Search Results (726)

As global super-large open-pit mines expand in scale and extraction depth, conventional single-stage planning cannot meet the combined demands of productivity and resource recovery, making the shift to underground block caving inevitable. This study outlines the systemic challenges of block-scale extraction and the rationale for adopting multi-level spatial design decision-making. Four core model categories are briefly proposed: ultimate pit limit optimization, gravity flow simulation for draw strategy, long-term production scheduling for large-scale computation, and probabilistic frameworks addressing geological and market uncertainty. A Bayesian network-based block decision model is then proposed and decoupled into three physical decision tiers. The first tier incorporates energy prices, transport costs, and ore prices to establish an economic boundary rating robust to market volatility. The second tier aggregates mining units with discrete-event perturbations to produce a reliability-oriented production rating. The third tier integrates rock mechanics parameters with in situ monitoring data to derive a physics-informed safety rating. The three ratings are synthesized via Bayesian inference and evaluated within a multi-attribute utility function encompassing net present value, safety index, downside risk, and information risk. A feedback module quantifies the economic benefit of uncertainty reduction, yielding a closed-loop intelligent system spanning macroeconomic boundary definition to operational safety alerting. Finally, the main conclusion of this study is that integrating macro-economic volatility with rock mechanics through a dynamic Bayesian framework is essential for managing the open-pit to underground transition. The results indicate that leveraging the Value of Information for real-time risk diagnosis significantly reduces conservative design losses, providing a quantifiable and robust decision-making paradigm for super-large mining systems. Full article

Respiratory disease classification using lung sound recordings remains challenging due to signal interference, heterogeneous acquisition conditions, and substantial overlap among clinically related acoustic patterns. This study presents a framework for respiratory disease classification using NMF-enhanced log-mel spectrograms and deep neural classifiers. Respiratory sound recordings from two publicly available datasets were harmonized into a unified label space comprising Asthma, Bronchiectasis, Bronchiolitis, COPD, Healthy, Pneumonia and URTI. Following signal standardization and fixed-length segmentation, a non-negative matrix factorization (NMF)-based enhancement stage was applied to increase the salience of respiratory components prior to log-mel spectrogram generation. The proposed classifier was a convolutional recurrent neural network (CRNN) that combined convolutional feature extraction, bidirectional recurrent modelling, and attention-based temporal aggregation. For comparison, RDLINet, a conventional CNN, ResNet, and a YOLO-style backbone were implemented under the same preprocessing and training framework. Experimental results demonstrated that the proposed CRNN achieved the best overall performance, attaining 96.14 ± 0.50% accuracy and 94.05 ± 1.21% Macro-F1 on the unified seven-class cohort. Class-wise analysis, confusion-matrix evaluation, and output-space visualization further showed that the CRNN provided more balanced recognition across disease categories and clearer class separation than competing architectures. These findings indicate that NMF-enhanced spectro-temporal modelling combined with convolutional recurrent learning offers an effective approach for automated multi-class respiratory disease classification. Full article

Background/Objectives: Osteoporosis is a major public health concern associated with increased fracture risk and reduced quality of life if not detected at an early stage. Automated analysis of knee X-ray images using artificial intelligence has shown promising potential for opportunistic osteoporosis screening. This study aims to develop and evaluate a multi-stage deep learning and machine learning framework for osteoporosis classification, with particular emphasis on external validation, calibration drift, and cross-domain generalization performance. Methods: Knee X-ray images were categorized into three classes: Normal, Osteopenia, and Osteoporosis. Deep features were extracted using pretrained convolutional neural networks, including ResNet18, EfficientNetB0, and DenseNet121. The extracted features were subsequently classified using multiple machine learning models, including Neural Network, Efficient Linear, Support Vector Machine, and Naive Bayes classifiers. Two data augmentation strategies were investigated: targeted minority-class augmentation and full 3× dataset expansion with class balancing. Model performance was evaluated using accuracy, precision, recall, F1-score, and AUC on internal validation, independent test sets, and external validation datasets. Additional analyses included reliability calibration assessment, isotonic recalibration, and class-prior boosting with cross-validated threshold optimization to address external domain shift. Results: EfficientNetB0 and DenseNet121 consistently outperformed ResNet18 across most evaluation metrics. Under the balanced augmentation strategy, EfficientNetB0 combined with Efficient Linear demonstrated strong and stable performance, while DenseNet121 paired with a Neural Network achieved the highest overall classification performance. External validation revealed a substantial discrepancy between AUC and threshold-based metrics, indicating the presence of calibration drift and class-prior mismatch across imaging domains. Reliability analysis showed severe probability collapse in the Osteopenia class during external testing. Post-hoc recalibration improved probability reliability, while class-prior boosting substantially increased Osteopenia sensitivity and improved balanced accuracy and macro F1-score under external validation conditions. Conclusions: The proposed framework demonstrates the feasibility of combining pretrained CNN-based deep feature extraction with machine learning classifiers for osteoporosis classification from knee X-ray images. The findings further highlight that maintaining model performance under external testing conditions may require not only strong feature extraction capability but also adaptive recalibration and deployment-aware threshold optimization to address calibration drift and cross-domain variability. While the results are encouraging, the present study should be considered a proof-of-concept investigation. Although the framework was evaluated using an independent public external dataset, further validation using larger and more diverse multi-center clinical cohorts is necessary to establish generalizability and clinical utility before routine clinical implementation can be considered. Full article

