Validation

blabla, validation is super blabla
Description...

Methods

k-fold Cross-Validation

cross_validation(sample_list, model, k_fold=3, epochs=20,
                 iterations=None, evaluation_path="evaluation",
                 draw_figures=True, run_detailed_evaluation=True,
                 callbacks=[], direct_output=False)

Function for an automatic k-fold Cross-Validation of the Neural Network model by running the whole pipeline several times with different data set combinations.

Arguments:

• sample_list: A list of sample indicies which will be used for validation.
• model: Instance of a Neural Network model class instance.
• k_fold: The number of k-folds for the Cross-Validation.
• epochs: Number of epochs. A single epoch is defined as one iteration through the complete data set.
• iterations: Number of iterations (batches) in a single epoch.
• evaluation_path: Path to the evaluation data directory. This directory will be created and used for storing all kinds of evaluation results during the validation processes.
• draw_figures: Boolean, whether evaluation figures should be automatically plotted in the evaluation directory.
• run_detailed_evaluation: Boolean, whether a detailed evaluation (additional prediction) should be performed.
• callbacks: A list of Callback classes for custom evaluation.
• direct_output: Boolean, whether computed evaluations will be output as the return of this function or if the evaluations will be saved on disk in the evaluation directory.

Returns:
None or validation results if direct_output is true. The variable validation_results is a list of Keras History objects containing all kinds of information. For each fold, a history object will created and appended to the list.

Example:

model = miscnn.Neural_Network(preprocessor=pp)

from miscnn.evaluation import cross_validation
cross_validation(sample_list, model, k_fold=3, epochs=50)

---

Percentage-Split Validation

get_indiceslist()

1-2 sentence description

Arguments:

• a: asd
• b: asd
• c: asd

Returns:
asd

Example:

function

---

Leave-One-Out Validation

get_indiceslist()

1-2 sentence description

Arguments:

• a: asd
• b: asd
• c: asd

Returns:
asd

Example:

function

---

Detailed Validation

get_indiceslist()

1-2 sentence description

Arguments:

• a: asd
• b: asd
• c: asd

Returns:
asd

Example:

function

---