CLUSTERIZATION PROBLEM SOLUTION BY COMPETITIVE LEARNING FOR INCOMPLETE STATISTICAL DATA |
1 | |
2010 |
scientific article | 004.93'14:004.853 | ||
220-225 | clusterization, competitive learning, missing values in statistical data |
A review of modern methods to restore missing values in incomplete statistical data has been presented. A procedure
to combine statistical data clustering with the retrieval of missing values based on a modification of the Kohonen
network competitive learning algorithm has been proposed. The efficiency of the proposed procedure has
been demonstrated by solving a practical problem to determine individual target doses of radioiodine (131I) in the
treatment of diffuse toxic goiter. |
![]() |