Accurate automated interpretation of upper-extremity musculoskeletal radiographs remains challenging because fracture appearance varies across anatomical regions and can be subtle under class imbalance. This study proposes a two-stage deep learning framework for MURA-based X-ray analysis, aiming to improve body-part recognition and body-part-wise abnormality detection. Multiple architectures were first compared for seven-class body-part classification, after which the selected hybrid Xception-Swin model was fine-tuned for abnormality detection within each anatomical subset. The framework combines Xception-derived local structural features with Swin Transformer contextual features using attention-based fusion, and performance was evaluated using accuracy, F1-score, AUC-ROC, Cohen’s kappa, calibration, component-level ablation, post hoc explainability, and zero-shot FracAtlas validation. For body-part classification, the model achieved accuracy = 0.9643, macro F1 = 0.9574, AUC-ROC = 0.9963, and kappa = 0.9579. For abnormality detection, accuracy ranged from 0.7289 to 0.8538, F1 from 0.7191 to 0.8508, AUC from 0.7693 to 0.9080, and kappa from 0.4449 to 0.7071. Ablation on hand and humerus radiographs showed the highest macro F1 with Hybrid Attention, while FracAtlas validation yielded AUC = 0.8247 and kappa = 0.5812. The results support complementary CNN-Transformer fusion and indicate preliminary cross-dataset generalizability. Implementation resources are available at Zenodo. Full article

Respiratory diseases are a major global health challenge. However, identification of respiratory diseases is often limited by subjectivity, environmental noise and inter-clinician variability. This study presents an explainable multimodal deep learning framework for recording-level multiclass classification of respiratory audio signals. The proposed system integrates two complementary representations—a spectro-temporal encoder based on a CNN–BiLSTM-attention architecture and a handcrafted acoustic-feature encoder capturing acoustic descriptors commonly used in respiratory-audio analysis, including MFCCs, zero-crossing rate, spectral centroid, spectral bandwidth, chroma, RMS energy, and spectral rolloff features. These branches are combined through late-stage fusion to leverage both data-driven representation learning and domain-informed acoustic cues. The proposed model was trained and internally evaluated on the Asthma Detection Dataset Version 2, comprising five respiratory categories: bronchial disease, asthma, COPD, healthy, and pneumonia. Mono conversion, resampling to 16 kHz, 100–2000 Hz band-pass filtering, amplitude normalisation, fixed 4 s trimming or zero-padding, training-only augmentation, handcrafted-feature extraction, mel-spectrogram generation, quality control auditing, and stratified recording-level partitioning have been applied in the pre-processing steps. Across five repeated experiments with different random seeds, the proposed hybrid model achieved a mean held-out recording-level test accuracy of $0.9099\pm 0.0163$, balanced accuracy of $0.8936\pm 0.0152$, macro F1-score of $0.8937\pm 0.0177$, macro ROC–AUC of $0.9867\pm 0.0010$, and macro PR–AUC of $0.9489\pm 0.0044$. Conventional machine learning baseline comparisons showed that the proposed model achieved stronger internal accuracy, balanced accuracy, macro recall, macro F1-score, and macro ROC–AUC than classical machine learning algorithms trained on handcrafted acoustic features, although Random Forest remained competitive in macro PR–AUC. Ablation analysis shows that the deep spectro-temporal branch was the primary contributor to predictive performance, while the handcrafted branch provided complementary interpretable acoustic information rather than consistently improving all classification metrics. Explainability was incorporated using Grad-CAM and Integrated Gradients for spectrogram-based interpretation and SHAP for handcrafted-feature attribution. Domain-shift evaluation on the ICBHI Respiratory Sound Database and a COPD-focused cohort revealed substantial dataset shift effects, including poor healthy-case recognition on ICBHI and seed-dependent COPD recognition in the COPD-focused cohort. Identifier-aware sensitivity analyses showed lower performance than the main recording-level split, suggesting that subject-like or source-level overlap may inflate internal performance estimates. The findings should be interpreted as promising internal held-out recording-level algorithmic performance with limited external transfer, rather than evidence of readiness for clinical use. Full article

