Logistic Classification Model Research Articles

The role of Allee effects for Gaussian and Lévy dispersals in an environmental niche.

In the study of biological populations, the Allee effect detects a critical density below which the population is severely endangered and at risk of extinction. This effect supersedes the classical logistic model, in which low densities are favorable due to lack of competition, and includes situations related to deficit of genetic pools, inbreeding depression, mate limitations, unavailability of collaborative strategies due to lack of conspecifics, etc. The goal of this paper is to provide a detailed mathematical analysis of the Allee effect. After recalling the ordinary differential equation related to the Allee effect, we will consider the situation of a diffusive population. The dispersal of this population is quite general and can include the classical Brownian motion, as well as a Lévy flight pattern, and also a "mixed" situation in which some individuals perform classical random walks and others adopt Lévy flights (which is also a case observed in nature). We study the existence and nonexistence of stationary solutions, which are an indication of the survival chance of a population at the equilibrium. We also analyze the associated evolution problem, in view of monotonicity in time of the total population, energy consideration, and long-time asymptotics. Furthermore, we also consider the case of an "inverse" Allee effect, in which low density populations may access additional benefits.

Exploring miRNAs' Based Modeling Approach for Predicting PIRA in Multiple Sclerosis: A Comprehensive Analysis.

The current hypothesis on the pathophysiology of multiple sclerosis (MS) suggests the involvement of both inflammatory and neurodegenerative mechanisms. Disease Modifying Therapies (DMTs) effectively decrease relapse rates, thus reducing relapse-associated disability in people with MS. In some patients, disability progression, however, is not solely linked to new lesions and clinical relapses but can manifest independently. Progression Independent of Relapse Activity (PIRA) significantly contributes to long-term disability, stressing the urge to unveil biomarkers to forecast disease progression. Twenty-five adult patients with relapsing-remitting multiple sclerosis (RRMS) were enrolled in a cohort study, according to the latest McDonald criteria, and tested before and after high-efficacy Disease Modifying Therapies (DMTs) (6-24 months). Through Agilent microarrays, we analyzed miRNA profiles from peripheral blood mononuclear cells. Multivariate logistic and linear models with interactions were generated. Robustness was assessed by randomization tests in R. A subset of miRNAs, correlated with PIRA, and the Expanded Disability Status Scale (EDSS), was selected. To refine the patient stratification connected to the disease trajectory, we computed a robust logistic classification model derived from baseline miRNA expression to predict PIRA status (AUC = 0.971). We built an optimal multilinear model by selecting four other miRNA predictors to describe EDSS changes compared to baseline. Multivariate modeling offers a promising avenue to uncover potential biomarkers essential for accurate prediction of disability progression in early MS stages. These models can provide valuable insights into developing personalized and effective treatment strategies.

Assessing risk factors of bypass graft surgery through the implementation of Bayesian and non-Bayesian methodologies

The aim of this research is to determine the key risk factors that influence the need for Bypass Graft surgery (BGS). A cohort of one hundred adult patients underwent coronary angiography operations at the Erbil Cardiac Center from January 2016 to December 2020. Multiple logistic regression (MLR) analysis was used to identify the determinants of BGS. Upon detailed statistical analysis, a significant association with BGS was found for several variables, including Age, BMI, Eosinophil, WBC, DBP, HbA1c, MCH, Blood Sugar, PLT, and T3. It was found that the significant variables contributing the most to predicting the likelihood of Bypass Graft surgery through Bayesian logistic regression analysis were Age, WBC, and MCH. These variables were identified as the pivotal risk factors associated with the probability of undergoing BGS. Age, DBP, and Blood Sugar had minimal influence on the likelihood of Bypass surgery. However, for each unit increase in WBC, the odds of surgery reduced by 38 %, and for MCH, the odds reduced by 19 %. DBP and Blood Sugar had little effect on Bypass surgery (Bayesian odds ratio close to 1). This study revealed that Bayesian logistic models offer superiority over classical logistic models. The Bayesian-based models provide more accurate predictions by incorporating informative prior distributions. To improve the accuracy and generalizability of the findings, future research should consider a larger sample size and incorporate informative prior distributions within the Bayesian framework. Also, machine learning algorithms can offer new insights into BGS outcomes by considering various variables and assessing model sensitivity.

