Repozytorium publikacji - Politechnika Gdańska

Liquid biopsy is a useful, minimally invasive diagnostic and monitoring tool for cancer disease. Yet, developing accurate methods, given the potentially large number of input features, and usually small datasets size remains very challenging. Recently, a novel feature parameterization based on the RNA-sequenced platelet data which uses the biological knowledge from the Kyoto Encyclopedia of Genes and Genomes, combined with a classifier based on the Convolutional Neural Network (CNN), allowed significantly improving the classification accuracy. In this work, we take a closer look at this approach and find that similar results can be obtained using significantly smaller models. Additionally, competitive results were achieved using gradient boosting. Since it has another advantage of adding interpretability to the model, we further analyze it in this work.

Autorzy

• dr inż. Sebastian Cygert link otwiera się w nowej karcie ,
• inż. Franciszek Górski link otwiera się w nowej karcie ,
• Piotr Juszczyk link otwiera się w nowej karcie ,
• Sebastian Lewalski link otwiera się w nowej karcie ,
• mgr inż. Krzysztof Pastuszak link otwiera się w nowej karcie ,
• prof. dr hab. inż. Andrzej Czyżewski link otwiera się w nowej karcie ,
• dr Anna Supernat