Search Results (496)

Purpose: Predicting colorectal cancer liver metastasis (CRLM) is essential for prognostic assessment. This study aims to develop and validate an interpretable multimodal machine learning framework based on multiparametric MRI for predicting CRLM, and to enhance the clinical interpretability of the model through SHapley Additive exPlanations (SHAP) analysis and deep learning visualization. Methods: This multicenter retrospective study included 463 patients with pathologically confirmed colorectal cancer from two institutions, divided into training (n = 256), internal testing (n = 111), and external validation (n = 96) sets. Radiomics features were extracted from manually segmented regions on axial T2-weighted imaging (T2WI) and diffusion-weighted imaging (DWI). Deep learning features were obtained from a pretrained ResNet101 network using the same MRI inputs. A least absolute shrinkage and selection operator (LASSO) logistic regression classifier was developed for clinical, radiomics, deep learning, and combined models. Model performance was evaluated by AUC, sensitivity, specificity, and F1-score. SHAP was used to assess feature contributions, and Grad-CAM was applied to visualize deep feature attention. Results: The combined model integrating features across the three modalities achieved the highest performance across all datasets, with AUCs of 0.889 (training), 0.838 (internal test), and 0.822 (external validation), outperforming single-modality models. Decision curve analysis (DCA) revealed enhanced clinical net benefit from the integrated model, while calibration curves confirmed its good predictive consistency. SHAP analysis revealed that radiomic features related to T2WI texture (e.g., LargeDependenceLowGrayLevelEmphasis) and clinical biomarkers (e.g., CA19-9) were among the most predictive for CRLM. Grad-CAM visualizations confirmed that the deep learning model focused on tumor regions consistent with radiological interpretation. Conclusions: This study presents a robust and interpretable multiparametric MRI-based model for noninvasively predicting liver metastasis in colorectal cancer patients. By integrating handcrafted radiomics and deep learning features, and enhancing transparency through SHAP and Grad-CAM, the model provides both high predictive performance and clinically meaningful explanations. These findings highlight its potential value as a decision-support tool for individualized risk assessment and treatment planning in the management of colorectal cancer. Full article

Background: Acute myeloid leukemia (AML) is a common and aggressive adults hematological malignancies. This study explored megakaryocyte–erythroid progenitors (MEPs) signature genes and constructed a prognostic model. Methods: Uniform manifold approximation and projection (UMAP) identified distinct cell types, with differential analysis between AML-MEP and normal MEP groups. Univariate and the least absolute shrinkage and selection operator (LASSO) Cox regression selected biomarkers to build a risk model and nomogram for 1-, 3-, and 5-year survival prediction. Results: Ten differentially expressed genes (DEGs) related to overall survival (OS), six (AHSP, MYB, VCL, PIM1, CDK6, as well as SNHG3) were retained post-LASSO. The model exhibited excellent efficiency (the area under the curve values: 0.788, 0.77, and 0.847). Pseudotime analysis of UMAP-defined subpopulations revealed that MYB and CDK6 exert stage-specific regulatory effects during MEP differentiation, with MYB involved in early commitment and CDK6 in terminal maturation. Finally, although VCL, PIM1, CDK6, and SNHG3 showed significant associations with AML survival and prognosis, they failed to exhibit pathological differential expression in quantitative real-time polymerase chain reaction (qRT-PCR) experimental validations. In contrast, the downregulation of AHSP and upregulation of MYB in AML samples were consistently validated by both qRT-PCR and Western blotting, showing the consistency between the transcriptional level changes and protein expression of these two genes (p < 0.05). Conclusions: In summary, the integration of single-cell/transcriptome analysis with targeted expression validation using clinical samples reveals that the combined AHSP-MYB signature effectively identifies high-risk MEP-AML patients, who may benefit from early intensive therapy or targeted interventions. Full article