Analysis of the evolution of modern Chinese history based on data mining

Abstract In this paper, data mining is proposed to study the evolution of China’s modern history by addressing the problem of incomplete content of the evolution process. The data mining technique mainly preprocesses the data set of Chinese modern history by logistic regression algorithm, and its purpose is to detect the accuracy of the data so as to provide accurate and high-quality data for the data mining process. The process of visualization using information related to the evolution of modern Chinese history and the influence of modern Chinese historical events is applied to the visualization analysis, and the final influence of the evolutionary development of modern Chinese history is obtained and saved to the database by weighting and summing the influence factors of modern Chinese historical figures. The logistic regression algorithm uses modern historical persons and things as input data, and the weights of modern historical persons and things are the predictions carried out by classification. The results show that the highest accuracy is 0.67 when the threshold value is set to 1. The logistic classification model predicts better for the case of weight type 2 of modern Chinese history people and weight type 6 of modern history things. This study makes a certain contribution to the study of modern history so that the study of modern history can gradually move toward completeness and objectivity.

Radiomic-Based Prediction of Lesion-Specific Systemic Treatment Response in Metastatic Disease.

Cancer is a heterogenous, dynamic target. Individual patients, lesions, and cell populations may present with different characteristics and varying sensitivities to anticancer therapies. The purpose of this study was to investigate the utility of image-derived (radiomic) biomarkers for prediction of lesion-specific treatment response. Using a database of serial imaging from a Phase III clinical trial (NCT01440088), all visible lung metastases were contoured on chest computed tomography performed at baseline and after 2 cycles of chemotherapy. Metastatic leiomyosarcoma patients received either doxorubicin monotherapy (DM) or doxorubicin plus evofosfamide (DE). Percent change in volume with respect to baseline was evaluated as a metric of treatment response. Lesions that experienced a positive volume change greater than 50% were labeled as 'progressors'; all other lesions were labeled as 'non-progressors'. Radiomic features were extracted at baseline from the whole lesion and three peritumoral regions using PyRadiomics (version 3.0.1). Features were compared using the Mann-Whitney U-test and further used to classify between response categories. Logistic classification models were evaluated with 10-fold cross-validation using Area Under the Curve for the Receiver Operating Characteristic (AUROC) and Precision-Recall (AUPRC) curves as measures of model performance. Significance was evaluated by permutation tests. All code used to derive features is open source and available for download. In total, 204 lung metastases from 80 patients were evaluated. The fraction of progressing lesions was 19% and 27% for the DM and DE regimes, respectively. Differences in individual lesion response were observed in 20% of patients with multiple lung metastases. Significant differences in entropy were observed between response categories (false discovery rate < 5% after multiple testing correction): interior regions for lesions receiving DM, and peripheral regions for lesions receiving DE. The best performing models for the DM and DE regimens had AUROC of 0.87 and 0.76 and AUPRC 0.60 and 0.50, respectively (p < 0.05). Predicting lesion-specific responses using radiomic features represents a paradigm shift. Lesion-specific radiomic models indicate a 2 to 3-fold increase in predictive capacity in comparison to a no-skill classifier. These models, although preliminary, achieved a strong predictive value and could be used to predict lesion-specific treatment response.

ВЫБОР МЕТОДОВ СТАТИСТИЧЕСКОЙ ОБРАБОТКИ РЕЗУЛЬТАТОВ РАДИОМИЧЕСКОГО АНАЛИЗА КТ-ИЗОБРАЖЕНИЙ ОПУХОЛЕЙ ГОЛОВЫ И ШЕИ

Purpose: Selection of the optimal method for statistical processing of the results of texture analysis of conventional CT images in patients with head and neck tumors. Material and methods: A total of 118 patients aged from 4 to 80 years with a verified diagnosis of 37 benign and 81 malignant head and neck tumors were studied. Texture analysis was performed using LIFEx program, version 7.10, with statistical processing using SPSS, MedCalc, XLSTAT, R. Results: The 39 texture indicators extracted from CT images were subjected to statistical processing by different methods, including Mann-Whitney U test, correlation matrix, factor analysis, LASSO-regression, ending with the development of a logistic classification model. Of the multiple processing methods, LASSO-regression followed by logistic model was optimal; according to its results, the percentage of correct classification of benign and malignant patient groups was – 81.3 %, area under the ROC curve was 0.902±0.029 (p&lt;0.0001), sensitivity – 82.7 %, specificity – 87.5 %. Conclusion: Texture analysis of medical images allows non-invasive prediction of benign or malignant nature of the imaged head and neck mass. The choice of the correct method for statistical processing of texture analysis results is critical to assess and classify patients according to the nature of the tumor.

