General Latent Structure Models and Statistical Validation of Osteoporosis syndromes feature of community women

摘要

Objective: Defining bask syndromes feature of Osteoporosis (OP) is an important avenue for prevention and treatment of TCM (traditional Chinese medicine) for OP patients. We use machine learning techniques to establish objective and quantitative diagnosis standards for syndrome differentiation of OP and to give statistical validation of OP syndromes feature of community women. Method: One crucial step of TCM diagnosis is syndrome differentiation which is simply a classifier that classifies patients into different classes based on their symptoms. Previous clustering methods are unable to cope with the complexity of TCM. We therefore present a new multidimensional clustering method in the form of General latent structure (GLS) models, which is a useful tool of latent structure analysis. In this paper, the notion of GLS model is provided. We use an efficient learning algorithm to achieve an optimal GLS model from the OP real data set. Further, we use GLS model to analyze and present a case study of TCM syndrome differentiation for OP from qualitative and quantitative contents. Results: Our analysis has found natural clusters that correspond well to TCM basic syndrome types of OP population. The GLS model reflects much better model quality, provides statistical validation for TCM syndrome types of OP patients of community women and suggests the possibility of establishing objective and quantitative diagnosis standards, for syndrome differentiation on OP.