ABSTRACT
The application progress of machine learning in research of acupuncture and moxibustion was reviewed from three aspects: mining of acupuncture and moxibustion prescription and indications, acupuncture efficacy prediction and its influencing factors, acupoint specificity and acupuncture manipulation research, and the existing problems in current research and future research trends were discussed. It is believed that the appropriate machine learning algorithm should be selected to build the model according to the research purpose and data characteristics in the future research; attention should be paid to feature design, feature selection and feature cleaning; sample data collection should be a priority, and data sharing platform and standardized data collection should be developed to improve the data quality.
Subject(s)
Acupuncture , Acupuncture Points , Acupuncture Therapy , Machine Learning , MoxibustionABSTRACT
The article "Effect of acupuncture vs sham acupuncture on live births among women undergoing in vitro fertilization: a randomized clinical trial", published in in May 2018, has concluded that acupuncture does not improve the rate of live births among women undergoing IVF. Through careful study of the article, the author analyzes its reliability from acupuncture therapeutic plan and specific acupuncture operation. As a result, although the research showed no significant difference between the acupuncture group and the sham acupuncture group, it could not prove no therapeutic effect in the sham acupuncture group, so the conclusion that the acupuncture did not improve the therapeutic effect could not be drawn; the compatibility of acupoints was inconsistent with the previous protocol, and its rationality was controversial; whether the frequency and duration of acupuncture treatment could highlight the live birth rate should be further discussed. In addition, the selection of acupuncturists may be another reason for the failure of the research aim.