Establishment and Validation of Predictive Model of Tophus in Gout Patients

(1) Background: A tophus is a clinical manifestation of advanced gout, and in some patients could lead to joint deformities, fractures, and even serious complications in unusual sites. Therefore, to explore the factors related to the occurrence of tophi and establish a prediction model is clinically significant. (2) Objective: to study the occurrence of tophi in patients with gout and to construct a predictive model to evaluate its predictive efficacy. (3) Methods: The clinical data of 702 gout patients were analyzed by using cross-sectional data of North Sichuan Medical College. The least absolute shrinkage and selection operator (LASSO) and multivariate logistic regression were used to analyze predictors. Multiple machine learning (ML) classification models are integrated to analyze and identify the optimal model, and Shapley Additive exPlanations (SHAP) interpretation was developed for personalized risk assessment. (4) Results: Compliance of urate-lowering therapy (ULT), Body Mass Index (BMI), course of disease, annual attack frequency, polyjoint involvement, history of drinking, family history of gout, estimated glomerular filtration rate (eGFR), and erythrocyte sedimentation rate (ESR) were the predictors of the occurrence of tophi. Logistic classification model was the optimal model, test set area under curve (AUC) (95% confidence interval, CI): 0.888 (0.839-0.937), accuracy: 0.763, sensitivity: 0.852, and specificity: 0.803. (5) Conclusions: We constructed a logistic regression model and explained it with the SHAP method, providing evidence for preventing tophus and guidance for individual treatment of different patients.

Model fitting of the seasonal population dynamics of the soybean aphid, Aphis glycines Matsumura, in the field

The soybean aphid, Aphis glycines Matsumura (Hemiptera: Aphididae), is one of the greatest threats to soybean production, and both trend analysis and periodic analysis of its population dynamics are important for integrated pest management (IPM). Based on systematically investigating soybean aphid populations in the field from 2018 to 2020, this study adopted the inverse logistic model for the first time, and combined it with the classical logistic model to describe the changes in seasonal population abundance from colonization to extinction in the field. Then, the increasing and decreasing phases of the population fluctuation were divided by calculating the inflection points of the models, which exhibited distinct seasonal trends of the soybean aphid populations in each year. In addition, multifactor logistic models were then established for the first time, in which the abundance of soybean aphids in the field changed with time and relevant environmental conditions. This model enabled the prediction of instantaneous aphid abundance at a given time based on relevant meteorological data. Taken as a whole, the successful approaches implemented in this study could be used to build a theoretical framework for practical IPM strategies for controlling soybean aphids.

Modelling factors associated with therapeutic inertia in hypertensive patients: Analysis using repeated data from a hospital registry in West Africa.

The proportion of poorly controlled hypertensives still remains high in the general African population. This is largely due to therapeutic inertia (TI), defined as the failure to intensify or modify treatment in a patient with poorly controlled blood pressure (BP). The objective of this study was to identify the determinants of TI. We conducted a retrospective cohort study from March 2012 to February 2014 of hypertensive patients followed during 4 medical visits. The TI score was the number of visits with TI divided by the number of visits where a therapeutic change was indicated. A random-effects logistic model was used to identify the determinants of TI. A total of 200 subjects were included, with a mean age of 57.98 years and 67% men. The TI score was measured at 85.57% (confidence interval [CI] 95% = [82.41-88.92]). Measured individual heterogeneity was significantly significant (0.78). Three factors were associated with treatment inertia, namely the number of antihypertensive drugs (odd ratios [OR] = 1.27; CI = [1.02-1.58]), the time between consultations (OR = 0.94; CI = [0.91-0.97]), and treatment noncompliance (OR = 15.18; CI = [3.13-73.70]). The random-effects model performed better in predicting high-risk patients with TI than the classical logistic model (P value < .001). Our study showed a high TI score in patients followed in cardiology in Burkina Faso. Reduction of the TI score through targeted interventions is necessary to better control hypertension in our cohort patients.

Serum metabolic profiling of coal worker's pneumoconiosis using untargeted lipidomics.

