Classification Accuracy Scores for Binary Transformed Probabilities Under Multiple Models

Victoria Sivill's dataset provides classification accuracy scores with standard deviation for the binary transformed P(M), P(P), and P(P|M) under a variety of classification models. The data is stored in an XLS file sized 5.5 KB and was last updated on May 28, 2026. The description notes that the most successful model on all three target variables was the neural network classifier.

Use Cases

• Compare neural network performance against other models based on the reported accuracy scores.
• Benchmark classification models on binary transformed probability targets P(M), P(P), and P(P|M).
• Analyze the standard deviation of accuracy scores to assess model stability across the described feature set.

Strengths

• Dataset is small (5.5 KB), facilitating quick download and inspection.
• Provides accuracy scores with standard deviation, offering a measure of performance variability.
• Explicitly identifies the neural network as the most successful model across three target variables.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

figshare, author Victoria Sivill

Collection Method

Likely results from a machine learning experiment comparing classifiers.

Freshness

Last updated 2026-05-28 17:21:25; freshness should be verified.