본문 바로가기
  • Home

A Study on Deduction Method of Significant Symptoms Combination based on Null Hypotheses

  • Journal of Knowledge Information Technology and Systems
  • Abbr : JKITS
  • 2016, 11(4), pp.301-314
  • Publisher : Korea Knowledge Information Technology Society
  • Research Area : Interdisciplinary Studies > Interdisciplinary Research
  • Published : August 31, 2016

Oh yongtaek 1 KIM,ANNA 2

1아이누리한의원
2한국한의학연구원

Accredited

ABSTRACT

This study aims to propose an objective analysis method for significant symptoms combination applicable to vast amount of data including big data or clinical data instantly generated in real time, as a substitute for methods based on TKM experts' analysis. A mathematical method based on null hypotheses was used for analyzing relative multi frequency symptom-treatment combinations. These results were used for comparison with TKM experts' analysis on significant symptom-treatment combinations. Based on the correspondence of comparison of the two, for certain purposes relative multi frequency combinations may be considered and used as significant combinations. Therefore, the use of the mathematical method based on null hypotheses for analysing relative multi frequency combinations on text based combinations was appropriate. This method was applied to treatment related formula from Donguibogam and resulted in 39 combinations of significant symptom-symptom, symptom→condition, condition→symptom combinations. This study has some limitations using data only from pain treatment formula. Nevertheless, this limitation may be solved and lead to extended research using various sources of data. Combinations that were not applicable to the hypothesis "Combinations with frequent appearance in certain study subjects are more significant than combinations that are not" were not found. Also, combinations with relatively low frequency in appearance but was significant in result were not found. The limitations of this study requires further research to verify the hypothesis and analysis for combinations that cannot be covered by the hypothesis.

Citation status

* References for papers published after 2023 are currently being built.