This study develops and evaluates a two-stage machine learning framework for forecasting the condominium price index of pre-pandemic market data of Frankfurt am Main, Germany, one quarter ahead. To the best of the author’s knowledge, it is the first study to combine German micro-level transaction and listing data, socioeconomic variables and macro-financial indicators in a single residential price-forecasting framework. Furthermore, it provides the first evidence on machine learning-based transaction price index forecasting in Germany. Methodologically, the framework links disaggregated and aggregate forecasting. In stage 1, prices per square metre are estimated for four market segments using ordinary least squares, random forest, extreme gradient boosting, and a stacked ensemble in a strictly out-of-sample expanding-window design. In stage 2, these predictions are combined with lagged index values and macro-financial indicators to forecast the city-wide index. The stage 2 model achieves a relative root mean squared error of 2.25% and a mean absolute percentage error of 1.85%, outperforming a naïve persistence benchmark by reducing root mean squared error by 23%. Model interpretation indicates that price persistence dominates stage 1, reflecting market inertia, while lagged macro-financial variables and location quality composition drive index forecasts, pointing to delayed financial market transmission and heterogeneous submarket dynamics. Full article

Vehicular Named Data Networking (VNDN) inherits the broadcast-oriented forwarding of NDN, which exposes safety messages to position-falsification attacks. Existing detectors rely either on static fuzzy thresholds, which drift as traffic patterns change, or on opaque deep models, which are accurate but uninterpretable to safety auditors. We propose a two-stage detector that combines an Adaptive Fuzzy Membership Tuning (AFMT) pre-filter with an attention-augmented bidirectional LSTM. AFMT is a Mamdani fuzzy classifier whose triangular membership-function parameters are updated online by gradient descent on a prediction-error feedback signal from the downstream BiLSTM, replacing offline-fixed thresholds. The BiLSTM consumes the fuzzy suspicion score as an extra feature and produces interpretable per-time-step attention weights aligned with attack onsets. On a simulator-synthesized VNDN benchmark following the five canonical VeReMi attack types, the detector attains F1-scores between 0.955 and 0.979 (macro-average 0.964), ties the strongest baselines on the hardest Random-Offset attack while achieving the highest ROC-AUC of all models (0.984), and runs in 0.44 ms per sample on a CPU. On a live OMNeT++/Veins/SUMO testbed running the five attacks on the LuST scenario, the detector attains an F1 value of 0.986. A leave-one-feature-out study shows that detection does not hinge on the Kalman plausibility feature, and on the real public VeReMi v1.0 dataset the architecture transfers to four of the five attack types at an F1 near 1.0, while the Constant Offset stays invisible to kinematics-only features, and this quantifies the value of the named-data-plane features. Every number reported here is measured from the running detector. Full article

Vibration sensor networks deployed on bridges continuously generate large volumes of unlabelled measurements under healthy operation, whereas labelled damage records on any specific target bridge remain extremely scarce—a chronic data asymmetry that constrains data-driven structural health monitoring (SHM). Existing remedies either require labelled source-bridge data or borrow augmentation pipelines and encoders from computer vision that are poorly matched to one-dimensional vibration signals. This study proposes a two-stage framework—siamese contrastive pretraining followed by few-shot fine-tuning on the target bridge—that learns environment-invariant representations from unlabelled source-side sensor signals and transfers them to a new bridge using only a handful of labelled samples. Three contributions are advanced: (i) a signal-domain augmentation policy that decouples sensor-level corruptions from operational-level fluctuations, including a frequency-band stochastic masking scheme designed to emulate cross-bridge perturbations; (ii) a one-dimensional self-calibrated convolutional encoder embedded in a stop-gradient siamese learner, providing the enlarged receptive field and inter-channel coupling required to capture sparse damage signatures in multi-sensor recordings; and (iii) a transferability analysis that formally links the contrastive invariance objective to a bound on the expected cross-bridge risk. On the Z24 benchmark and an in-house four-configuration laboratory bridge population, the method attains a 5-shot macro-F1 of 0.913 (Z24 → Lab) and 0.892 (Lab → Z24), outperforming eleven baselines by 3.4–37.1 percentage points. Full article

