Development of an Acupoint Recommendation Algorithm Based on the Analysis of Parkinson’s Disease Case Reports

Objectives : This study aimed to analyze individual cases where acupuncture therapy has been used to improve symptoms of Parkinson's disease and to develop a system that achieves the following objectives: (1) recommending acupuncture points used in cases with similar combinations of symptoms, and (2) suggesting frequently used acupoints for specific symptoms.Methods : We conducted a comprehensive literature search of papers published up to July 8, 2025, utilizing databases such as ScienceON, OASIS, and CNKI.Case reports on Parkinson's disease treatment were identified and compiled into a structured database.Using the Pandas, NumPy, and sklearn libraries, two algorithms were developed.In the first algorithm, the user inputs the patient's symptoms, and the algorithm finds the most similar cases, recommending the acupuncture treatments and acupoints used in those cases.In the second algorithm, the user inputs specific symptoms and acupuncture treatments, and the algorithm recommends the most frequently used acupoints for those inputs.Results : Both algorithms achieved high mean accuracy of K-fold cross validation in providing reliable recommendations.These validation results showed that the algorithms have the ability to accurately recommend acupoints and treatment methods based on Parkinson's disease symptoms.This indicates that they have potential for effective clinical application.Conclusions : The developed algorithms showed high accuracy in recommending acupuncture points and treatments based on patient symptoms, with promising results for clinical application.Future research with more diverse cases could enhance the generalizability of these recommendations.This case-based algorithmic research can make both technical and clinical contributions to the effective acupuncture treatment of Parkinson's disease, and furthermore, it holds the potential to evolve into various related platforms and studies.