223. Use of Machine Learning techniques in predicting the course of relapsing-remitting MS in individual patients
Invited abstract in session WC-4: Optimization in regression, classification and learning I, stream Optimization in regression, classification and learning.
Wednesday, 10:05 - 11:20Room: M:M
Authors (first author is the speaker)
| 1. | Raffaele Mariosa
|
| DIAG, Sapienza, University of Rome | |
| 2. | Laura Palagi
|
| Department of Computer, Control, and Management Engineering A. Ruberti, Sapienza University of Rome |
Abstract
We aim to use ML tools to derive a robust and accurate prognostic tool to personalize the treatment of multiple sclerosis (MS). MS is the leading cause of progressive neurological disability in young people, having a high impact on social and economic costs. An early prediction of disease course would allow to differentiate treatment strategies based on the expected severity of the disease. Most ML studies use unconventional data, hindering real-world applicability regardless of performance. Here, we applied ML on data collected by the Italian National MS Registry (IMSR) to predict medium-term disease course. We employed traditional supervised ML algorithms (XG-Boost, Support Vector Machine, Balanced Random Forest Classifiers), ensembles and auto-ML tools to predict the state of an RR patient after time T (180, 360 and 720 days) from the current visit. Moreover, to address the critical need for model interpretability and explainability, we use optimal tree mathematical programming models to extract the main rules for classification. In particular, we use MIRET [1] that offers also a hierarchical visualization tool based on a heatmap representation of the tree ensemble's feature use. Our model of active collaboration between doctors and data analyst has been successful in optimizing the analytical workflow, emphasizing the importance of data quality and suggesting strategies for enhancing both database management and analysis.
Keywords
- Artificial intelligence based optimization methods and appl
- Complexity and efficiency of optimization algorithms
- Optimization for learning and data analysis
Status: accepted
Back to the list of papers