Digital infrastructure expansion in Europe raises a spatial planning problem: early-stage screening needs to compare regional conditions while also checking whether rankings remain stable when decision priorities change. This study evaluates 24 European Nomenclature of Territorial Units for Statistics level 2 (NUTS-2) regions for geo-distributed data center development. The 2022 decision matrix uses five Eurostat criteria: information and communications technology (ICT) specialists’ share in employment, average hourly labor cost, renewable electricity share, non-household electricity price and population density. Four criteria are national intensive proxies assigned to the selected NUTS-2 regions, while population density is directly observed at the NUTS-2 level. After a log₁₀ transformation of population density and min–max normalization, we compare the weighted sum model (WSM), TOPSIS and VIKOR across four weighting scenarios. We then apply a random-weighting audit based on Stochastic Multicriteria Acceptability Analysis (SMAA) principles, using 10,000 Dirichlet weight draws, followed by a local Dirichlet sensitivity analysis around the Balanced profile. Results show that the most stable high-performing profiles are not limited to the established FLAP-D market reference. Latvija (LV00), Stockholm (SE11), Helsinki-Uusimaa (FI1B), Eesti (EE00) and Área Metropolitana de Lisboa (PT17) form the main high-performing set across stochastic rank metrics, while several mature Western metropolitan regions remain more sensitive to cost and territorial-pressure criteria. The study provides a reproducible spatial decision support framework for macro-regional screening rather than micro-siting. Full article

Background/Objectives: Type 2 diabetes mellitus (T2DM) is a major global health challenge due to its high prevalence and association with chronic complications, highlighting the need for reliable predictive tools to support clinical decision-making. Methods: This study proposes a two-stage hierarchical prediction system based on a Random Forest (RF) classifier. In Stage 1, the model performs multiclass classification into healthy (H), T2DM without complications (D), and T2DM with complications (C). In Stage 2, patients classified as C are further stratified into microvascular or macrovascular complications. The dataset included 31 biochemical, molecular, inflammatory, and oxidative stress variables from Mexican and Spanish cohorts. Feature selection was performed using Pearson correlation, and feature relevance was further assessed using RF importance measures. Model training used stratified cross-validation, with additional evaluation on a hold-out set to approximate real-world performance. Results: The optimized RF achieved an accuracy of 92% and a macro F1-score of 0.92, outperforming baseline models, with an AUC-ROC of 0.89 for complication prediction. Key predictive features included IL-18, miR-126, duration of T2DM, HbA1c, and IL-10. Conclusions: The novelty of this study lies in integrating heterogeneous biomarkers within a hierarchical predictive framework, rather than in the machine learning algorithm itself. This multimodal approach, combined with interpretable machine learning techniques, is designed to deliver clinically meaningful insights for patient stratification and personalized management in T2DM. Full article

Low-resource hate speech detection remains a challenging task for Indonesian social media due to limited labeled annotations, highly informal linguistic expressions, and substantial lexical variability. Under such conditions, purely supervised transformer models often suffer from unstable semantic generalization, while conventional pseudo-labeling methods are vulnerable to noisy unlabeled sample propagation. To address these limitations, this study proposes a hybrid feature-enhanced IndoBERT framework integrated with a controlled semi-supervised learning strategy. The proposed model combines contextual IndoBERT embeddings with abusive lexicon cues, handcrafted linguistic indicators, and TF-IDF–SVD statistical representations through a lightweight concatenation–projection feature fusion mechanism, while unlabeled data are incorporated via adaptive confidence thresholding and class-balanced pseudo-label selection to improve pseudo-label reliability. Extensive experiments were conducted under realistic low-resource supervision settings using only 5%, 10%, and 20% labeled data, and the proposed framework was systematically compared against representative baselines, including sparse lexical machine learning models, shallow neural architectures, multilingual transformers, IndoBERTweet, naive pseudo-labeling, and LLM-based prompting. The results show that model effectiveness is strongly supervision-dependent. Under the most extreme low-resource setting, compact statistical augmentation provides the most stable complementary signal, whereas under moderate low-resource supervision, the full hybrid representation combined with controlled semi-supervised learning yields the strongest and most consistent gains. The proposed Hybrid IndoBERT + controlled SSL framework outperforms all baselines at the 20% labeled setting, reaching an accuracy of 0.8654, Macro-F1 of 0.8633, and ROC-AUC of 0.9334. Additional analyses of pseudo-label reliability, calibration behavior, computational efficiency, and qualitative error patterns further show that the proposed framework improves low-resource robustness while maintaining comparable inference-time efficiency. These findings demonstrate that low-resource hate speech detection benefits most from the staged integration of contextual semantic modeling, interpretable linguistic cues, global lexical–statistical structure, and carefully regulated unlabeled data exploitation. Additional experiments using GPT-4o-mini and Llama-3.1-8B further demonstrate that the proposed framework remains competitive against general-purpose large language model prompting approaches under low-resource Indonesian hate speech detection scenarios. The proposed framework provides a practical and reproducible direction for hate speech detection in annotation-constrained social media environments. Full article

