1.Joint Relation Extraction of Famous Medical Cases with CasRel Model Combining Entity Mapping and Data Augmentation
Yuxin LI ; Xinghua XIANG ; Hang YANG ; Dasheng LIU ; Jiaheng WANG ; Zhiwei ZHAO ; Jiaxu HAN ; Mengjie WU ; Qianzi CHE ; Wei YANG
Chinese Journal of Experimental Traditional Medical Formulae 2026;32(2):218-225
ObjectiveTo address the challenges of unstructured classical Chinese expressions, nested entity relationships, and limited annotated data in famous traditional Chinese medicine(TCM) case records, this study proposes a joint relation extraction framework that integrates data augmentation and entity mapping, aiming to support the construction of TCM diagnostic knowledge graphs and clinical pattern mining. MethodsWe developed an annotation structure for entities and their relationships in TCM case texts and applied a data augmentation strategy by incorporating multiple ancient texts to expand the relation extraction dataset. A cascade binary tagging framework for relation triple extraction(CasRel) model for TCM semantics was designed, integrating a pre-trained bidirectional encoder representations from transformers(BERT) layer for classical TCM texts to enhance semantic representation, and using a head entity-relation-tail entity mapping mechanism to address entity nesting and relation overlapping issues. ResultsExperimental results showed that the CasRel model, combining data augmentation and entity mapping, outperformed the pipeline-based Bert-Radical-Lexicon(BRL)-bidirectional long short-term memory(BiLSTM)-Attention model. The overall precision, recall, and F1-score across 12 relation types reached 65.73%, 64.03%, and 64.87%, which represent improvements of 14.26%, 7.98%, and 11.21% compared to the BRL-BiLSTM-Attention model, respectively. Notably, the F1-score for tongue syndrome relations increased by 22.68%(69.32%), and the prescription-syndrome relations performed the best with the F1-score of 70.10%. ConclusionThe proposed framework significantly improves the semantic representation and complex dependencies in TCM texts, offering a reusable technical framework for structured mining of TCM case records. The constructed knowledge graph can support clinical syndrome differentiation, prescription optimization, and drug compatibility, providing a methodological reference for TCM artificial intelligence research.
2.Pharmacoeconomic evaluation of finerenone combined with standard treatment regimen in the treatment of diabetic nephropathy
Hai LIANG ; Runan XIA ; Panpan DI ; Mengmeng ZHAO ; Pengcheng ZHANG ; Yashen HOU ; Hong ZHANG ; Wei WU ; Miao YANG
China Pharmacy 2025;36(1):86-90
OBJECTIVE To evaluate the cost-effectiveness of finerenone combined with standard treatment regimen in the treatment of diabetic nephropathy (DN). METHODS From the perspective of healthcare service providers, a Markov model was established to simulate the dynamic changes of each stage in DN patients who received finerenone combined with the standard treatment regimen or the standard treatment regimen alone based on the phase Ⅲ clinical trial study of finerenone for DN. Markov model was used to perform the cost-effectiveness of long-term effects and the costs of the two therapies with a simulation cycle of 4 months, a simulation period of 15 years and an annual discount rate of 5%. At the same time, one-way sensitivity analysis and probability sensitivity analysis were performed, and the stability of the results was validated. RESULTS Accumulative cost of the standard treatment regimen was 579 329.54 yuan, and the accumulative utility was 8.052 4 quality-adjusted life year (QALYs); the accumulative cost of finerenone combined with the standard treatment regimen was 332 520.61 yuan, and the accumulative utility was 8.187 4 QALYs. Finerenone combined with the standard treatment regimen was more cost-effective. The results of one-way sensitivity analysis showed that dialysis status utility value, DN stage 3 utility value and DN stage 4 utility value had a great influence on the incremental cost-effectiveness ratio, but did not affect the robustness of the model. The results of probability sensitivity analysis showed that finerenone combined with the standard treatment regimen was more cost-effective with 100% probability. CONCLUSIONS For DN patients, finerenone combined with the standard treatment regimen is more cost-effective as an absolute advantage option.
