Classification Performance Metrics for Digital Pathology

Comprehensive evaluation of machine learning models and algorithms for digital pathology applications. Provides classification metrics, ROC analysis, segmentation quality assessment, and statistical model comparison for AI validation and algorithm comparison studies.

Usage

mlpathology(
  data,
  analysis_type = "classification",
  actual_labels,
  predicted_labels,
  predicted_probabilities,
  reference_segmentation,
  predicted_segmentation,
  model1_predictions,
  model2_predictions,
  model1_probabilities,
  model2_probabilities,
  roc_analysis = TRUE,
  roc_comparison = FALSE,
  confidence_level = 0.95,
  confusion_matrix_plot = TRUE,
  roc_plot = TRUE,
  bootstrap_validation = FALSE,
  bootstrap_runs = 1000
)

Arguments

data

the data as a data frame

analysis_type

Type of performance analysis to conduct

actual_labels

True/actual classification labels

predicted_labels

Predicted classification labels from model

predicted_probabilities

Predicted probabilities for ROC analysis (binary classification)

reference_segmentation

Reference/ground truth segmentation masks (binary)

predicted_segmentation

Predicted segmentation masks from model (binary)

model1_predictions

Predictions from first model for comparison

model2_predictions

Predictions from second model for comparison

model1_probabilities

Probabilities from first model for ROC comparison

model2_probabilities

Probabilities from second model for ROC comparison

roc_analysis

Perform ROC curve analysis for binary classification

roc_comparison

Compare ROC curves between two models using DeLong's test

confidence_level

Confidence level for performance metrics

confusion_matrix_plot

Generate confusion matrix heatmap

roc_plot

Generate ROC curve plot

bootstrap_validation

Use bootstrap sampling for confidence intervals

bootstrap_runs

Number of bootstrap replicates for validation

Value

A results object containing:

results$instructions					a html
results$confusionmatrix					a table
results$performancemetrics					a table
results$rocanalysis					a table
results$segmentationmetrics					a table
results$modelcomparison					a table
results$roccomparison					a table
results$confusionmatrixplot					an image
results$rocplot					an image
results$interpretation					a html

Tables can be converted to data frames with asDF or as.data.frame. For example:

results$confusionmatrix$asDF

as.data.frame(results$confusionmatrix)

Examples

data('classification_results')

mlpathology(data = classification_results,
           analysis_type = 'classification',
           actual_labels = actual,
           predicted_labels = predicted,
           predicted_probabilities = prob)