Supervised Machine Learning Classification Research Articles

A new fast ADMM for kernelless SVM classifier with truncated fraction loss

Support vector machine (SVM) is one of well-known supervised machine learning classifier and is used widely in image classification, pattern recognition, disease diagnosis, etc. However, the high computational complexity is a major issue for large-scale classification problems. To address this issue, we propose a new truncated fraction loss function to construct a new sparse and robust SVM model (Ltf-SVM). By newly developed support vectors and a working set based on the proximal operator of truncated fraction loss, we prove the sparsity and robustness to outliers of Ltf-SVM from a theoretical perspective. When it comes to solve Ltf-SVM problems, we develop an alternating direction method of multipliers with a working set for addressing Ltf-SVM. The proposed algorithm is proven to converge globally and enjoys very low computational complexity. Moreover, numerical experiments verify its superior performance with respect to its high accuracy and high computational speed in comparison with other leading solvers.

Machine learning classification of polycystic ovary syndrome based on radial pulse wave analysis.

Patients with Polycystic ovary syndrome (PCOS) experienced endocrine disorders that may present vascular function changes. This study aimed to classify and predict PCOS by radial pulse wave parameters using machine learning (ML) methods and to provide evidence for objectifying pulse diagnosis in traditional Chinese medicine (TCM). A case-control study with 459 subjects divided into a PCOS group and a healthy (non-PCOS) group. The pulse wave parameters were measured and analyzed between the two groups. Seven supervised ML classification models were applied, including K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Decision Trees, Random Forest, Logistic Regression, Voting, and Long Short Term Memory networks (LSTM). Parameters that were significantly different were selected as input features and stratified k-fold cross-validations training was applied to the models. There were 316 subjects in the PCOS group and 143 subjects in the healthy group. Compared to the healthy group, the pulse wave parameters h3/h1 and w/t from both left and right sides were increased while h4, t4, t, As, h4/h1 from both sides and right t1 were decreased in the PCOS group (P < 0.01). Among the ML models evaluated, both the Voting and LSTM with ensemble learning capabilities, demonstrated competitive performance. These models achieved the highest results across all evaluation metrics. Specifically, they both attained a testing accuracy of 72.174% and an F1 score of 0.818, their respective AUC values were 0.715 for the Voting and 0.722 for the LSTM. Radial pulse wave signal could identify most PCOS patients accurately (with a good F1 score) and is valuable for early detection and monitoring of PCOS with acceptable overall accuracy. This technique can stimulate the development of individualized PCOS risk assessment using mobile detection technology, furthermore, gives physicians an intuitive understanding of the objective pulse diagnosis of TCM. Not applicable.

Impact of Satellite-Derived Land Cover Resolution Using Machine Learning and Hydrological Simulations

This study carefully assesses the capability of supervised machine learning classification algorithms in identifying land cover (LC) in the context of the Jhelum River basin in Kashmir. Sentinel 2 and Landsat 8 high-resolution data from two satellite sources were used. Through preprocessing techniques, we removed any potential noise inherent to satellite imagery and assured data consistency. The study then utilized and compared the skills of the supervised algorithms random forest (RF) and support vector machine (SVM). A hybrid approach, amalgamating classifications from both methods, was also tested for potential synergistic enhancements in accuracy. Using a stratified random sampling approach for validation, the SVM algorithm emerged with a commendable accuracy rate of 82.5%. Using simulations from 2000 to 2015, the soil and water assessment tool (SWAT) model was used to further explore the hydrological effects of LC alterations. Between 2009 and 2019, there were discernible changes in the land cover, with a greater emphasis on ranges, forests, and agricultural plains. When these changes were combined with the results of the hydrologic simulation, a resultant fall in average annual runoff—from above 700 mm to below 600 mm—was seen. With runoff values possibly ranging between 547 mm and 747 mm, the statistics emphasize the direct effects of urban communities encroaching upon forest, agricultural, and barren lands. This study concludes by highlighting the crucial role that technical pipelines play in enhancing LC classifications and by providing suggestions for future water resource estimation and hydrological impact evaluations.

