DATA ANALISYS IN DETERMINATION OF APPLICABILITY OF DOGS FOR SPECIFIC SERVICE
The object of development and research is mathematical classification models that are applied to the initial data set with the characteristics of dogs performing a certain type of activity. The purpose of the work is to investigate the possibility of using classification methods to assess the suitability of dogs for service dogs. As a result of the work, several classification methods were investigated: k-nearest neighbors, classification tree, logistic regression. A program has been developed to prepare the initial data set for further mining using selected methods. Experiments have been carried out related to assessing the accuracy of each method, and the criteria that have the greatest influence on the results of testing dogs have been identified.
Authors: D. V. Kozlov, Ya. A. Bekeneva
Direction: Informatics, Computer Technologies And Control
Keywords: Classification, k-nearest neighbors, classification tree, logistic regression
View full article