Background: Programmed cell death-related genes (PCDRGs) have been reported to play an important role in diagnosis, treatment and immunity regarding cancer, but their prognostic value and therapeutic potential in acute myeloid leukemia (AML) patients still need to be fully explored. Methods: Cox regression analysis and Least Absolute Shrinkage and Selection Operator (LASSO) analysis were used to identify PCDRGs significantly associated with the prognosis of AML patients. Furthermore, a prognostic risk model for AML patients was constructed based on the selected PCDRGs, and their immune microenvironment and biological pathways were analyzed. Cell experiments ultimately confirmed the potential role of PCDRGs in AML. Results: The results yielded four PCDRGs that were used to develop a prognostic risk model, and the prognostic significance of this model was confirmed using an independent external AML patient cohort. This prognostic risk model provides an independent prognostic risk factor for AML patients. This prognostic feature is related to immune cell infiltration in AML patients. The inhibition of solute carrier family 39 member 14 (SLC39A14) expression enhanced apoptosis and inhibited cell cycle progression in AML cells. Conclusions: This study integrates bioinformatics analysis and cellular experiments to reveal potential gene therapy targets and prognostic gene markers in AML. Full article

Background: Personalized anti-tumor therapy that activates the immune response has demonstrated clinical benefits in various cancers. However, its efficacy against testicular cancer (TC) remains uncertain. This study aims to identify suitable patients for anti-tumor immunotherapy and to uncover potential therapeutic targets in TC for the development of tailored anti-tumor immunotherapy. Methods: Consensus clustering analysis was conducted to delineate immune subtypes, while weighted gene co-expression network analysis (WGCNA), least absolute shrinkage and selection operator (LASSO) regression, and support vector machine (SVM) algorithms were employed to evaluate the potential efficacy of anti-tumor immunotherapy. Candidate immunotherapy targets were systematically identified through multi-gene panel analyses and subsequently validated using molecular biology assays. A prioritized target emerging from cellular screening was further evaluated for its capacity to potentiate anti-tumor immunity. The therapeutic efficacy of this candidate was rigorously confirmed through a comprehensive suite of immunological experiments. Results: Following systematic screening of five candidate genes (WNT11, FAM181B, GRK5, FSCN1, and ECHS1), GRK5 emerged as a promising therapeutic target for immunotherapy based on its distinct functional and molecular associations with immune evasion mechanisms. Cellular functional assays revealed that GRK5 knockdown significantly attenuated the malignant phenotype of testicular cancer cells, as evidenced by reduced proliferative capacity and invasive potential. Complementary immunological validation established that specific targeting of GRK5 with the selective antagonist GRK5-IN-2 disrupts immune evasion pathways in testicular cancer, as quantified by T-cell-mediated cytotoxicity. Conclusions: These findings position GRK5 as a critical modulator of tumor-immune escape, warranting further preclinical exploration of GRK5-IN-2 as a candidate immunotherapeutic agent. Full article

Integrating multi-source remote sensing can improve the accuracy of forest aboveground biomass (AGB) estimation. However, the accuracy and stability of the forest AGB estimation results are affected by multiple remote sensing feature variables as well as parameter tuning of machine learning algorithms. To this end, this study employed six types of remote sensing data—Landsat 8 OLI, Sentinel-2A, GEDI, ICESat-2, ALOS-2, and SAOCOM. A dual-variable selection strategy based on SHapley Additive exPlanations (SHAP) was developed, and a genetic algorithm (GA) was used to optimize the parameters of five machine learning models—elastic net (EN), least absolute shrinkage and selection operator (Lasso), support vector regression (SVR), Random Forest (RF), and Categorical Boosting (CatBoost)—to estimate the AGB of Pinus kesiya var. langbianensis forest in Wuyi Village, Zhenyuan County. The dual-variable selection strategy integrates SHAP with the Pearson correlation coefficient (PC), RF, EN, and Lasso to enhance feature screening robustness and interpretability. The results of the study showed that Lasso-SHAP dual-variate screening was more stable than SHAP univariate screening. In particular, the Lasso-SHAP strategy improved the average R² from 0.59 (using SHAP alone) to above 0.70, achieving an enhancement of 11%. Among GA-optimized parametric machine learning models, the linear GA-Lasso achieved the best performance, with an R² of 0.91 and an RMSE of 12.94 Mg/ha, followed by the GA-EN model (R² = 0.89, RMSE = 14.46 Mg/ha). For nonlinear models, GA-SVR performed the best (R² = 0.74, RMSE = 22.07 Mg/ha), surpassing the GA-CatBoost model (R² = 0.64, RMSE = 25.88 Mg/ha). In summary, the Lasso-SHAP dual-variable selection strategy effectively improves the estimation accuracy of AGB for Pinus kesiya var. langbianensis forests, while GA-optimized machine learning models demonstrate excellent performance, providing strong support for regional-scale forest resource monitoring and carbon stock assessment. Full article

