Based on the classification analysis of aerobic lung diseases (such as chronic obstructive pulmonary

A new international medical study has found that the associated outcomes are of important reference value for disease prevention. There is a lack of risk grades for and identification of co-morbidity associated with autism spectrum disorder (AS) and pulmonary disease (PD) and risk factors for these co-morbidities. The study builds on an ML model based on immune-related cells and immune-induced markers to identify co-morbidity associated with autism spectrum disorder (AS) and pulmonary disease.

Four types of machine learning algorithms have been developed and evaluated, including decision tree, random forest, XGBoost, and support vector machines (SVMs). These algorithms were used to distinguish between AS and AS+PD. A subset analysis of lung clots (PN) and chronic obstructive pulmonary disease (COPD) was used for sensitivity analysis, using unexplained data.

363 AS patients were grouped, including only 193 AS and 170 AS+PD patients (including 123 with PN and 47 with COPD). The LASSO index returns to identifying age, neutrophils, IgG, ESR, CRP, NLR, PNR, and PIV as the key predictive variables (AS+PD). In the general population, SVM models showed excellent generalization and resistance to over-adaptation, while Random Forest and XGBoost showed over-adaptation.

In subgroup analysis, XGBoost became the best classifier for AS+PN and AS+COPD, and its good correction and net clinical validation are direct evidence of its robustness and consistency. In patients with AS+PN and AS+COPD pulmonary complications, specific immune-related cells and immune-mediating markers (especially white cells, IgG, NLR, and PI) were shown to be present. In the four ML models, SVM performed the best overall generalization, while XGBoost performed extremely well in the PN and COPD subgroups.

These findings support the integration of immune-inflammatory markers with the ML algorithm to improve the recognition of current pulmonary and symptomatic states to help identify AS pulmonary and symptomatic states. It should be pointed out that the above content is based on international public research findings and is for reference in HealthCop and does not constitute clinical advice. If you have any health questions, please consult professional medical personnel.