3.Mechanism of Modified Shaofu Zhuyutang in Antagonising Ectopic Endometrial Tissue Fibrosis Based on Cellular Pyroptosis Mediated by TRL4/NF-κB/NLPR3 Signaling Pathway
Zuoliang ZHANG ; Jiaxing WANG ; Wanrun WANG ; Xiangyu LIN ; Bin YUE ; Zhirui ZHANG ; Yinan WANG ; Yaling YANG ; Dongqing WEI ; Cancan HUANG ; Quansheng WU
Chinese Journal of Experimental Traditional Medical Formulae 2025;31(4):19-28
ObjectiveTo investigate the mechanism of action of modified Shaofu Zhuyutang in antagonizing cellular pyroptosis and fibrosis in ectopic endometrial tissues of endometriosis through the Toll-like receptor 4/nuclear factor-κB/NOD-like receptor protein 3 (TRL4/NF-κB/NLPR3) signaling pathway. MethodsSeventy-two SPF-grade female SD rats were randomly divided into a sham-operated group (n = 12) and a modeling group (n = 60). The rats in the sham-operated group underwent a caesarean section, while the rats in the modeling group were used to establish an endometriosis model through the auto-transplantation method. After successful modeling, the animals were randomly divided into the model group, progesterone group (0.25 mg·kg-1), and modified Shaofu Zhuyutang low-, medium-, and high-dose groups (7.5, 15, 30 g·kg-1), with 12 animals in each group. After 4 weeks of drug administration, voluntary activity and heat pain latency were observed. The rats were sacrificed for tissue collection, and Masson staining were used to observe histopathological changes in the endometrial tissues. Enzyme-linked immunosorbent assay (ELISA) was used to measure serum levels of interleukin-18 (IL-18), interleukin-1β (IL-1β), tumor necrosis factor-α (TNF-α), and transforming growth factor-β (TGF-β). Immunohistochemistry (IHC) was used to detect the protein expression area of tumor necrosis factor-related factor 6 (TRAF6) and NLPR3 in the endometrial tissues. Immunofluorescence (IF) was used to detect the relative fluorescence intensity of Caspase-1 and gasdermin D (GSDMD) in the endometrial tissues. Western blot was employed to measure the relative expression of TRL4, myeloid differentiation factor 88 (MyD88), TRAF6, NF-κB p65, phosphorylated NF-κB p65 (p-NF-κB p65), and NLPR3 proteins in endometrial tissues. Real-time fluorescence quantitative polymerase chain reaction (Real-time PCR) was used to detect the mRNA expression of TRL4, MyD88, TRAF6, NF-κB, and NLPR3 in the endometrial tissues. ResultsCompared with the sham-operated group, rats in the model group showed reduced voluntary activity and shorter heat pain latency. Serum levels of IL-18, IL-1β, TNF-α, and TGF-β were elevated. The relative expression areas of TRAF6 and NLPR3 proteins were increased, and the relative fluorescence intensity of Caspase-1 and GSDMD was enhanced. The relative expression of TRL4, MyD88, TRAF6, NF-κB p65, p-NF-κB p65, and NLPR3 proteins, along with the expression of TRL4, MyD88, TRAF6, NF-κB, and NLPR3 mRNA, were significantly increased (P<0.01). Compared with the model group, rats in the progesterone group and the modified Shaofu Zhuyutang medium- and high-dose groups exhibited improved voluntary activity, longer heat pain latency, the fibrosis of endometrial tissue is alleviated. Serum levels of IL-18, IL-1β, TNF-α, and TGF-β were decreased. The relative expression areas of TRAF6 and NLPR3 proteins decreased, and the relative fluorescence intensity of Caspase-1 and GSDMD weakened. The relative expression of TRL4, MyD88, TRAF6, p-NF-κB p65, NLPR3 proteins, and TRL4, MyD88, TRAF6, NF-κB, and NLPR3 mRNA expression were reduced (P<0.05, P<0.01). ConclusionModified Shaofu Zhuyutang may play a therapeutic role in endometriosis by interfering with key proteins in the TRL4/NF-κB/NLPR3 signaling pathway, reducing NLRP3 inflammasome-induced cellular pyroptosis, antagonizing the fibrosis process in ectopic endometrial tissues, improving the inflammatory microenvironment in the pelvic cavity, and alleviating pain.
