This data was originally used in a competition in the IEEE World
Congress on Computational Intelligence, 2008. The classification
problem is to diagnose whether a certain symptom exists or does not
exist in an automotive subsystem. Each case consists of 500
measurements of engine noise and a classification. There are two
separate problems:

For FordA the Train and test data set were collected in typical
operating conditions, with minimal noise contamination.

%The winning entry achieved 100\% test accuracy with Ford A and
86.2\% on Ford B. The best results for Ford B were obtained by
constructing a classifiers on the ACF features and/or the Power
Spectrum. The data has not been normalised. The original
competition website seems to no longer work.

http://home.comcast.net/~nn_classification/}
