Cost-Sensitive Analysis: Optimal Classification Thresholds for TravisTorrent Builds

A 5.5 KB Excel file contains results from a cost-sensitive analysis using a Random Forest model on the TravisTorrent dataset. The analysis calculates optimal classification thresholds across varying false-negative to false-positive cost ratios. Author Lalit Narayan Mishra published the dataset on figshare under a CC-BY-4.0 license, with a last update timestamp of 2026-05-27.

Use Cases

• Selecting optimal decision thresholds for a Random Forest classifier based on varying cost ratios.
• Analyzing the trade-off between false negatives and false positives in a software build outcome prediction task.
• Benchmarking cost-sensitive classification methods using the TravisTorrent dataset as a reference point.

Strengths

• Dataset is small (5.5 KB), facilitating quick download and inspection.
• Results are shared under a permissive CC-BY-4.0 license.
• The analysis is focused on a specific, well-defined task of threshold optimization.

Limitations

• The dataset's row count and specific column definitions are unknown, limiting initial assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.
• The data appears to be a derived analysis output rather than a primary observational dataset.

Provenance

Source

figshare, author Lalit Narayan Mishra.

Collection Method

Likely derived from analysis of the TravisTorrent dataset using a Random Forest model.

Freshness

Last updated 2026-05-27 17:39:30; freshness should be verified.