::::::: Matrix - Extracting Knowledge out of Data :::::::

corporate - patient data analysis

the project:

-	performed in collaboration with a German university hospital

the goal:

-	classify alcoholics (before and after treatment) and non-alcoholics using results of routine blood parameters

the data:

-	89 blinded patient datasets, each containing 96 blood parameters
-	all blood parameters within 'normal range' (for alcoholics and non-alcoholics)

the results:

-	Matrix indentified 5 blood parameters that allowed researchers to discriminate 2 groups from each other, and identified 10 parameters that allowed for three groups to be characterised simultaneously
-	Matrix's algorithm was 100% accurate in its ability to predict which blood tests corresponded to alcoholics before and after treatment

classification based on clinical blood parameters

(training datasets indicated by dark colors;
examination (blind) datasets indicated by light colors)

2009 Matrix Advanced Solutions | Terms of use | site notice/Impressum