Data analysis on patients affected by Myelodysplastic Syndrome

Project of course Applied Statistics @ PoliMi, A.A. 2020/2021, Second Semester

Team

• Luca Caivano (luca.caivano@mail.polimi.it)
• Manfred Nesti (manfred.nesti@mail.polimi.it)
• Taguhi Mesropyan (taguhi.mesropyan@mail.polimi.it)
• Michele Precuzzi (michele.precuzzi@mail.polimi.it)

The project, in collaboration with Humanitas Research Hospital, had the purpose of analyzing medical data coming from patients affected by Myelodisplastic Syndrome: a rare deasese that can progress into Acute Myeloid Leukemia, had very high mortality and for which only cure is trasplantation. Our main goal was to investigate the effect of trasplantation relying on classical statistical inference method and some non-parametrized statistical method such as survival analysis.

Moreover we used an innovative approach based on combination between Neural Networks and clustering in order to clusterize patients based on theit mortality risk and trying to predict the time of survival (Survival Clustering).

The code can be found in the different directories, taking into account tha data are missing for NDA agreement. The main results obtained can be found in the final presentation.