Scafuri, Luca (2023) Supervised machine learning methodologies for bladder cancer progression risk classification and clinical patient management. [Tesi di dottorato]
Anteprima |
Testo
Scafuri_Luca_36.pdf Download (3MB) | Anteprima |
| Tipologia del documento: | Tesi di dottorato |
|---|---|
| Lingua: | English |
| Titolo: | Supervised machine learning methodologies for bladder cancer progression risk classification and clinical patient management |
| Autori: | Autore Email Scafuri, Luca lucaluca@hotmail.it |
| Data: | 7 Dicembre 2023 |
| Numero di pagine: | 217 |
| Istituzione: | Università degli Studi di Napoli Federico II |
| Dipartimento: | Ingegneria Elettrica e delle Tecnologie dell'Informazione |
| Dottorato: | Information and Communication Technology for Health |
| Ciclo di dottorato: | 36 |
| Coordinatore del Corso di dottorato: | nome email Riccio, Daniele daniele.riccio@unina.it |
| Tutor: | nome email Pasquino, Nicola [non definito] Marinelli, Alfredo [non definito] |
| Data: | 7 Dicembre 2023 |
| Numero di pagine: | 217 |
| Parole chiave: | bladder cancer; artificial intelligence; machine learning; classification; progression risk; |
| Settori scientifico-disciplinari del MIUR: | Area 09 - Ingegneria industriale e dell'informazione > ING-INF/07 - Misure elettriche e elettroniche Area 06 - Scienze mediche > MED/06 - Oncologia medica |
| Depositato il: | 23 Gen 2024 22:45 |
| Ultima modifica: | 09 Mar 2026 14:01 |
| URI: | http://www.fedoa.unina.it/id/eprint/15690 |
Abstract
In this thesis work, supervised learning methodologies were compared for the classification of the risk of progression of bladder cancer. In particular, three supervised learning algorithms were applied (Decision Tree, Random Forest and Naive Bayes) and the accuracy results were compared both considering the data set composed of all the variables and the data set containing a lower number of variables through the use of a dimensionality reduction technique (Feature Selection). The data refers to a sample of 111 patients and both qualitative and quantitative variables were considered as input: sex, body mass index, smoking, family history, age, muscular invasiveness of the tumor, dimensions of the bladder wall, number and dimensions of lymph nodes, number and size of liver and bone lesions, etc. These variables were used to predict the risk of progression of bladder cancer, which is divided into four classes from 1 to 4: Low (risk 1), Medium-low (risk 2), Medium-high (risk 3), High (risk 4). The main goal is to give the medical oncologist objective support for the diagnosis of disease progression, a diagnosis which is not always easy to carry out, especially in the case of minimal and/or subtle disease progression, thus avoiding: exposing the patient to unnecessary side effects resulting from the use of a drug that is starting to lose its effectiveness, and to spend the budget relating to healthcare costs unnecessarily. From the analysis of the results it can be seen that the predictions for the risk of progression of bladder cancer are satisfactory for the three algorithms used. Comparing the predictions obtained, the dimensionality reduction technique is reliable, the loss of accuracy is balanced by the advantage of having a lower number of predictors and this represents an advantage for the computational effort of the algorithm.
Downloads
Downloads per month over past year
Actions (login required)
 |
Modifica documento |
