Machine learning

I found WEKA (Waikato Environment for Knowledge Analysis) when I was searching on the web on related topics ("machine learning" "software"). It is a collection of machine learning algorithms for data mining tasks [Dimov 07]. It is written in Java, developed at the University of Waikato (New Zealand).

[Available with GNU GPL licence at: https://www.cs.waikato.ac.nz/ml/weka/]

As an example I have taken a database from the collection of the University of California, Irvine (USA) under the name Post-Operative Patient Database.

Explanation of the nature of the problem

This is based on a real problem. Given a current inpatient at the post-operative stage, the idea is to determine the best place for the patient to continue his/her recovery on the basis of certain data (list of attributes) obtained from the relevant medical tests.

Patient-associated information (list of attributes):

• Internal temperature.
• Surface temperature.
• Oxygen concentration.
• Blood pressure.
• Stability of surface temperature.
• Stability of internal temperature.
• Level of wellness shown.

The decision taken will lead to the patient being sent to (distribution of classes):

• (I) ICU (Intensive Care Unit).
• (S) Home (discharge).
• (A) A general hospital ward.

I used the following files: post-operative.arff, instance.arff.