Advances in Chinese Medicine Diagnosis: From Traditional Methods to Computational Models

Effects of Diagnostic Errors in Pattern Differentiation and Acupuncture Prescription: A Single-Blinded, Interrater Agreement Study

“…To objectively test whether the simulated patient could have its pattern differentiated (either correctly or not) among several candidates in a dataset, an automated model for pattern differentiation must be used. PDA was chosen among other automated methods because it is the most accurate automated method for pattern differentiation considering a large set of patterns regardless of an underlying medical disease [ 23 ].…”

“…Parameters related to the diagnostic performance of each rater were obtained from 2 × 2 contingency tables [ 23 ] made from the comparison between the results of the simulation by SimTCM (gold-standard method) and the pattern differentiation by each rater: true positives ( TP ): manifestation profiles simulated with the target-pattern that were correctly identified by the rater as present; false negatives ( FN ): manifestation profiles simulated with the target-pattern that were erroneously identified by the rater as absent; false positives ( FP ): manifestation profiles not simulated with the target-pattern that were erroneously identified by the rater as present; true negatives ( TN ): manifestation profiles not simulated with the target-pattern that were correctly identified by the rater as absent. …”

Promoting integrative medicine by computerization of traditional Chinese medicine for scientific research and clinical practice: The SuiteTCM Project

“…In this scenario, computer methods are the best choice for management, processing and integration of several aspects that might be overlooked if considered in a case-by-case fashion [15] . Indeed, computerization of TCM as a scientific approach is not a new topic since several computational models were already proposed for acquisition, processing, and analysis of data [16][17][18][19][20] . However, the majority of existent models were developed to perform specific tasks (e.g., pattern differentiation, dataset organization, data mining, computer-aided drug design) and thus loose the integrative nature of health care in TCM.…”

“…It was suggested that modern medicine should fully incorporate TCM theories and practices provided that they are supported by scientific evidence [3] . In this direction, an interesting scientific approach is to seek for evidence that corroborates data reported by TCM practitioners or not -ancient and contemporary ones included [17] -instead of testing new theories based on TCM. The epidemiologic approach is well suited to study the morbidity of TCM in human populations.…”

Advances in Patient Classification for Traditional Chinese Medicine: A Machine Learning Perspective

“…Gu and Chen [ 6 ] discussed the relationship between modern bioinformatics and TCM and concluded that the major focuses of current TCM researches are to understand the pathological mechanisms of TCM from the systems biology perspective and facilitate novel drug design based on the analysis of TCM herbal medicine. Ferreira [ 7 ] introduced the validation methods of computational models for diagnostic assessment and also analyzed some potential reasons of misdiagnosis results. In another work [ 10 ], Ferreira and Lopes reviewed pattern classification researches regarding specific disorders such as genitourinary, cardiovascular, neurologic, surgical, and the like.…”