Data modeling to obtain a classifier for recognizing the output feature

This article discusses the application of machine learning in medical field, namely, classification acute pancreatitis. disease was chosen as a priority solution for number reasons: presence subjective indicators on part patient, requirement high speed provision necessary treatment accordance with set severity, lack equipment correct diagnosis severity. Thus, course study, classifier built that recognizes severe and mild With help selection model's hyper-parameters, it happened at each stage model optimization, leading kernel radial-basis function. Based results feature space by thinning method core, sets features are different from other, both quality quantity, were identified. radial basis function selected itself average quantity between polynomial hyperbolic tangent. At end to check accuracy adjusted regressions, rolling exam conducted, during which certain established 1vR 3vR led conclusion first third class labels linearly separable, but second label cannot be distinguished. Moreover, two regressions aggregate clearly distinguish three clusters.