Empirical Analysis of Ranking Models: Performance Measures for Dataset Search

Performance measures for dataset ranking models, provided as complementary material for a research paper. The data was contributed by author Angelo Batista Neves Júnior. The repository includes results for 15 models, including Bayesian, cosine similarity, J48, JRip, and SN variants.

Use Cases

• Comparing ranking model performance based on the listed model labels (e.g., 'bayesian-12C', 'cos-5L')
• Visualizing performance trends by generating 2D line plots for selected models
• Benchmarking different algorithmic approaches (Bayesian, cosine, decision trees) for dataset search tasks

Strengths

• Includes results for 15 distinct model configurations, allowing for comparative analysis
• Provides a script to generate visual plots from the performance data

Limitations

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

Provenance

Source

Angelo Batista Neves Júnior

Collection Method

Likely contains experimental results from evaluating dataset ranking models.