Validation and interpretation of a multimodal drowsiness detection system using explainable machine learning

Background and objectiveDrowsiness behind the wheel is a major road safety issue with efforts focused on developing drowsy driving detection systems. However, most drowsy driving detection studies using physiological signals have focused on developing a 'black box' machine learning classifier, with much less focus on 'robustness' and 'explainability'—two crucial properties of a trustworthy machine learning model. Therefore, this study has focused on using multiple validation techniques to evaluate the overall performance of such a system using multiple supervised machine learning-based classifiers and then unbox the black box model using explainable machine learning. MethodsDriving was simulated via a 30-minute psychomotor vigilance task while the participants reported their level of subjective sleepiness with their physiological signals: electroencephalogram (EEG), electrooculogram (EOG) and electrocardiogram (ECG) being recorded. Six different techniques, comprising subject-dependent and independent techniques were applied for model validation and robustness testing with three supervised machine learning classifiers, namely K-nearest neighbours (KNN), support vector machines (SVM) and random forest (RF), and two explainable methods, namely SHapley Additive exPlanation (SHAP) analysis and partial dependency analysis (PDA) were leveraged for model interpretation. ResultsThe study identified the leave one participant out, a subject-independent validation technique to be most useful, with the best sensitivity of 70.3 %, specificity of 82.2 %, and an accuracy of 80.1 % using the random forest classifier in addressing the autocorrelation issue due to inter-individual differences in physiological signals. Moreover, the explainable results suggest most important physiological features for drowsiness detection, with a clear cut-off in the decision boundary. ConclusionsThe implication of the study will ensure a rigorous validation for robustness testing and an explainable machine learning approach to developing a trustworthy drowsiness detection system and enhancing road safety. The explainable machine learning-based results show promise in real-life deployment of the physiological-signal based in-vehicle trustworthy drowsiness detection system, with higher reliability and explainability, along with a lower system cost.

Diagnosis of Alzheimer's disease via resting-state EEG: integration of spectrum, complexity, and synchronization signal features.

Alzheimer's disease (AD) is the most common neurogenerative disorder, making up 70% of total dementia cases with a prevalence of more than 55 million people. Electroencephalogram (EEG) has become a suitable, accurate, and highly sensitive biomarker for the identification and diagnosis of AD. In this study, a public database of EEG resting state-closed eye recordings containing 36 AD subjects and 29 normal subjects was used. And then, three types of signal features of resting-state EEG, i.e., spectrum, complexity, and synchronization, were performed by applying various signal processing and statistical methods, to obtain a total of 18 features for each signal epoch. Next, the supervised machine learning classification algorithms of decision trees, random forests, and support vector machine (SVM) were compared in categorizing processed EEG signal features of AD and normal cases with leave-one-person-out cross-validation. The results showed that compared to normal cases, the major change in EEG characteristics in AD cases was an EEG slowing, a reduced complexity, and a decrease in synchrony. The proposed methodology achieved a relatively high classification accuracy of 95.65, 95.86, and 88.54% between AD and normal cases for decision trees, random forests, and SVM, respectively, showing that the integration of spectrum, complexity, and synchronization features for EEG signals can enhance the performance of identifying AD and normal subjects. This study recommended the integration of EEG features of spectrum, complexity, and synchronization for aiding the diagnosis of AD.

Machine Learning-Based Approach for Predicting Diabetes Employing Socio-Demographic Characteristics

Diabetes is one of the fatal diseases that play a vital role in the growth of other diseases in the human body. From a clinical perspective, the most significant approach to mitigating the effects of diabetes is early-stage control and management, with the aim of a potential cure. However, lack of awareness and expensive clinical tests are the primary reasons why clinical diagnosis and preventive measures are neglected in lower-income countries like Bangladesh, Pakistan, and India. From this perspective, this study aims to build an automated machine learning (ML) model, which will predict diabetes at an early stage using socio-demographic characteristics rather than clinical attributes, due to the fact that clinical features are not always accessible to all people from lower-income countries. To find the best fit of the supervised ML classifier of the model, we applied six classification algorithms and found that RF outperformed with an accuracy of 99.36%. In addition, the most significant risk factors were found based on the SHAP value by all the applied classifiers. This study reveals that polyuria, polydipsia, and delayed healing are the most significant risk factors for developing diabetes. The findings indicate that the proposed model is highly capable of predicting diabetes in the early stages.

