CRDI Corporate Fraud Multimodal Dataset with 276 Cases

276 cases of corporate fraud data integrating geospatial, vocal-stress, and linguistic features. The dataset is hosted on Kaggle, but the author, organization, and creation date are unknown. Its specific collection methodology and temporal coverage are not detailed in the provided metadata.

Use Cases

• Detecting deceptive communication patterns based on linguistic features mentioned in the description
• Analyzing speaker stress levels for fraud risk assessment based on vocal-stress features
• Investigating geographic patterns of fraudulent activity based on geospatial features
• Training multimodal classifiers that combine speech, text, and location signals for fraud prediction

Strengths

• Integrates three distinct data modalities (geospatial, vocal-stress, linguistic) for a holistic view
• Contains 276 documented cases of corporate fraud

Limitations

• Column-level documentation is absent; field semantics must be inferred after download
• Row count is unknown, which may limit suitability assessment
• Last update date is unknown; freshness unverified

Provenance

Source

Kaggle