Сохранить в:
Библиографические подробности
Главный автор: Di Rienzo, Mirco
Формат: Recurso digital
Язык:
Опубликовано: Zenodo 2025
Предметы:
Online-ссылка:https://doi.org/10.5281/zenodo.15849997
Метки: Добавить метку
Нет меток, Требуется 1-ая метка записи!
Оглавление:
  • <p>This app allows you to perform regression on images containing graph trends, using various fitting methods; approximate an equation for the chosen method; and make predictions. It is divided into three modules:</p> <ul> <li> <p><strong>Module A</strong><br>You can upload and edit an image, set editing and analysis parameters, run regressions or clustering, and then export the resulting data.</p> </li> <li> <p><strong>Module B</strong><br>You train a model using the CSV files generated in Module A. This module also performs statistical analysis on the scatterplot points to identify statistical outliers via a z-score and p-value on the principal features that defined the classes.</p> </li> <li> <p><strong>Module C</strong><br>You automatically generate CSVs from multiple images by either loading a reference CSV of analysis parameters or by using the sidebar (which is itself divided into submodules). Within Module C, you choose whether to reinforce the model trained manually in Module B or to classify using the current (unsupervised) model. Finally, you can select the incorrect predictions and calculate the model’s overall accuracy.</p> </li> </ul>