Improving distributed denial of service attack detection using supervised machine learning

Distributed denial-of-service (DDoS) attacks are a big problem for cyber security because they can cause a lot of damage to both people and companies. Distributed Denial of Service (DDoS) attacks have been seen to do a lot of damage to the networks and devices they are aimed at. These hacks slow down networks and use up buffer space, which makes resources unavailable. To solve this problem, “Supervised Machine Learning Models” have been used. Several machine learning techniques, such as Random Forest, K-Nearest Neighbors (KNN), and Logistic Regression, were used to figure out what was normal and what was an attack. This study used a sample of the CSE-CICIDS2018, CSE-CICIDS2017, and CICDoS datasets. The dataset was divided into two parts in which three fourth of the data was used for training and one fourth of the data for testing purpose. The proposed research attempt to classify the DDoS attack by using supervised machine learning classifiers. This approach employs three machine learning classifiers such as Random Forest, KNN and Logistic regression. Then we perform Feature Scaling by using Standard Scaler. Finally, the system was evaluated. Random forest classifier outperformed other classifiers with an accuracy of 97.6 % whereas KNN and Logistic regression achieved 97 % and 91.1 %. The study employed several Supervised Machine Learning techniques, including Random Forest, KNN, and Logistic Regression to identify the most effective algorithm for the test. Results demonstrate that Random Forest outperformed the other models.

Machine learning-based evaluation of application value of traditional Chinese medicine clinical index and pulse wave parameters in the diagnosis of polycystic ovary syndrome

IntroductionPolycystic ovary syndrome (PCOS) is a prevalent endocrine disorder in women that often leads to ovulatory infertility. This study aims to establish and validate an effective predictive model for PCOS and to explore the correlation of features between patients with irregular menstruation and those with PCOS, using pulse wave parameters and traditional Chinese medicine (TCM) clinical indices. MethodsFrom August 2018 to January 2022, women with irregular menstruation were enrolled in this study. Subjects who met the inclusion criteria were categorized into PCOS and non-PCOS groups based on diagnostic criteria. Pulse wave parameters and TCM clinical indices were collected by two medical professionals. After data cleaning, recursive feature elimination with cross-validation (RFECV) was used for feature selection. Four supervised machine learning classifiers were used to build PCOS prediction models, including Extra Trees (ET), Random Forest (RF), Extreme Gradient Boosting (XGB/XGBoost), and Support Vector Machine (SVM). The SHapley Additive exPlanation (SHAP) values based on the optimal model were visualized for further feature explanation. ResultsA total of 450 women with irregular periods were enrolled in the study, consisting of 294 patients with PCOS and 156 without PCOS. Based on RFECV, 31 features, including 12 pulse parameters and 19 TCM clinical indices, were selected for building prediction models. Using pulse and TCM clinical index was superior to using pulse parameters or TCM clinical index alone in prediction. SVM achieved the best PCOS prediction results (accuracy=0.837, AUC=0.878, F1 score=0.878). For pulse parameters, lower values of right As, right h4, left h1, left h3, left h5, left w/t, and higher right t5 showed an obvious positive PCOS predictive effect. Women with PCOS were more likely to experience delayed menstruation, negative emotions, a slightly jelly-like menstrual texture, higher BMI, and a TCM assessment of greasy tongue coating. ConclusionA PCOS prediction model based on the SVM algorithm was established and verified as the best model for distinguishing between patients with PCOS and patients without PCOS with irregular menstruation. The new prediction model that uses pulse wave parameters and TCM clinical indices offers a non-invasive and cost-effective way to diagnose PCOS, and the model provides objective evidence for TCM diagnosis.

