analyze-results

Comparing ML experiment results across models and hyperparameters involves computing statistics and generating a comparison table with delta versus baseline. The skill automates this by discovering structured result files and organizing them by independent variables and dependent metrics.

Analyzing ML experiment robustness across seeds requires computing statistics across varying models, hyperparameters, and seeds to reveal performance trends. The skill automates this by parsing structured result files to surface robust improvements.

Generating a comparison table from JSON and CSV files requires parsing standardized result outputs in directories like figures/ or results/. The skill automates discovery, applies basic validation, and formats structured data into an organized comparison table.

Generating research insights from experiment statistics involves producing observations, interpretations, and implications from analyzed metrics. The skill automates reporting by suggesting actionable next steps for research papers or dashboards.

Analyzing machine learning experiments requires standardized result files in JSON or CSV format placed in common directories. The skill locates these structured data outputs, implements basic validation, and handles formatting and parsing errors.