4.Expert Consensus on Clinical Application of Qinbaohong Zhike Oral Liquid in Treatment of Acute Bronchitis and Acute Attack of Chronic Bronchitis
Jian LIU ; Hongchun ZHANG ; Chengxiang WANG ; Hongsheng CUI ; Xia CUI ; Shunan ZHANG ; Daowen YANG ; Cuiling FENG ; Yubo GUO ; Zengtao SUN ; Huiyong ZHANG ; Guangxi LI ; Qing MIAO ; Sumei WANG ; Liqing SHI ; Hongjun YANG ; Ting LIU ; Fangbo ZHANG ; Sheng CHEN ; Wei CHEN ; Hai WANG ; Lin LIN ; Nini QU ; Lei WU ; Dengshan WU ; Yafeng LIU ; Wenyan ZHANG ; Yueying ZHANG ; Yongfen FAN
Chinese Journal of Experimental Traditional Medical Formulae 2025;31(4):182-188
The Expert Consensus on Clinical Application of Qinbaohong Zhike Oral Liquid in Treatment of Acute Bronchitis and Acute Attack of Chronic Bronchitis (GS/CACM 337-2023) was released by the China Association of Chinese Medicine on December 13th, 2023. This expert consensus was developed by experts in methodology, pharmacy, and Chinese medicine in strict accordance with the development requirements of the China Association of Chinese Medicine (CACM) and based on the latest medical evidence and the clinical medication experience of well-known experts in the fields of respiratory medicine (pulmonary diseases) and pediatrics. This expert consensus defines the application of Qinbaohong Zhike oral liquid in the treatment of cough and excessive sputum caused by phlegm-heat obstructing lung, acute bronchitis, and acute attack of chronic bronchitis from the aspects of applicable populations, efficacy evaluation, usage, dosage, drug combination, and safety. It is expected to guide the rational drug use in medical and health institutions, give full play to the unique value of Qinbaohong Zhike oral liquid, and vigorously promote the inheritance and innovation of Chinese patent medicines.
5.Effect modification of amino acid levels in association between polycyclic aromatic hydrocarbon exposure and metabolic syndrome: A nested case-control study among coking workers
Jinyu WU ; Jiajun WEI ; Shugang GUO ; Huixia XIONG ; Yong WANG ; Hongyue KONG ; Liuquan JIANG ; Baolong PAN ; Gaisheng LIU ; Fan YANG ; Jisheng NIE ; Jin YANG
Journal of Environmental and Occupational Medicine 2025;42(3):325-333
Background Exposure to polycyclic aromatic hydrocarbons (PAHs) is associated with the development of metabolic syndrome (MS). However, the role of amino acids in PAH-induced MS remains unclear. Objective To explore the impact of PAHs exposure on the incidence of MS among coking workers, and to determine potential modifying effect of amino acid on this relationship. Methods Unmatched nested case-control design was adopted and the baseline surveys of coking workers were conducted in two plants in Taiyuan in 2017 and 2019, followed by a 4-year follow-up. The cohort comprised 667 coking workers. A total of 362 participants were included in the study, with 84 newly diagnosed cases of MS identified as the case group and 278 as the control group. Urinary levels of 11 PAH metabolites and plasma levels of 17 amino acids were measured by ultrasensitive performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS). Logistic regression was used to estimate the association between individual PAH metabolites and MS. Stratified by the median concentration of amino acids, Bayesian kernel machine regression (BKMR) model was employed to assess the mixed effects of PAHs on MS. Due to the skewed data distribution, all PAH metabolites and amino acids in the analysis were converted by natural logarithm ln (expressed as lnv). Results The median age of the 362 participants was 37 years, and 83.2% were male. Compared to the control group, the case group exhibited higher concentrations of urinary 2-hydroxyphenanthrene (2-OHPhe), 9-hydroxyphenanthrene (9-OHPhe), and hydroxyphenanthrene (OHPhe) (P=0.005, P=0.049, and P=0.004, respectively), as well as elevated levels of plasma branched chain amino acid (BCAA) and aromatic amino acid (AAA) (P<0.05). After being adjusted for confounding factors, for every unit increase in lnv2-OHPhe in urine, the OR (95%CI) of MS was 1.57 (1.11, 2.26), and for every unit increase in lnvOHPhe, the OR (95%CI) of MS was 1.82 (1.16, 2.90). Tyrosine, leucine, and AAA all presented a significant nonlinear correlation with MS. At low levels, tyrosine, leucine, and AAA did not significantly increase the risk of MS, but at high levels, they increased the risk of MS. In the low amino acid concentration group, as well as in the low BCAA and low AAA concentration groups, it was found that compared to the PAH metabolite levels at the 50th percentile (P50), the log-odds of MS when the PAH metabolite levels was at the 75th percentile (P75) were 0.158 (95%CI: 0.150, 0.166), 0.218 (95%CI: 0.209, 0.227), and 0.262 (95% CI: 0.241, 0.282), respectively, However, no correlation between PAHs and MS was found in the high amino acid concentration group. Conclusion Amino acids modify the effect of PAHs exposure on the incidence of MS. In individuals with low plasma amino acid levels, the risk of developing MS increases with higher concentrations of mixed PAH exposure. This effect is partly due to the low concentrations of BCAA and AAA.
