Evaluation

This module contains an algorithm to evaluate proposed models. Author: Lucas Pavanelli

class evaluation.Evaluation(output_folder)

Evaluates proposed models.

output_folderstr

Path to output folder

output_folderstr

Path to output folder

static convert_output_to_text(y, out_id2w)

Converts output list containing integers to a text list.

ylist

Output list containing integers.

out_id2wdict

Map of index to word.

list

Text list.

evaluate(num_experiment, y_true, y_pred)

Evaluates model’s predictions using classification_report, generates a csv containing classification report, and return micro avg f1-score

num_experimentint

Current experiment number

y_truelist

True output list.

y_predlist

Predicted output list.

float

F1 score of current experiment

generate_output_csv(file_name, y_true, y_pred, test_tokens)

Generates a csv containing test results.

file_namestr

Name of csv file.

y_truelist

True output list.

y_predlist

Predicted output list.

test_tokenslist

List of tokens in test data

static get_single_output_id_list(y)

Converts a list of lists into a single list.

ylist

Output list.

list

Single list.