ML Model Training overview

Full specification is available here.
immuneML version: {{immuneML_version}}
Dataset details
General information
Name {{dataset_name}}
Type {{dataset_type}}
Dataset size {{dataset_size}}
Analysis labels
{{#labels}} {{/labels}}
Label name Label values (classes)
{{name}} {{values}}
Parameters for training ML model
Metrics
Optimization metric {{optimization_metric}}
Other metrics {{other_metrics}}
Cross-validation settings
assessment {{assessment_desc}}
selection {{selection_desc}}

Optimization results

{{#hp_per_label}}

{{label}}

The following table shows a summary of training ML models for the label {{label}}. For each of the data splits made in the assessment (outer) cross-validation loop, it shows which of combinations of ML settings had the best {{optimization_metric}} score out of all. By clicking ‘see details’ you will find more detailed information about the performance of the trained models on each data split, and generated reports.

{{#assessment_results}} {{/assessment_results}}
Split index Optimal settings (preprocessing, encoding, ML) Optimization metric ({{optimization_metric}}) Details
{{index}} {{hp_setting}} {{optimization_metric_val}} see details
{{/hp_per_label}}

Trained models

Trained models are available as zip files which can be directly provided as input for the MLApplication instruction and used to encode the data and predict the label on a new dataset. These zip files include trained ML model, encoder and preprocessing that were chosen as optimal for the given label, along with additional files showing the values of each parameter in the model and encoder.

{{#models_per_label}}
Download {{label}} model here.
{{/models_per_label}}
{{#show_hp_reports}}

Hyperparameter reports

Hyperparameter reports are executed on the trained models and have access to assessment and selection data (both the outer and the inner loop of nested cross-validation) and typically show trends in the models, over different splits to training, validation or test datasets.

{{#hp_reports}}

Report {{name}}

{{info}}

{{#output_figures}} {{#is_embed}}
{{#name}}

{{name}}

{{/name}}
{{/is_embed}} {{^is_embed}}
{{#name}}

{{name}}

{{/name}}
{{/is_embed}} {{/output_figures}}
{{/hp_reports}}
{{/show_hp_reports}}