6.Clinical Safety Monitoring of 3 035 Cases of Juvenile Feilike Mixture After Marketing in Hospital
Jian ZHU ; Zhong WANG ; Jing LIU ; Jun LIU ; Wei YANG ; Yanan YU ; Hongli WU ; Sha ZHOU ; Zhiyu PAN ; Guang WU ; Mengmeng WU ; Zhiwei JING
Chinese Journal of Experimental Traditional Medical Formulae 2025;31(10):194-200
ObjectiveTo explore the clinical safety of Feilike Mixture (FLK) in the real world. MethodsThe safety of all children who received FLK from 29 institutions in 12 provinces between January 21,2021 and December 25,2021 was evaluated through prospective centralized surveillance and a nested case control study. ResultsA total of 3 035 juveniles were included. There were 29 research centers involved,which are distributed across 12 provinces,including one traditional Chinese medicine (TCM) hospital and 28 general hospitals. The average age among the juveniles was (4.77±3.56) years old,and the average weight was (21.81±12.97) kg. Among them,119 cases (3.92%) of juveniles had a history of allergies. Acute bronchitis was the main diagnosis for juveniles,with 1 656 cases (54.46%). FLK was first used in 2 016 cases (66.43%),and 142 juvenile patients had special dosages,accounting for 4.68%. Among them,92 adverse drug reactions (ADRs) occurred,including 73 cases of gastrointestinal system disorders,10 cases of metabolic and nutritional disorders,eight cases of skin and subcutaneous tissue diseases,two cases of vascular and lymphatic disorders,and one case of systemic diseases and various reactions at the administration site. The manifestations of ADRs were mainly diarrhea,stool discoloration,and vomiting,and no serious ADRs occurred. The results of multi-factor analysis indicated that special dosages (the use of FLK)[odds ratio (OR) of 2.642, 95% confidence interval (CI) of 1.105-6.323],combined administration: spleen aminopeptide (OR of 4.978, 95%CI of 1.200-20.655),and reason for combined administration: anti-infection (OR of 1.814, 95%CI of 1.071-3.075) were the risk factors for ADRs caused by FLK. Conclusion92 ADRs occurred among 3 035 juveniles using FLK. The incidence of ADRs caused by FLK was 3.03%,and the severity was mainly mild or moderate. Generally,the prognosis was favorable after symptomatic treatment such as drug withdrawal or dosage reduction,suggesting that FLK has good clinical safety.
