Opening the black box: interpretable machine learning for predictor finding of metabolic syndrome.

OBJECTIVE: The internal workings ofmachine learning algorithms are complex and considered as low-interpretation "black box" models, making it difficult for domain experts to understand and trust these complex models. The study uses metabolic syndrome (MetS) as the entry point to analyze and evaluate the application value of model interpretability methods in dealing with difficult interpretation of predictive models. METHODS: The study collects data from a chain of health examination institution in Urumqi from 2017 ~ 2019, and performs 39,134 remaining data after preprocessing such as deletion and filling. RFE is used for feature selection to reduce redundancy; MetS risk prediction models (logistic, random forest, XGBoost) are built based on a feature subset, and accuracy, sensitivity, specificity, Youden index, and AUROC value are used to evaluate the model classification performance; post-hoc model-agnostic interpretation methods (variable importance, LIME) are used to interpret the results of the predictive model. RESULTS: Eighteen physical examination indicators are screened out by RFE, which can effectively solve the problem of physical examination data redundancy. Random forest and XGBoost models have higher accuracy, sensitivity, specificity, Youden index, and AUROC values compared with logistic regression. XGBoost models have higher sensitivity, Youden index, and AUROC values compared with random forest. The study uses variable importance, LIME and PDP for global and local interpretation of the optimal MetS risk prediction model (XGBoost), and different interpretation methods have different insights into the interpretation of model results, which are more flexible in model selection and can visualize the process and reasons for the model to make decisions. The interpretable risk prediction model in this study can help to identify risk factors associated with MetS, and the results showed that in addition to the traditional risk factors such as overweight and obesity, hyperglycemia, hypertension, and dyslipidemia, MetS was also associated with other factors, including age, creatinine, uric acid, and alkaline phosphatase. CONCLUSION: The model interpretability methods are applied to the black box model, which can not only realize the flexibility of model application, but also make up for the uninterpretable defects of the model. Model interpretability methods can be used as a novel means of identifying variables that are more likely to be good predictors.

Construction of Xinjiang metabolic syndrome risk prediction model based on interpretable models.

BACKGROUND: We aimed to construct simple and practical metabolic syndrome (MetS) risk prediction models based on the data of inhabitants of Urumqi and to provide a methodological reference for the prevention and control of MetS. METHODS: This is a cross-sectional study conducted in the Xinjiang Uygur Autonomous Region of China. We collected data from inhabitants of Urumqi from 2018 to 2019, including demographic characteristics, anthropometric indicators, living habits and family history. Resampling technology was used to preprocess the data imbalance problems, and then MetS risk prediction models were constructed based on logistic regression (LR) and decision tree (DT). In addition, nomograms and tree diagrams of DT were used to explain and visualize the model. RESULTS: Of the 25,542 participants included in the study, 3,267 (12.8%) were diagnosed with MetS, and 22,275 (87.2%) were diagnosed with non-MetS. Both the LR and DT models based on the random undersampling dataset had good AUROC values (0.846 and 0.913, respectively). The accuracy, sensitivity, specificity, and AUROC values of the DT model were higher than those of the LR model. Based on a random undersampling dataset, the LR model showed that exercises such as walking (OR=0.769) and running (OR= 0.736) were protective factors against MetS. Age 60 ~ 74 years (OR=1.388), previous diabetes (OR=8.902), previous hypertension (OR=2.830), fatty liver (OR=3.306), smoking (OR=1.541), high systolic blood pressure (OR=1.044), and high diastolic blood pressure (OR=1.072) were risk factors for MetS; the DT model had 7 depth layers and 18 leaves, with BMI as the root node of the DT being the most important factor affecting MetS, and the other variables in descending order of importance: SBP, previous diabetes, previous hypertension, DBP, fatty liver, smoking, and exercise. CONCLUSIONS: Both DT and LR MetS risk prediction models have good prediction performance and their respective characteristics. Combining these two methods to construct an interpretable risk prediction model of MetS can provide methodological references for the prevention and control of MetS.

Relationship between vitamin D receptor gene polymorphisms and second acid-fast bacilli smear-positive during treatment for tuberculosis patients.

PURPOSE: Tuberculosis (TB) treatment is associated with Vitamin D. This study aimed to explore the relationship between Vitamin D receptor (VDR) gene polymorphisms and second acid-fast bacilli (AFB) smear-positive during treatment for TB patients. METHODS: This was a cross-sectional study. Seven hundred and thirty-one TB patients whose single nucleotide polymorphism site (SNPs) of VDR gene were detected from December 2019 to December 2020 in XinJiang of China. The genotypic distributions in each group were tested separately for Hardy-Weinberg equilibrium. The tetragram test was used to construct haplotypes to evaluate the association between each haplotype and second AFB smear-positive occurrence. RESULTS: No significant deviations were observed with all the four polymorphism sites in the genotypic distributions (P>0.05). Linkage disequilibrium (LD) analysis showed that there was LD between SNPs of VDR gene (r2=0.74, D`>0.9). Each haplotype was not considered to be the influencing factor of second AFB smear-positive. CONCLUSIONS: There is no association between VDR gene polymorphism (ApaI, BsmI, FokI and TaqI) and second AFB smear-positive.

The relationship between vitamin D level and second acid-fast bacilli (AFB) smear-positive during treatment for TB patients was inferred by Bayesian network.

BACKGROUND: Vitamin D is related to human immunity, so we used Bayesian network model to analyze and infer the relationship between vitamin D level and the acid-fast bacilli (AFB) smear-positive after two months treatment among pulmonary tuberculosis (TB) patients. METHODS: This is a cross-sectional study. 731 TB patients whose vitamin D level were detected and medical records were collected from December 2019 to December 2020 in XinJiang of China. Logistic regression was used to analyze the influencing factors of second AFB smear-positive. Bayesian network was used to further analyze the causal relationship among vitamin D level and the second AFB smear-positive. RESULTS: Baseline AFB smear-positive (OR = 6.481, 95%CI: 1.604~26.184), combined cavity (OR = 3.204, 95%CI: 1.586~6.472), full supervision (OR = 8.173, 95%CI:1.536~43.492) and full management (OR = 6.231, 95%CI:1.031~37.636) were not only the risk factors and can also be considered as the reasons for second AFB smear-positive in TB patients (Ensemnle > 0.5). There was no causal relationship between vitamin D level and second AFB smear-positive (Ensemnle = 0.0709). CONCLUSIONS: The risk factors of second AFB smear-positive were baseline AFB smear-positive, combined cavity, full supervision and full management. The vitamin D level in TB patients was not considered as one of the reasons for the AFB smear-positive.

Simulation Study of 4H-SiC Trench Insulated Gate Bipolar Transistor with Low Turn-Off Loss.

In this work, an insulated gate bipolar transistor (IGBT) is proposed that introduces a portion of the p-polySi/p-SiC heterojunction on the collector side to reduce the tail current during device turn-offs. By adjusting the doping concentration on both sides of the heterojunction, the turn-off loss is further reduced without sacrificing other characteristics of the device. The electrical characteristics of the device were simulated through the Silvaco ATLAS 2D simulation tool and compared with the traditional structure to verify the design idea. The simulation results show that, compared with the traditional structure, the turn-off loss of the proposed structure was reduced by 58.4%, the breakdown voltage increased by 13.3%, and the forward characteristics sacrificed 8.3%.