Induction of Similarity Measures and Medical Diagnosis Support Rules through Separable, Linear Data Transformations

Summary

Objectives: To improve the medical diagnosis support rules based on comparisons of diagnosed patients with similar cases (precedents) archived in a clinical database. The case-based reasoning (CBR) or the nearest neighbors (K-NN) classifications, which operate on referencing (learning) data sets, belong to this scheme.

Methods: Inducing similarity measure through special linear transformations of the referencing sets aimed at the best separation of these sets. Designing separable transformations can be based on dipolar models and minimization of the convex and piecewise linear (CPL) criterion functions in accordance with the basis exchange algorithm.

Results: Separable linear transformations allow for some data sets to decrease the error rate of the K-NNclassification rule based on the Euclidean distance. Such results can be seen on the example of data sets taken from the Heparsystem of diagnosis support.

Conclusions: Medical diagnosis support based on the CBRor the K-NNrules can be improved through separable transformations of the referencing sets.

Keywords

Medical diagnosis support - case-based reasoning - the nearest neighbors technique - the Euclidean distance function - linear transformations

• References

• 1 Perner P. Data Mining on Multimedia Data. Berlin:: Springer; 2002
• 2 Duda OR, Hart PE. Pattern Classification, Sec.. Edition. New York:: J. Wiley; 2001
• 3 Johnson RA, Wichern DW. Applied Multivariate Statistical Analysis. Englewood Cliffs New York:: Prentice-Hall, Inc.; 1991
• 4 Bobrowski l, Topczewska M. Improving the K-NN classification with the Euclidean distance through linear data transformations. ICDM 2004 Permer P. et al. Lipsk, Germany, Springer Verlag: 2004. Springer Lecture Notes in Computer Science (to appear).
• 5 Bobrowski l, Topczewska M. Linear visualising transformations and convex, piecewise linear criterion functions. Bioc and Biom Eng 2002; 22 (01) 69-78.
• 5 Bobrowski L.. Design of piecewise linear classifiers from formal neurons by some basis exchange. Pattern Recognition 1991; 24 (09) 863-70.
• 5 Bobrowski L., Wasyluk H.. Diagnosis supporting rules of the Hepar system. In: Petel VL, Rogers R, Haux R (eds). Pattern Recognition 2001; 1309-13.