Bank Marketing Dataset for Imbalanced Customer Conversion Forecasting

A dataset used for forecasting customer conversion in bank marketing, which is challenged by imbalanced class distributions. The dataset, shared by Fahim Nasir on figshare and last updated in May 2026, was used to evaluate ensemble and deep learning models under five sampling strategies, including Borderline2SMOTE, to quantify trade-offs between performance and synthetic distortion.

Use Cases

• Evaluating sampling techniques for imbalanced classification based on the described use of Borderline2SMOTE and other strategies
• Benchmarking model performance across discrimination, calibration, and computational cost as per the quadrilateral evaluation framework
• Applying SHAP for model interpretability to identify key drivers of customer responses mentioned in the description

Strengths

• Dataset is shared under a permissive CC-BY-4.0 license
• The associated research proposes a specific evaluation framework assessing models across discrimination, calibration, cost, and explainability

Limitations

• Dataset size is only 9.5 KB, indicating a very limited scope
• Row count and column-level documentation are unknown, which limits suitability assessment
• Description metadata is limited; actual data quality and features require manual inspection after download

Provenance

Source

figshare

Freshness

Last updated 2026-05-15 17:43:45