Risk Prediction as a Service: A RESTful Machine Learning Pipeline for the Healthcare Domain


The thesis aims at designing a prototype of a configurable Machine Learning pipeline for the healthcare domain, where risk prediction is provided as a RESTful web service. The pipeline should enable selection of the model to train, the desired performance metrics to achieve, the structure of input data compliant with the model application. Promising tools are Smile and SparkML.

Keywords:  machine learning, as a service, REST, healthcare, risk prediction

Tags:
    

Thesis

Supervision

— supervisor
Andrea Omicini
— co-supervisor
Stefano Mariani

Category

2nd-Cycle Thesis

Status

out-of-date

Language

wgb.gif

Dates

— available since
15/03/2018

Partita IVA: 01131710376 - Copyright © 2008-2021 APICe@DISI Research Group - PRIVACY