TY - GEN
T1 - Neural network signal analysis in immunology
AU - Theis, Fabian J.
AU - Hartl, Dominic
AU - Krauss-Etschmann, Susanne
AU - Lang, Elmar W.
PY - 2003
Y1 - 2003
N2 - This paper aims to investigate whether both supervised and unsupervised signal analysis contributes to the interpretation of immunological data. For this purpose a data base was set up containing measured data from bronchoalveolarlavage fluid which was obtained from 37 children with pulmonary diseases. The children were dichotomized into two groups: 20 children suffered from chronic bronchitis whereas 17 children had an interstitial lung disease. A self-organizing map (SOM) was utilized to test higher-order correlations between cellular subsets and the patient groups. Furthermore, a supervised approach with a perceptron trained to the patients' diagnosis was applied. The SOM confirmed the results that were expected from previous statistical analyses and shed light on formerly not considered relationships. The supervised perceptron learning after principal component analysis for dimension reduction turned out to be highly successful by linearly separating the patients into two groups with different diagnoses. The simplicity of the perceptron made it easy to extract diagnosis rules, which partly were known already and is now readily be tested on larger data sets.
AB - This paper aims to investigate whether both supervised and unsupervised signal analysis contributes to the interpretation of immunological data. For this purpose a data base was set up containing measured data from bronchoalveolarlavage fluid which was obtained from 37 children with pulmonary diseases. The children were dichotomized into two groups: 20 children suffered from chronic bronchitis whereas 17 children had an interstitial lung disease. A self-organizing map (SOM) was utilized to test higher-order correlations between cellular subsets and the patient groups. Furthermore, a supervised approach with a perceptron trained to the patients' diagnosis was applied. The SOM confirmed the results that were expected from previous statistical analyses and shed light on formerly not considered relationships. The supervised perceptron learning after principal component analysis for dimension reduction turned out to be highly successful by linearly separating the patients into two groups with different diagnoses. The simplicity of the perceptron made it easy to extract diagnosis rules, which partly were known already and is now readily be tested on larger data sets.
UR - http://www.scopus.com/inward/record.url?scp=84904306811&partnerID=8YFLogxK
U2 - 10.1109/ISSPA.2003.1224857
DO - 10.1109/ISSPA.2003.1224857
M3 - Conference contribution
AN - SCOPUS:84904306811
SN - 0780379462
SN - 9780780379466
T3 - Proceedings - 7th International Symposium on Signal Processing and Its Applications, ISSPA 2003
SP - 235
EP - 238
BT - Proceedings - 7th International Symposium on Signal Processing and Its Applications, ISSPA 2003
PB - IEEE Computer Society
T2 - 7th International Symposium on Signal Processing and Its Applications, ISSPA 2003
Y2 - 1 July 2003 through 4 July 2003
ER -