In this work, untargeted lipidomics was employed to analyze the effects of coal dust exposure on serum metabolite profiles. Furthermore, the potential of differential metabolites as novel biomarkers for diagnosis was investigated by binary logistic classification model. Nineteen differential metabolites were found among the three groups. The compounds were enriched in pathways associated with linoleic acid metabolism and pyrimidine metabolism. Fifty-three differential metabolites were found in coal dust-exposed people and CWP patients, and they were mainly enriched in glycerophospholipid metabolism. Three differential metabolites were correlated with lung function values. The diagnostic model, composed of lysoPI (16:0/0:0), bilirubin, and lysoPC (24:1/0:0), showed strong discrimination ability between dust-exposed people and CWP patients. The sensitivity, specificity, and AUC values of the model were 0.869, 0.600, and 0.750, respectively. The results suggest that coal worker's pneumoconiosis causes abnormal lipid metabolism in the body. A diagnostic model may aid current CWP diagnostic methods, and lysoPI (16:0/0:0), bilirubin, and lysoPC (24:1/0:0) can be used as potential CWP biomarkers. Further study is warranted to validate the findings in larger populations.

ML-Based Texture and Wavelet Features Extraction Technique to Predict Gastric Mesothelioma Cancer.

Microsatellites are small, repetitive sequences found all across the human genome. Microsatellite instability is the phenomenon of variations in the length of microsatellites induced by the insertion or deletion of repeat units in tumor tissue (MSI). MSI-type stomach malignancy has distinct genetic phenotypes and clinic pathological characteristics, and the stability of microsatellites influences whether or not patients with gastric mesothelioma react to immunotherapy. As a result, determining MSI status prior to surgery is critical for developing treatment options for individuals with gastric cancer. Traditional MSI detection approaches need immunological histochemistry and genetic analysis, which adds to the expense and makes it difficult to apply to every patient in clinical practice. In this study, to predict the MSI status of gastric cancer patients, researchers used image feature extraction technology and a machine learning algorithm to evaluate high-resolution histopathology pictures of patients. 279 cases of raw data were obtained from the TCGA database, 442 samples were obtained after preprocessing and upsampling, and 445 quantitative image features, including first-order statistics of impressions, texture features, and wavelet features, were extracted from the histopathological images of each sample. To filter the characteristics and provide a prediction label (risk score) for MSI status of gastric cancer, Lasso regression was utilized. The predictive label's classification performance was evaluated using a logistic classification model, which was then coupled with the clinical data of each patient to create a customized nomogram for MSI status prediction using multivariate analysis.

Kinetics of Bacterial Adaptation, Growth, and Death at Didecyldimethylammonium Chloride sub-MIC Concentrations.

Minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) are standard indexes for determining disinfection effectiveness. Nevertheless, they are static values disregarding the kinetics at sub-MIC concentrations where adaptation, growth, stationary, and death phases can be observed. The understanding of these dynamic mechanisms is crucial to designing effective disinfection strategies. In this study, we studied the 48 h kinetics of Bacillus cereus and Escherichia coli cells exposed to sub-MIC concentrations of didecyldimethylammonium chloride (DDAC). Two mathematical models were employed to reproduce the experiments: the only-growth classical logistic model and a mechanistic model including growth and death dynamics. Although both models reproduce the lag, exponential and stationary phases, only the mechanistic model is able to reproduce the death phase and reveals the concentration dependence of the bactericidal/bacteriostatic activity of DDAC. This model could potentially be extended to study other antimicrobials and reproduce changes in optical density (OD) and colony-forming units (CFUs) with the same parameters and mechanisms of action.

Predicting the Risk of Mortality in Children using a Fuzzy-Probabilistic Hybrid Model.

Introduction The mortality risk in children admitted to Pediatric Intensive Care Units (PICU) is usually estimated by means of validated scales, which only include objective data among their items. Human perceptions may also add relevant information to prognosticate the risk of death, and the tool to use this subjective data is fuzzy logic. The objective of our study was to develop a mathematical model to predict mortality risk based on the subjective perception of PICU staff and to evaluate its accuracy compared to validated scales. Methods A prospective observational study in two PICUs (one in Spain and another in Latvia) was performed. Children were consecutively included regardless of the cause of admission along a two-year period. A fuzzy set program was developed for the PICU staff to record the subjective assessment of the patients' mortality risk expressed through a short range and a long range, both between 0% and 100%. Pediatric Index of Mortality 2 (PIM2) and Therapeutic Intervention Scoring System 28 (TISS28) were also prospectively calculated for each patient. Subjective and objective predictions were compared using the logistic regression analysis. To assess the prognostication ability of the models a stratified B-random K-fold cross-validation was performed. Results Five hundred ninety-nine patients were included, 308 in Spain (293 survivors, 15 nonsurvivors) and 291 in Latvia (282 survivors, 9 nonsurvivors). The best logistic classification model for subjective information was the one based on MID (midpoint of the short range), whereas objective information was the one based on PIM2. Mortality estimation performance was 86.3% for PIM2, 92.6% for MID, and the combination of MID and PIM2 reached 96.4%. Conclusions Subjective assessment was as useful as validated scales to estimate the risk of mortality. A hybrid model including fuzzy information and probabilistic scales (PIM2) seems to increase the accuracy of prognosticating mortality in PICU.

