2023

A dataset of 1,271 companies listed on Korea's KOSPI and KOSDAQ markets from 2016 to 2023. It was created by Xiao Wang to develop an AI model for predicting corporate management performance. The data combines financial variables with strategic indicators derived from text mining CEO messages in sustainability reports.

Use Cases

• Training machine learning classifiers for corporate performance prediction based on combined financial and textual data.
• Comparing model performance (e.g., Transformer vs. LSTM) on a structured financial-text hybrid dataset.
• Analyzing the relationship between CEO communication themes in sustainability reports and corporate outcomes.
• Benchmarking the predictive value of Sustainable Balanced Scorecard (SBSC) framework indicators against traditional financial metrics.

Strengths

• Contains data for 1,271 distinct companies over an 8-year period (2016-2023).
• Model performance metrics are provided, with the best hybrid model achieving Accuracy=0.8467, AUC=0.8481, and F1=0.8572.
• The dataset demonstrates a measurable performance improvement from adding strategic variables (ΔAccuracy=+0.0121).

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The 9.5 KB file size suggests a very limited scope, likely containing aggregated or summary data rather than raw records.

Provenance

Source

figshare, author Xiao Wang.

Collection Method

Financial variables selected from prior accounting research; strategic variables derived via text mining of CEO messages from sustainability reports.

Time Range

2016-2023

Freshness

Last updated 2026-05-06 17:32:43; freshness should be verified.

Geography

South Korea (KOSPI and KOSDAQ markets)