Background/Objectives: The precise identification of Alzheimer’s disease (AD) stages through clinical data is crucial for early diagnosis and suitable therapy. This classification remains troublesome due to overlap in cognitive profiles across different phases of illness progression. This study presents a comprehensive and advanced diagnostic system, termed FLAME, featuring an enhanced federated learning architecture for privacy-preserving multi-institutional implementation. It provides a systematic review of machine learning (ML) and deep learning (DL) models for the classification of five stages of Alzheimer’s disease (AD). The models include cognitively normal (CN), subjective memory complaints (SMC), early mild cognitive impairment (EMCI), late mild cognitive impairment (LMCI), and Alzheimer’s disease (AD). Methods: Sixteen traditional machine learning models and eleven deep learning architectures—including FT-Transformer and NODE—were evaluated using a structured clinical dataset comprising 362 features. A hybrid ensemble was created at the probability level by combining the two top-performing models, LightGBM and a five-layer DNN. The weights of this ensemble were automatically optimised using a Genetic Algorithm (GA) with Macro-F1 as the fitness criterion, confirmed stable across 30 independent runs ($w^{★}=0.5024\pm 0.0001$). A federated learning architecture was then established, deploying the DNN across non-IID clients while keeping LightGBM centralised. We examine four distinct aggregation algorithms: FedAvg, FedProx, FedNova, and SCAFFOLD. Results: Among all deep learning architectures, FT-Transformer achieved the highest standalone performance (accuracy = 0.7810, $\kappa$ = 0.7081). The five-layer deep neural network (DNN) was selected as the DL representative for the hybrid ensemble. LightGBM attained superior machine learning performance (accuracy = 0.8156, $\kappa$ = 0.7537), confirmed deterministic across 10 seeds. The LightGBM vs. XGBoost difference is not statistically significant (McNemar $p=0.4227$). The GA-optimised hybrid ensemble (w = 0.685) surpassed both individual baselines across all evaluation metrics. The FedNova hybrid design achieved superior overall performance in federated configurations, surpassing all centralised arrangements in accuracy (accuracy = 0.8213, $\kappa$ 0.7614). Conclusions: Evolutionary ensemble optimisation combined with federated learning provides a robust, scalable, and privacy-preserving solution for AD stage classification, offering a clinically viable framework for real-world multi-institutional decision-support systems. However, the AD class remains severely under-recalled across all configurations (F1 ≤ 0.21), identifying this as the primary open challenge for clinical translation. Full article

Reusing industrial byproducts to prepare recycled aggregate concrete (RAC) is a sustainable approach that can protect the ecological environment. This study tested the possibility of preparing an eco-efficient recycled concrete containing steel slag (SS), ground-granulated blast-furnace slag (GGBS), and polypropylene (PP) fibers to avoid resource waste and depletion and decrease CO₂ emissions. To this end, 12 mix proportions were designed to analyze the effects of SS, GGBS, and PP fibers on the macro- and micro-performances of the developed RAC. The experimental results showed that increasing the SS content decreased the RAC mechanical strength, whereas partially substituting SS with GGBS in the RAC improved the mechanical properties, especially at a later stage. Adding PP fibers to the RAC containing SS and GGBS significantly increased the splitting tensile strength. However, it had little effect on the compressive strength as the PP fiber content was less than 0.6%. The microscopic experiment revealed that adding GGBS promoted the degree of hydration of SS, reduced the Ca (OH)₂ content, made the ITZ structure more compact, and optimized the pore characteristics of the RAC. Furthermore, according to the raw materials and results of mechanical properties, a hybrid Genetic Algorithm/Artificial Neural Network (GA-ANN) technique was proposed to predict the compressive strength of the RAC containing SS, GGBS, and PP fibers. We found that the proposed GA-ANN model effectively predicts the compressive strength. The findings of this study demonstrate that preparing RAC incorporating SS, GGBS, and PP fibers is promising for the reuse of industrial byproducts and construction waste. Full article