Robust estimates of regional treatment effects in multiregional randomized clinical trials with semiparametric logistic model.

In multiregional randomized clinical trials (MRCTs), determining the regional treatment effect of a new treatment over an existing one is important to both the sponsor and related regulatory agencies. Also of particular interest is to test the null hypothesis that the treatment benefit is the same among all the regions. Existing methods are mainly for continuous endpoint and use parametric models, which are not robust. MRCTs are known for facing increased variation and heterogeneity and a robust model for its design and analysis would be desirable. We consider clinical trials with a binary primary endpoint and propose a robust semiparametric logistic model which has a known parametric and an unknown nonparametric component. The parametric component represents our prior knowledge about the model, and the nonparametric part reflects uncertainty. Compared to the classic logistic model for this problem, the proposed model has the following advantages: robust to model assumption, more flexible and accurate to model the relationship between the response and covariates, and possibly more accurate parameter estimates. The model parameters are estimated by profile maximum likelihood approach, and the null hypothesis of regional treatment difference being the same is tested by the profile likelihood ratio statistic. Asymptotic properties of the estimates are derived. Simulation studies are conducted to evaluate the performance of the proposed model, which demonstrated clear advantages over the classic logistic model. The method is then applied to analyzing a real MRCT.

Observational Characterization of the Ecological and Environmental Features Associated with the Presence of Oropouche Virus and the Primary Vector Culicoides paraenesis: Data Synthesis and Systematic Review.

Oropouche virus (OROV), a member of the Orthobunyavirus genus, is an arthropod-borne virus (arbovirus) and is the etiologic agent of human and animal disease. The primary vector of OROV is presumed to be the biting midge, Culicoides paraenesis, though Culex quinquefasciatus, Cq. venezuelensis, and Aedes serratus mosquitoes are considered secondary vectors. The objective of this systematic review is to characterize locations where OROV and/or its primary vector have been detected. Synthesis of known data through review of published literature regarding OROV and vectors was carried out through two independent searches: one search targeted to OROV, and another targeted towards the primary vector. A total of 911 records were returned, but only 90 (9.9%) articles satisfied all inclusion criteria. When locations were characterized, some common features were noted more frequently than others, though no one characteristic was significantly associated with presence of OROV using a logistic classification model. In a separate correlation analysis, vector presence was significantly correlated only with the presence of restingas. The lack of significant relationships is likely due to the paucity of data regarding OROV and its eco-epidemiology and highlights the importance of continued focus on characterizing this and other neglected tropical diseases.

A machine learning model pipeline for detecting wet pavement condition from live scenes of traffic cameras

Highway safety is largely influenced by weather conditions that have become increasingly volatile due to the climate change. It well known that wet pavement significantly reduces surface friction, leading to inflated collision risk. Thus, timely knowledge of the road surface condition is critical for safe driving. In this paper, a novel machine learning model pipeline is proposed to detect the wetness of pavement based on live images of highway scenes captured by publicly accessible traffic cameras. To simplify the learning task, we finetuned the state-of-the-art instance segmentation baseline models to extract background instance targets, including pavement, sky, and vegetation, which are common in highway scenes. Then, the color mixture attributes in HSV (hue, saturation and value) of each segmented instance were extracted and used as visual cues for inferring pavement condition. Finally, gradient boosting ensemble classifiers are constructed and trained using the HSV features to predict the wetness of pavement. For the segmentation task, we leveraged Detectron2 baseline models (Mask R-CNN) and evaluated three backbone networks: R50-FPN, R101-FPN, and X101-FPN. For the classification task, two most popular gradient boosting algorithms (XGBoost and CatBoost) were evaluated together with a classic logistic model. Based on experiments with our custom dataset, the best performance (F1 score: 0.927, AUC: 0.975) was achieved by the R101-FPN backbone coupled with the CatBoost classifier.