Background/Objectives: Postoperative deep vein thrombosis (DVT) is a common and serious complication after tibial fracture surgery. This study aimed to develop and evaluate machine learning (ML) models to predict the occurrence of DVT following tibia fracture surgery. Methods: A retrospective analysis was conducted on patients who had undergone surgery for isolated tibial fractures. A total of 42 predictive models were developed using combinations of six ML algorithms—logistic regression, support vector machine, random forest, extreme gradient boosting, Light Gradient Boosting Machine (LightGBM), and neural networks—and seven feature selection methods, including SHapley Additive exPlanations (SHAP), Least Absolute Shrinkage and Selection Operator (LASSO), Boruta, recursive feature elimination, univariate filtering, and full-variable inclusion. Model performance was assessed based on discrimination, quantified by the area under the receiver operating characteristic curve (AUC-ROC), and calibration, measured using Brier scores, with internal validation performed via bootstrapping. Results: Of 471 patients, 80 (17.0%) developed postoperative DVT. The ML models achieved high overall accuracy in predicting DVT. Twenty-four models showed similarly excellent discrimination (pairwise AUC comparisons, p > 0.05). The top-performing model (random forest with RFE) attained an AUC of ~0.99, while several others (including LightGBM and SVM-based models) also reached AUC values in the 0.97–0.99 range. Notably, support vector machine models paired with Boruta or LASSO feature selection demonstrated the best calibration (lowest Brier scores), indicating reliable risk estimation. The final selected SVM models achieved high specificity (≥95%) with moderate sensitivity (~75–80%) for DVT detection. Conclusions: ML models demonstrated high accuracy in predicting postoperative DVT following tibial fracture surgery. Support vector machine-based models showed particularly favorable discrimination and calibration. These results suggest the potential utility of ML-based risk stratification to guide individualized prophylaxis, warranting further validation in prospective clinical settings. Full article

Background: Bacterial infections are frequent complications in critically ill COVID-19 patients, and are associated with increased morbidity, antibiotic use, and healthcare burden. Early and accurate identification of infection remains challenging. Pancreatic Stone Protein (PSP) has emerged as a promising biomarker of infection. In this study, PSP was evaluated alongside C-reactive protein (CRP). Methods: We conducted a prospective study including 105 critically ill COVID-19 patients admitted to the intensive care unit (ICU). Blood samples were collected at admission to measure PSP and CRP. A LASSO Least Absolute Shrinkage and Selection Operator (LASSO) regression model was used to identify independent predictors of proven or suspected bacterial infection. Mixed-effects models were applied to account for repeated measures and clinical confounders. Results: Among 105 patients, 57 (54%) developed bacterial infections. PSP levels were significantly higher in infected patients (median 100 ng/mL) than in non-infected patients (median 37 ng/mL, p < 0.001). CRP was also elevated in infected patients (median 125 vs. 70 mg/L, p = 0.015). The LASSO model retained PSP as the most informative predictor. In mixed-effects logistic regression, PSP remained significantly associated with infection (OR 1.017, 95% CI 1.006–1.027, p = 0.001). The AUC for PSP was 0.87. Conclusion: PSP appears to be a useful biomarker for early detection of bacterial infection in critically ill COVID-19 patients. Its integration into infection surveillance protocols could support antibiotic stewardship efforts and improve clinical decision-making. Full article

