Evaluating Misclassification Ratios in Region Identification in Flow Cytometry Data Using an SVM Based on Neural Network

摘要

This innovative approach presents the development of new algorithms to optimize the pattern recognition of different white blood cell types within real-world flow cytometry data. The behavior of parametric data clusters in a multidimensional space is analyzed using the learning systems known as Support Vector Machines (SVM). Beckman-Coulter Corporation supplied flow cytometry data of 30 patients to be used in the development of the algorithm. Subsequently, the characteristics of the cells provided in these sets were used to train an SVM neural network. The goal will be to write an automatic program to train the network to identify the different white blood cell subpopulations of a sample and provide information to medical doctors in the form of diagnostic references. With the application of the hypothesis space, the learning bias and the learning algorithm, the SVM network will be trained to evaluate misclassification ratios in flow cytometry data in an effort to assess the ubiquitous problem of data overlap.