Raxol.Benchmark.StatisticalAnalyzer (Raxol v2.0.1)

View Source

Advanced statistical analysis for benchmark results. Provides percentiles, distributions, outlier detection, and statistical significance testing.

Summary

Functions

Perform comprehensive statistical analysis on benchmark results.

Analyze the distribution of data.

Calculate confidence intervals using bootstrap method.

Calculate a specific percentile value.

Calculate all standard percentiles (P50, P75, P90, P95, P99, P99.9).

Detect outliers using the IQR (Interquartile Range) method.

Detect outliers using MAD (Median Absolute Deviation) method.

Generate histogram data for visualization.

Perform Mann-Whitney U test (non-parametric).

Perform statistical significance testing (t-test).

Functions

analyze(results)

Perform comprehensive statistical analysis on benchmark results.

analyze_distribution(data)

Analyze the distribution of data.

calculate_confidence_intervals(data, confidence_level \\ 0.95, iterations \\ 1000)

Calculate confidence intervals using bootstrap method.

calculate_percentile(sorted_data, percentile)

Calculate a specific percentile value.

calculate_percentiles(sorted_data)

Calculate all standard percentiles (P50, P75, P90, P95, P99, P99.9).

detect_outliers(sorted_data)

Detect outliers using the IQR (Interquartile Range) method.

detect_outliers_mad(data, threshold \\ 3)

Detect outliers using MAD (Median Absolute Deviation) method.

generate_histogram(data, bins \\ 20)

Generate histogram data for visualization.

mann_whitney_u_test(sample1, sample2)

Perform Mann-Whitney U test (non-parametric).

t_test(sample1, sample2, opts \\ [])

Perform statistical significance testing (t-test).