Background: Clostridioides difficile infection (CDI) is a major cause of healthcare-associated morbidity and mortality. Understanding the predictors of in-hospital mortality and recurrence of CDI is key for improving outcomes. This study combined traditional statistical methods and machine learning approaches to identify risk factors for these outcomes. Methods: We conducted a nationwide, retrospective study using the Spanish Minimum Basic Data Set at Hospitalization, analyzing 34,557 admissions with CDI from 2020 to 2022. Logistic regression combined with the least absolute shrinkage and selection operator (LASSO) was used to identify the most relevant predictors. Survival analyses using Cox regression and LASSO were also performed to assess time-to-mortality predictors. Results: Mortality and recurrence rates increased over the study period, reflecting the increasing burden of CDI. LASSO identified a parsimonious subset of predictors while maintaining predictive accuracy (area under the curve: 0.71). Older age (OR = 2.10, 95%CI: 1.98–2.22), Charlson Comorbidity Index ≥ 2 (OR = 1.42, 95%CI: 1.33–1.52), admission to the intensive care unit (OR = 3.09, 95%CI: 2.88–3.32), congestive heart failure (OR = 1.71, 95%CI: 1.61–1.82), malignancies (OR = 1.76, 95%CI: 1.66–1.87), and dementia (OR = 1.36, 95%CI: 1.25–1.48) were strongly associated with all-cause hospital mortality. For recurrence, age ≥ 75 years (OR = 1.19, 95%CI: 1.12–1.27), chronic kidney disease (OR = 1.15, 95%CI: 1.08–1.23), and chronic liver disease (OR = 1.43, 95%CI: 1.16–1.74) were the strongest predictors, while malignancy appeared protective, likely due to survivor bias. Conclusions: Our study provides a robust framework for predicting CDI outcomes. The integration of traditional statistical methods and machine learning applied to a large dataset may improve the reliability of the results. Our findings highlight the need for targeted interventions in high-risk populations and emphasize the potential utility of machine learning in risk stratification. Future studies should validate these models in external cohorts and explore survivor bias in malignancy-associated outcomes. Full article

Soil salinization severely limits global agricultural sustainability, particularly across the saline–alkaline landscapes of the Qinghai–Tibet Plateau. We examined how potassium fulvate (PF) modulates oat (Avena sativa L.) performance, soil chemistry, and rhizospheric microbiota in the saline–alkaline soils of the Qaidam Basin. PF markedly boosted shoot and root biomass, with the greatest response observed at 150 kg hm⁻². At the same time, it enhanced soil fertility by increasing organic matter, nitrate-N, ammonium-N, and available potassium, and improved ionic balance by lowering Na⁺ concentrations and the sodium adsorption ratio (SAR), while increasing Ca²⁺ levels and soil moisture content. Under the high-dose treatment (F2), endogenous fungal contributions declined sharply, exogenous replacements increased, and fungal α-diversity fell; multivariate ordinations confirmed that PF reshaped both bacterial and fungal communities, with fungi exhibiting the stronger response. We integrated three machine learning algorithms—least absolute shrinkage and selection operator (LASSO), Random Forest (RF), and eXtreme Gradient Boosting (XGBoost)—to minimize the bias inherent in any single method. We identified microbial β-diversity, organic matter, and Na⁺ and Ca²⁺ concentrations as the most robust predictors of the Soil Salinization and Alkalization Index (SSAI). Structural equation modeling further showed that PF mitigates salinity chiefly by improving soil physicochemical properties (path coefficient = −0.77; p < 0.001), with microbial assemblages acting as key intermediaries. These findings provide compelling theoretical and empirical support for deploying PF to rehabilitate saline–alkaline soils in alpine environments and offer practical guidance for sustainable land management in the Qaidam Basin. Full article

Objectives: Evidence suggests that COVID-19 infection contributes to elevated risks of psychiatric disorders, including depression and anxiety, however, this association remains underexplored. This study aimed to examine the incidence of depression and anxiety in individuals with COVID-19 compared to those without any infection. Method: Using the Utah All Payers Claims Database (2019 to 2021), we examined adult patients with continuous insurance enrollment. Individuals with pre-existing depression or anxiety were excluded. COVID-19 infection in 2020 was identified using diagnostic and procedural codes. The Least Absolute Shrinkage and Selection Operator (LASSO) method was applied to select covariates, followed by entropy balancing to adjust for baseline differences. Weighted logistic regression models were used to estimate the association between COVID-19 infection and incident mental health diagnoses in 2021. Results: Among 356,985 adults included in the final analytic sample for depression analysis, 37.6 percent had a documented COVID-19 infection in 2020. Individuals with prior infection had significantly higher odds of receiving a depression diagnosis in 2021 compared to those without infection (OR = 1.48, p < 0.01). A similar pattern was observed for anxiety: among 371,491 adults, 38.1 percent had a COVID-19 infection, and infected individuals had 46 percent greater odds of receiving an anxiety diagnosis (OR = 1.46, p < 0.01), after adjusting for demographic and clinical characteristics. Conclusions: This study highlights the elevated risk of depression and anxiety among patients who had been infected with COVID-19, emphasizing the importance of addressing the mental health needs of individuals affected by the virus. Full article

