Health Evaluation and Promotion
Online ISSN : 1884-4103
Print ISSN : 1347-0086
ISSN-L : 1347-0086
48th JHEP conference 2020
Significance of Gender-Specific Aging Profiles at the Highest Similarity of Paired Tests between Two Data Components on Medical Checkup Examinees: DIM Mode Analysis of Spherical Surface SOM
Nobuhiko KasezawaHeizou TokutakaMasaaki OhkitaGen Niina
Author information
JOURNAL OPEN ACCESS

2020 Volume 47 Issue 5 Pages 578-584

Details
Abstract

 Recently, research based on self-organization has seen rapid worldwide development. The Self-Organizing Map (SOM) analysis technique, a practical application, has the benefit that it is not affected by distribution patterns as a non-linear regression method against complex system event data. Furthermore, SOM has the feature of outlining the entirety of the data while maintaining the mutual similarity relationship.

 Recent technical developments undertaken by the authors have allowed simultaneous on-screen display of the front and hidden rear sides of a spherical surface of SOM. This enables us to visually grasp the overall similarity relationships between components. In addition, the introduction of the DIM mode into SOM analysis, which calculates the degree of similarity between components, enables quantification of such similarity between components. This also allows for the comparison of strength among components. Through these processes, the SOM analysis technique has been improved. Such advancement facilitates development of an environment where it is feasible to apply data of the complex system analysis obtained from medical checkups with a wide range of variable factors to clinical practice.

 In this symposium, the authors examined data such as laboratory results, lifestyle interviews, and nutrient intake status in Japanese adults who received a medical checkup. Thereafter, we exhaustively calculated the degree of similarity by the latest SOM analysis method in all possible pairs between two components. We then searched the highest degree of similarity in aging between men and women among key components and the results were stored in the profiles. This enables us to dynamically understand the situation in which the strength of similarity between two components changes depending on aging in men and women. Additionally, in terms of clinical significance of the component-pairing, it can be used not only to predict the age at which symptoms often develop, and detect signs of symptoms without delay, but also to provide medical support and health guidance in a timely and appropriately manner.

 We evaluate the biological trend information of men and women associated with aging in detail based on our results and apply it to the health promotion system used for medical checkups. We believe that this will contribute to enabling more qualitative improvement of medical checkups.

Content from these authors
© 2020 Japan Society of Health Evaluation and Promotion
Previous article Next article
feedback
Top