7.Modified Buwangsan Ameliorates Cognitive Dysfunction in Rat Model of Type 2 Diabetes Mellitus by Regulating Autophagy in Hippocampus via PI3K/Akt/mTOR Pathway
Jie YANG ; Tonghua LIU ; Wei LIU ; Lili WU ; Lingling QIN
Chinese Journal of Experimental Traditional Medical Formulae 2025;31(13):104-113
ObjectiveTo evaluate the therapeutic effects of modified Buwangsan on cognitive dysfunction in the rat model of type 2 diabetes mellitus with mild cognitive impairment (T2DM-MCI) and explore the underlying mechanism. MethodsThirty-six 5-week-old SPF-grade SD rats were randomly assigned into 6 groups: Normal (Con, fed with a normal diet), model (DM, fed with a high-sugar and high-fat diet), low-dose modified Buwangsan (L-BWS, 1.86 g·kg-1), medium-dose modified Buwangsan (M-BWS, 3.72 g·kg-1), high-dose modified Buwangsan (H-BWS,7.44 g·kg-1), and huperzine A (SSJJ, 0.018 mg·kg-1). The rats were treated by gavage once a day for 12 weeks. The body weight and blood glucose level were monitored dynamically. Morris water maze was employed to test the cognitive function of rats. Hematoxylin-eosin and Nissl staining were employed to observe the pathological changes of the hippocampus. The levels of tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), and interleukin-6 (IL-6) in the serum and hippocampus were assessed by enzyme-linked immunosorbent assay. Western blotting was employed to determine the expression levels of key autophagy-related proteins including microtubule-associated protein 1 light chain 3 (LC3), type Ⅲ phosphatidylinositol 3-kinase complex regulatory subunit (Beclin1), and phosphorylated UNC-51-like kinase (p-ULK) 1/2 in the hippocampus. Immunofluorescence staining was employed to observe the regulation of p-PI3K/PI3K, p-mTOR/mTOR, and p-Akt/Akt ratios. ResultsCompared with the DM group, the L-BWS, M-BWS, H-BWS, and SSJJ groups showed increases in body weight at the end of the experiment (P<0.05), and the M-BWS, H-BWS and SSJJ groups showed declines in fasting blood glucose level (P<0.05). In the water maze test, compared with the DM group, the M-BWS, H-BWS, and SSJJ groups presented shortened escape latency (P<0.001). The L-BWS, M-BWS, H-BWS, and SSJJ group showcased regularly arranged cells in the hippocampus and cortex, markedly increased number of neurons, and significantly recovered Nissl bodies. Compared with the DM group, the L-BWS, M-BWS, H-BWS, and SSJJ groups had reductions in the levels of IL-1β and IL-6 in the serum and hippocampus (P<0.05), increases in the LC3-II/LC3-I ratio and expression level of beclin1 in the hippocampus (P<0.05) and the p-ULK level (P<0.05). The p-PI3K/PI3K, p-Akt/Akt, and p-mTOR/mTOR ratios in the hippocampus decreased in the M-BWS, H-BWS, and SSJJ groups (P<0.01). ConclusionModified Buwangsan significantly ameliorates cognitive dysfunction and neurological damage in the rat model of T2DM through multiple mechanisms. It regulates metabolic disorders, lowers the blood glucose level, improves lipid metabolism, and alleviates oxidative stress. It promotes the protection and repair of neurons by inhibiting inflammatory responses and activating the autophagy pathway in the hippocampus. At the same time, modified Buwangsan relieves autophagy inhibition by regulating the PI3K/Akt/mTOR signaling pathway to alleviate the brain tissue injury.
8.Modified Buwangsan Ameliorates Cognitive Dysfunction in Rat Model of Type 2 Diabetes Mellitus by Regulating Autophagy in Hippocampus via PI3K/Akt/mTOR Pathway
Jie YANG ; Tonghua LIU ; Wei LIU ; Lili WU ; Lingling QIN
Chinese Journal of Experimental Traditional Medical Formulae 2025;31(13):104-113
ObjectiveTo evaluate the therapeutic effects of modified Buwangsan on cognitive dysfunction in the rat model of type 2 diabetes mellitus with mild cognitive impairment (T2DM-MCI) and explore the underlying mechanism. MethodsThirty-six 5-week-old SPF-grade SD rats were randomly assigned into 6 groups: Normal (Con, fed with a normal diet), model (DM, fed with a high-sugar and high-fat diet), low-dose modified Buwangsan (L-BWS, 1.86 g·kg-1), medium-dose modified Buwangsan (M-BWS, 3.72 g·kg-1), high-dose modified Buwangsan (H-BWS,7.44 g·kg-1), and huperzine A (SSJJ, 0.018 mg·kg-1). The rats were treated by gavage once a day for 12 weeks. The body weight and blood glucose level were monitored dynamically. Morris water maze was employed to test