|
Prediction of late postoperative complications based on the results of discriminant and correlation analyses of the early postoperative glycemia characteristics
T. V. Zakharovaab, A. V. Slivkinaa a Faculty of Computational Mathematics and Cybernetics,
M.V. Lomonosov Moscow State University, 1-52 Leninskiye Gory, GSP-1, Moscow 119991, Russian Federation
b Institute of Informatics Problems, Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow, 119333, Russian Federation
Abstract:
The application of the statistical methods for multidimensional observations classification, including discriminant analysis and data correlation analysis which make it possible to evaluate the strength of the factors' joint influence on the result, is considered. The main task is to predict complications development among patients, who underwent pancreas surgery. The research hypothesis suggests that the average early postoperative blood glucose level is of a crucial value for predicting carbohydrate metabolism disorders in the late postoperative period. The considered solution is based on a set of predictors (variables) associated with the blood glucose level, i. e., variance, sample size, mean, maximum, and minimum. The results of the discriminant analysis, performed using the STATISTICA software package on the basis of the available experimental data, do not confirm the research hypothesis.
Keywords:
Lambda Wilks; data sample; sample variance; range; discriminant analysis.
Received: 23.04.2017
Citation:
T. V. Zakharova, A. V. Slivkina, “Prediction of late postoperative complications based on the results of discriminant and correlation analyses of the early postoperative glycemia characteristics”, Sistemy i Sredstva Inform., 27:3 (2017), 63–73
Linking options:
https://www.mathnet.ru/eng/ssi529 https://www.mathnet.ru/eng/ssi/v27/i3/p63
|
|