Regionally, compound agroecological stress arising from both natural and anthropogenic emergy inputs may influence maize transcriptomic responses; however, evidence across multiple scales remains limited. We developed a reproducible five-step framework integrating a macro-level compound stress index, molecular response modules, cross-scale coupling, spatial continuity, and independent field validation. Nine variables (emergy indicators ELR, Fn, and NEYR; climate; soil; and terrain) were PCA-weighted into a Composite Abiotic Stress Intensity Index (CASI; first three PCs = 83.7%; and prefecture-level Moran’s I = 0.463). Across 15 public RNA-seq datasets (286 samples), WGCNA identified five separable modules (drought–heat, reproductive stage heat, low nitrogen/phosphorus, osmotic salt, and chronic compound), 270 core genes, and four cross-module hubs (ZmDREB2A, ZmHSFA2, ZmWRKY33, and ZmNRT2.1). With n = 21, the sCCA (r₁ = 0.81, permutation p = 0.003; LOO-CV r = 0.71), random forest, and spatial error model all confirmed coupling between ELR and the drought–heat module (β = 0.51, p = 0.008). PLS-DA four-zone partitioning (Q² = 0.548) and a county-level second-order trend surface (R² = 0.67) verified spatial continuity. GSVA on five independent field RNA-seq datasets yielded 74.4 to 82.8% core gene directional consistency and Cliff’s δ of 0.59 to 0.68 (large effect), avoiding circular reasoning. The framework enables molecular analysis for precision agriculture and climate-resilient breeding. Full article

The Tajik Basin is located on the eastern edge of the Central Asian segment of the Tethyan tectonic domain. The basin underwent intense tectonic transformation during the Himalayan period, resulting in complex structural styles, unclear original sedimentary characteristics and oil and gas geological conditions, and a complex process of oil and gas accumulation, which restricts the further evaluation of the basin’s exploration potential. Studying the Tajik Basin in the macro background of the Tethys tectonic domain, the tectonic sedimentary evolution of the Tethys tectonic domain has a significant effect on the basin’s tectonic evolution, sedimentary characteristics, and oil and gas accumulation conditions. The Tajik Basin has gone through four stages of tectonic evolution: the Late Permian to Triassic was the stage of back arc foreland basin; the Jurassic period was the stage of back arc extensional faulting depression; the Cretaceous–Paleogene period was the stage of depression basins; and the Neogene is the stage of the regenerated foreland basins. Through field geological surveys and analysis of outcrop samples, it has been determined that the Tajik Basin has developed three sets of source rocks: the Middle and Lower Jurassic, Cretaceous, and Paleogene. Among them, the organic matter abundance of the Middle and Lower Jurassic is relatively high, most of them are in the mature stage, and they are primarily gas-generating source rocks. The Cretaceous and Paleogene source rocks are mainly oil generating and in a low-mature state. There are four sets of reservoirs developed in the Tajik Basin: Middle-Upper Jurassic carbonate rocks, Lower Cretaceous clastic rocks, Upper Cretaceous carbonate rocks and Paleogene carbonate rocks. Comprehensive research shows that the Tajik Basin mainly develops three types of oil and gas reservoirs: Jurassic carbonate gas reservoirs, distributed in the southwestern Gissar Uplift and Surhan Depression in the western part of the basin; Paleogene carbonate reservoirs, distributed in the southern Vakhsh Depression and the eastern Kuliabu Depression; and multi layer–multi lithology oil and gas reservoirs, distributed in the northern Dushanbe Depression. The primary controlling factor for the three types of oil and gas reservoirs is tectonic movement, which forms traps and simultaneously reshapes the reservoirs, ultimately leading to effective accumulation of oil and gas. The distribution of oil and gas in the Tajik Basin is characterized by “west gas and east oil, west more and east less, west pre-salt and east post-salt, and pre-salt gas and post-salt oil”. Affected by the regional tectonic movements of the Tethys rich oil and gas tectonic domain, the basin has high-quality hydrocarbon source rocks, reservoirs, and cap rock conditions. The pre-salt Jurassic has the potential to form large natural gas reservoirs, while the post-salt Cretaceous and Paleogene still have further potential for exploration. Full article