Detection of Confused Blood Samples by Self-Organaizing Maps

Abstract

In this paper, we propose a detection of confusion for blood samples based on SOM(Self-Organizing Maps). We apply the differentials of time-series CBC(Complete Blood Count) as blood test data, and it is assumed that a confusion is occurred between subjects. The SOM of our method classifies input data into two categories, namely confused data and non-confused ones. Experimental results show that our method achieves the high accuracy of detection especially when the input data, not to be employed during the learning, are applied.