Compelling evidence has demonstrated a bidirectional relationship between Sjögren’s syndrome (SS) and periodontitis (PD). Nevertheless, the underlying mechanisms driving their co-occurrence remain unclear, highlighting the need for finding the hub gene. This study sought to examine the common genes and any connections between SS and PD. Differently expressed genes (DEGs) were analyzed by means of gene set enrichment analysis (GSEA), weighted gene co-expression network analysis (WGCNA), and least absolute shrinkage and selection operator (LASSO) methods. The test and validation sets were used to depict the receiver operating characteristic (ROC) curves. The immune cell infiltration was performed via the cell-type identification by estimating relative subsets of RNA transcripts (CIBERSORT) methodology. The relationships between immune infiltrating cells and the common gene were examined. Ninety-five common genes with similar expression trends were obtained after DEGs analysis, which were enriched in cytokine—cytokine receptor interaction, chemokine signaling pathway, proteasome, intestinal immune network for IgA production, and cytosolic DNA sensing pathway. Thirty-nine common genes were obtained after WGCNA. Sixteen shared genes of DEGs analysis and WGCNA were incorporated into the LASSO model to obtain the unique shared gene, C-C motif chemokine receptor 1 (CCR1), which overexpressed and owned predictable ROC curves in test and validation sets. The examination of immune cell infiltration underscored its crucial roles in the disturbance of immune homeostasis and the emergence of pathogenic circumstances with the simultaneous occurrence of SS and PD. CCR1 overexpresses and serves as a critical pathogenic hub linking SS and PD, which may play a role through immune cell infiltration. Full article

Background: The N⁶-methyladenosine (m⁶A) modification of eukaryotic mRNA is the most prevalent of such epigenetic modifications and has recently been identified as a potential player in the pathogenesis and progression of hepatocellular carcinoma (HCC). With the increasing emergence of immunotherapy in the treatment of HCC, we have evaluated the potential of m⁶A-related genes in predicting overall survival and the therapeutic efficacy of immunotherapy in HCC patients. Methods: We employed transcriptomic data from TCGA-LIHC and GSE76427, comprising a total of 485 HCC patients, as the training set. Based on 23 recognized m⁶A regulators, we performed clustering analysis on HCC patients. The intersecting differentially expressed genes (DEGs) among subtypes were used in least absolute shrinkage and selection operator (LASSO) Cox and multivariate Cox regression analyses to construct the risk model. For the quantification of a risk model of HCC patients, a risk score was developed and correlated with clinical and immunological parameters. Furthermore, a single-cell transcriptomic atlas was used to analyze the relationship between model genes and immune cell subpopulations. Mechanistic studies included in vitro assays to validate the association between the m⁶A-related gene ANLN and the progression of HCC. Results: Internal (TCGA and GEO) and external validation (ICGC) suggested that an 8-gene risk score provides an accurate and stable prognostic assessment for HCC. Furthermore, the high-risk score, characterized by elevated TP53 mutation frequency, tumor mutation burden (TMB), and tumor stem cell characteristics indicated a poor prognosis. The prognostic signature was associated with immune cell infiltration in HCC. Those patients with a high-risk score had lower immune tolerance with a better prediction of the efficacy of immunotherapy. The risk model helps to assess and predict the response and prognosis of HCC patients to immune checkpoint inhibitors (ICIs). Additionally, single-cell RNA sequencing data revealed that the high-risk group had a higher proportion of T cells and fewer immunosuppressive T cells, potentially correlating with a better response to immunotherapy. Finally, in vitro experiments showed that ANLN, an m⁶A-related gene, promoted the proliferation and migration of HCC cells. Conclusions: In this study, we identified and validated an m⁶A gene signature consisting of eight genes that can be used to predict prognosis and immunotherapy efficacy in HCC patients. Full article

