Echocardiographic Data Mining

E-chocardiography Journal: Alphabetical List / Chronological List / Images / Home Page

Data mining techniques include neural networks and inductive decision trees. We employed neural networks to analyze integrated backscatter from the interventricular septum in patients undergoing dobutamine stress echocardiography. Inductive decision trees were employed in the analysis of the Studies of Left Ventricular Dysfunction (SOLVD) Registry Echocardiography substudy database. The See5 program was used to generate rules and desision trees for the presence or absence of left ventricular hypertrophy in patients with left ventricular dysfunction.

Links to data mining sites:

See5 tutorial.
Machine learning repository at UCI includes an echocardiographic post myocardial infarct prognosis database from the Reed Institute in Miami. The description and the data are available using ftp (file transfer protocol).

Back to E-chocardiography Home Page.

e-mail:shindler@umdnj.edu