Physics-guided logistic classification for tool life modeling and process parameter optimization in machining

This paper describes a physics-guided logistic classification method for tool life modeling and process parameter optimization in machining. Tool life is modeled using a classification method since the exact tool life cannot be measured in a typical production environment where tool wear can only be directly measured when the tool is replaced. In this study, laboratory tool wear experiments are used to simulate tool wear data normally collected during part production. Two states are defined: tool not worn (class 0) and tool worn (class 1). The non-linear reduction in tool life with cutting speed is modeled by applying a logarithmic transformation to the inputs for the logistic classification model. A method for interpretability of the logistic model coefficients is provided by comparison with the empirical Taylor tool life model. The method is validated using tool wear experiments for milling. Results show that the physics-guided logistic classification method can predict tool life using limited datasets. A method for pre-process optimization of machining parameters using a probabilistic machining cost model is presented. The proposed method offers a robust and practical approach to tool life modeling and process parameter optimization in a production environment.

Ultimate COVID-19 Detection Protocol Based on Saliva Sampling and qRT-PCR with Risk Probability Assessment

In the era of COVID-19 outbreak, various efforts are undertaken to develop a quick, easy, inexpensive, and accurate way for diagnosis. Although many commercial diagnostic kits are available, detailed scientific evaluation is lacking, making the public vulnerable to fear of false-positive results. Moreover, current tissue sampling method from respiratory tract requires personal contact of medical staff with a potential asymptomatic SARSCOV-2 carrier and calls for safe and less invasive sampling method. Here, we have developed a convenient detection protocol for SARS-COV-2 based on a non-invasive saliva self-sampling method by extending our previous studies on development of a laboratory-safe and low-cost detection protocol based on qRT-PCR. We tested and compared various self-sampling methods of self-pharyngeal swab and self-saliva sampling from non-carrier volunteers. We found that the self-saliva sampling procedure gave expected negative results from all of the non-carrier volunteers within 2 hours, indicating cost-effectiveness, speed and reliability of the saliva-based method. For an automated assessment of the sampling quality and degree of positivity for COVID-19, we developed scalable formulae based on a logistic classification model using both cycle threshold and melting temperature from the qRT-PCR results. Our newly developed protocol will allow easy sampling and spatial-separation between patient and experimenter for guaranteed safety. Furthermore, our newly established risk assessment formula can be applied to a large-scale diagnosis in health institutions and agencies around the world.

The value of meteorological data in marine risk assessment

The objective of this paper is to investigate whether the use of meteorological data improves the prediction of marine incidents, as represented by marine insurance claims for a vessel's voyage, both on a stand-alone basis and when combined with vessel-specific features and ship tracking data from the Automated Identification System (AIS). Furthermore, the paper investigates whether predictive performance improves when using machine learning algorithms, such as logistic LASSO regression and eXtreme Gradient Boosted Trees over classical logistic models, and identify dependencies and interaction effects among the risk factors within the SHapley Additive exPlanations framework. The data sample includes weather and AIS data for 42,000 voyages in the North Pacific between January 2013 and August 2019. The results suggest that meteorological information adds value in claims prediction and that short-term complex interactions between the vessel and weather conditions impact marine risk. The research is important for the improved modelling of marine risk on the basis of high-frequency, high-resolution ship tracking and weather data.

Explainable machine learning for materials discovery: predicting the potentially formable Nd-Fe-B crystal structures and extracting the structure-stability relationship.

New Nd-Fe-B crystal structures can be formed via the elemental substitution of LA-T-X host structures, including lanthanides (LA), transition metals (T) and light elements, X = B, C, N and O. The 5967 samples of ternary LA-T-X materials that are collected are then used as the host structures. For each host crystal structure, a substituted crystal structure is created by substituting all lanthanide sites with Nd, all transition metal sites with Fe and all light-element sites with B. High-throughput first-principles calculations are applied to evaluate the phase stability of the newly created crystal structures, and 20 of them are found to be potentially formable. A data-driven approach based on supervised and unsupervised learning techniques is applied to estimate the stability and analyze the structure-stability relationship of the newly created Nd-Fe-B crystal structures. For predicting the stability for the newly created Nd-Fe-B structures, three supervised learning models: kernel ridge regression, logistic classification and decision tree model, are learned from the LA-T-X host crystal structures; the models achieved maximum accuracy and recall scores of 70.4 and 68.7%, respectively. On the other hand, our proposed unsupervised learning model based on the integration of descriptor-relevance analysis and a Gaussian mixture model achieved an accuracy and recall score of 72.9 and 82.1%, respectively, which are significantly better than those of the supervised models. While capturing and interpreting the structure-stability relationship of the Nd-Fe-B crystal structures, the unsupervised learning model indicates that the average atomic coordination number and coordination number of the Fe sites are the most important factors in determining the phase stability of the new substituted Nd-Fe-B crystal structures.