Description

The ASPIDA research project will involve the development of novel pattern recognition models based on the algorithmic theory of randomness and the application of these models to the problem of medical diagnosis. More specifically, a tool will be created for the diagnosis of illnesses in children suffering from acute abdominal pain. This application field is considered as extremely important, since wrong diagnoses for this problem can result in devastating consequences (even to death).

An important advantage of the diagnostic tool that will be developed is that its diagnoses will be complemented with confidence measures, which will give the user the ability to evaluate the degree to which he can rely on each diagnosis. This will be achieved with the use of a new powerful technique called conformal prediction, which is one of the most important developments of the past few years in the field of machine learning.

The main expected results of the project are:

1. two new conformal prediction algorithms based on neural networks and genetic algorithms,
2. the application of these two algorithms to the problem of medical diagnosis and the evaluation of their results, and
3. the creation of a diagnostic system for the diagnosis of children suffering from acute abdominal pain.

NEWS

PUBLICATIONS