Читать книгу Bioinformatics and Medical Applications - Группа авторов - Страница 58

According to an article titled “Automatic classification of lung cancer from cellular images using deep symmetric neural networks”, we evaluated three types of cancer and trained for classification. However, pictures from the same sample belonged to the same group.

While executing cross-validation algorithm in different sets of images, we found that, in Set 1, there are 28 items and respective cross-validation score is 5,280 for adenocarcinoma. There are 42 items and cross-validation score is 5478 for squamous cell carcinoma. For small cell carcinoma, there are 26 images and cross-validation score is 5,070.

In Set 2, there are 28 items and respective cross-validation score is 5,184 for adenocarcinoma. There are 37 items and cross-validation score is 5,220 for squamous cell carcinoma. For small cell carcinoma, there are 33 images and cross-validation score is 5,280.

In Set 2, there are 26 items and respective cross-validation score is 5,040 for adenocarcinoma. There are 46 items and cross-validation score is 5,310 for squamous cell carcinoma. For small cell carcinoma, there are 32 images and cross-validation score is 5,214.