This study investigates the performance prediction of poly-crystalline photovoltaic (PV) systems in Jordan using experimental data, analytical models, and machine learning approaches. Two 5 kWp grid-connected PV systems at Applied Science Private University in Amman were analyzed: one south-oriented and another east–west (EW)-oriented. Both systems are fixed at an 11° tilt angle. Linear regression, Least Absolute Shrinkage and Selection Operator (LASSO), ElasticNet, and artificial neural networks (ANNs) were employed for performance prediction. Among these, linear regression outperformed the others due to its accuracy, interpretability, and computational efficiency, making it an effective baseline model. LASSO and ElasticNet were also explored for their regularization benefits in managing feature relevance and correlation. ANNs were utilized to capture complex nonlinear relationships, but their performance was limited, likely because of the small sample size and lack of temporal dynamics. Regularization and architecture choices are discussed in this paper. For the EW system, linear regression predicted an annual yield of 1510.45 kWh/kWp with a 2.1% error, compared to 1433.9 kWh/kWp analytically (3.12% error). The south-oriented system achieved 1658.15 kWh/kWp with a 1.5% error, outperforming its analytical estimate of 1772.9 kWh/kWp (7.89% error). Productivity gains for the south-facing system reached 23.64% (analytical), 10.43% (experimental), and 9.77% (predicted). These findings support the technical and economic assessment of poly-crystalline PV deployment in Jordan and regions with similar climatic conditions. Full article

Background: Gliomas represent the most prevalent and aggressive primary brain tumors, requiring precise classification to guide treatment strategies and improve patient outcomes. Purpose: This study aimed to develop and evaluate a machine learning-driven approach for glioma classification by identifying the most relevant genetic and clinical biomarkers while demonstrating clinical utility. Methods: A dataset from The Cancer Genome Atlas (TCGA) containing 23 features was analyzed using an integrative approach combining Boruta, Least Absolute Shrinkage and Selection Operator (LASSO), and SHapley Additive exPlanations (SHAP) for feature selection. The refined feature set was used to train four machine learning models: Random Forest, Support Vector Machine, XGBoost, and Logistic Regression. Comprehensive evaluation included class distribution analysis, calibration assessment, and decision curve analysis. Results: The feature selection approach identified 13 key predictors, including IDH1, TP53, ATRX, PTEN, NF1, EGFR, NOTCH1, PIK3R1, MUC16, CIC mutations, along with Age at Diagnosis and race. XGBoost achieved the highest AUC (0.93), while Logistic Regression recorded the highest testing accuracy (88.09%). Class distribution analysis revealed excellent GBM detection (Average Precision 0.840–0.880) with minimal false negatives (5–7 cases). Calibration analysis demonstrated reliable probability estimates (Brier scores 0.103–0.124), and decision curve analysis confirmed substantial clinical utility with net benefit values of 0.36–0.39 across clinically relevant thresholds. Conclusions: The integration of feature selection techniques with machine learning models enhances diagnostic precision, interpretability, and clinical utility in glioma classification, providing a clinically ready framework that bridges computational predictions with evidence-based medical decision-making. Full article

Lithology identification during drilling plays a vital role in geological and geotechnical exploration, as it facilitates the early detection of formation-related hazards and supports the development of optimized mining strategies. Traditional lithology identification research involves problems such as fuzzy indicator characteristics and unbalanced sample quantities, which affect the accuracy and interpretability of model identification. In order to solve these problems, the Shanxi Guoqiang Coal Mine was taken as the research object, and a combined machine learning model was used to conduct a study on lithology identification during drilling. First, the least absolute shrinkage and selection operator (LASSO) algorithm was used to screen the independent variables and retain the parameters that contributed the most to lithology identification. Then, the synthetic minority oversampling technique (SMOTE) algorithm was used to expand the data samples, increase the amounts of minority sample data, and keep the ratios of various lithology data at 1:1. Then, the Bayesian optimization (BO) algorithm was used to optimize the penalty factor C and kernel function hyperparameter γ—two important parameters of the support vector machine (SVM) model—and the BO-SVM lithology identification model was established. Finally, the data samples were processed, and the results were compared with those of single models and unbalanced sample processing to evaluate their effect. The results showed the following: during the drilling process, the four indicators of drilling speed, mud pressure, slurry flow rate, and torque are strongly correlated with the lithology and can be used for lithology identification and classification research. After the data set was oversampled using the SMOTE algorithm, each model had better robustness and generalization ability; the classification result evaluation indicators were also greatly improved, especially for the random forest model, which had a poor original evaluation effect. The BO algorithm was used to optimize the parameters of the SVM model and establish a combined model that correctly identified 95 groups of data out of 96 groups of test samples with an identification accuracy rate of 99%, which was better than that of the traditional machine learning model. The evaluation results were compared with measured data, which confirmed the reliability of the combined model classification method and its potential to be extended to lithology identification and classification work. Full article