the cognitive function of rats. Hematoxylin-eosin and Nissl staining were employed to observe the pathological changes of the hippocampus. The levels of tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), and interleukin-6 (IL-6) in the serum and hippocampus were assessed by enzyme-linked immunosorbent assay. Western blotting was employed to determine the expression levels of key autophagy-related proteins including microtubule-associated protein 1 light chain 3 (LC3), type Ⅲ phosphatidylinositol 3-kinase complex regulatory subunit (Beclin1), and phosphorylated UNC-51-like kinase (p-ULK) 1/2 in the hippocampus. Immunofluorescence staining was employed to observe the regulation of p-PI3K/PI3K, p-mTOR/mTOR, and p-Akt/Akt ratios. ResultsCompared with the DM group, the L-BWS, M-BWS, H-BWS, and SSJJ groups showed increases in body weight at the end of the experiment (P<0.05), and the M-BWS, H-BWS and SSJJ groups showed declines in fasting blood glucose level (P<0.05). In the water maze test, compared with the DM group, the M-BWS, H-BWS, and SSJJ groups presented shortened escape latency (P<0.001). The L-BWS, M-BWS, H-BWS, and SSJJ group showcased regularly arranged cells in the hippocampus and cortex, markedly increased number of neurons, and significantly recovered Nissl bodies. Compared with the DM group, the L-BWS, M-BWS, H-BWS, and SSJJ groups had reductions in the levels of IL-1β and IL-6 in the serum and hippocampus (P<0.05), increases in the LC3-II/LC3-I ratio and expression level of beclin1 in the hippocampus (P<0.05) and the p-ULK level (P<0.05). The p-PI3K/PI3K, p-Akt/Akt, and p-mTOR/mTOR ratios in the hippocampus decreased in the M-BWS, H-BWS, and SSJJ groups (P<0.01). ConclusionModified Buwangsan significantly ameliorates cognitive dysfunction and neurological damage in the rat model of T2DM through multiple mechanisms. It regulates metabolic disorders, lowers the blood glucose level, improves lipid metabolism, and alleviates oxidative stress. It promotes the protection and repair of neurons by inhibiting inflammatory responses and activating the autophagy pathway in the hippocampus. At the same time, modified Buwangsan relieves autophagy inhibition by regulating the PI3K/Akt/mTOR signaling pathway to alleviate the brain tissue injury.
9.Frontal and Parietal Alpha Asymmetry as Biomarkers for Negative Symptoms in Schizophrenia
Yao-Cheng WU ; Chih-Chung HUANG ; Yi-Guang WANG ; Chu-Ya YANG ; Wei-Chou CHANG ; Chuan-Chia CHANG ; Hsin-An CHANG
Psychiatry Investigation 2025;22(4):435-441
Objective:
Negative symptoms in schizophrenia indicate a poor prognosis. However, the mechanisms underlying the development of negative symptoms remain unclear. This study investigated the relationship between negative symptoms in schizophrenia and frontal alpha asymmetry (FAA).
Methods:
The study used a 32-channel electroencephalography to acquire alpha power in 4 target-paired sites in each patient. Regional alpha asymmetry was calculated based on the alpha power using EEGLAB Frontal Alpha Asymmetry Toolbox.
Results:
Sixty schizophrenia patients with predominant negative symptoms (PNS), 72 stabilized schizophrenia (SS) patients, and 73 healthy control (HC) participants were enrolled in this study. No significant differences were observed in FAA between the PNS and SS groups, although both groups exhibited reduced P3-P4 alpha asymmetry compared to HCs. A positive correlation was found between F7-F8 alpha asymmetry and illness duration. Additionally, a predictive model based on P3-P4 alpha asymmetry scores was able to differentiate schizophrenia patients from HCs, achieving a sensitivity of 71.2% and a specificity of 72.6%.
Conclusion
This study highlighted that parietal alpha asymmetry could serve as a valuable diagnostic tool for schizophrenia.
10.Comparison of Logistic Regression and Machine Learning Approaches in Predicting Depressive Symptoms: A National-Based Study
Xing-Xuan DONG ; Jian-Hua LIU ; Tian-Yang ZHANG ; Chen-Wei PAN ; Chun-Hua ZHAO ; Yi-Bo WU ; Dan-Dan CHEN
Psychiatry Investigation 2025;22(3):267-278
Objective:
Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic.
Methods:
Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC).
Results:
LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models.
Conclusion
Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Result Analysis
Print
Save
E-mail