Opportunities and challenges of supervised machine learning for the classification of motor evoked potentials according to muscles

BackgroundEven for an experienced neurophysiologist, it is challenging to look at a single graph of an unlabeled motor evoked potential (MEP) and identify the corresponding muscle. We demonstrate that supervised machine learning (ML) can successfully perform this task.MethodsIntraoperative MEP data from supratentorial surgery on 36 patients was included for the classification task with 4 muscles: Extensor digitorum (EXT), abductor pollicis brevis (APB), tibialis anterior (TA) and abductor hallucis (AH). Three different supervised ML classifiers (random forest (RF), k-nearest neighbors (kNN) and logistic regression (LogReg)) were trained and tested on either raw or compressed data. Patient data was classified considering either all 4 muscles simultaneously, 2 muscles within the same extremity (EXT versus APB), or 2 muscles from different extremities (EXT versus TA).ResultsIn all cases, RF classifiers performed best and kNN second best. The highest performances were achieved on raw data (4 muscles 83%, EXT versus APB 89%, EXT versus TA 97% accuracy).ConclusionsStandard ML methods show surprisingly high performance on a classification task with intraoperative MEP signals. This study illustrates the power and challenges of standard ML algorithms when handling intraoperative signals and may lead to intraoperative safety improvements.

Relative performance evaluation of machine learning algorithms for land use classification using multispectral moderate resolution data

Analyses of spatial and temporal patterns of land use and land cover through multi-resolution remote sensing data provide valuable insights into landscape dynamics. Land use changes leading to land degradation and deforestation have been a prime mover for changes in the climate. This necessitates accurately assessing land use dynamics using a machine-learning algorithm’s temporal remote sensing data. The current study investigates land use using the temporal Landsat data from 1973 to 2021 in Chikamagaluru district, Karnataka. The land cover analysis showed 2.77% decrease in vegetation cover. The performance of three supervised learning techniques, namely Random Forest (RF), Support Vector Machine (SVM), and Maximum Likelihood classifier (MLC) were assessed, and results reveal that RF has performed better with an overall accuracy of 90.22% and a kappa value of 0.85. Land use classification has been performed with supervised machine learning classifier Random Forest (RF), which showed a decrease in the forest cover (48.91%) with an increase of agriculture (6.13%), horticulture (43.14%) and built-up cover (2.10%). Forests have been shrinking due to anthropogenic forces, especially forest encroachment for agriculture and industrial development, resulting in forest fragmentation and habitat loss. The fragmentation analysis provided the structural change in the forest cover, where interior forest cover was lost by 27.67% from 1973 to 2021, which highlights intense anthropogenic pressure even in the core Western Ghats regions with dense forests. Temporal details of the extent and condition of land use form an information base for decision-makers.

An evaluation of nature-inspired optimization algorithms and machine learning classifiers for electricity fraud prediction

&lt;span&gt;This study evaluated the nature-inspired optimization algorithms to improve classification involving imbalanced class problems. The particle swarm optimization (PSO) and grey wolf optimizer (GWO) were used to adaptively balance the distribution and then four supervised machine learning classifiers artificial neural network (ANN), support vector machine (SVM), extreme gradient-boosted tree (XGBoost), and random forest (RF) were applied to maximize the classification performance for electricity fraud prediction. The imbalance data was balanced using random undersampling (RUS) and two nature-inspired algorithm techniques (PSO and GWO). Results showed that for the data balanced using random undersampling, ANN (Sentest = 50.31%), and XGBoost (Sentest = 66.32%) has better sensitivity than SVM (Sentest = 23.61%), while RF exhibits overfitting (Sentrain = 100%, Sentest = 71.25%). The classification performance of RF model hybrid with PSO improved tremendously (AccTest = 96.98%, Sentest = 94.87%, Spectest = 99.16%, Pretest = 99.14%, F1 Score = 96.96%, and area under the curve (AUC) = 0.989). This was closely followed by hybrid of XGBoost with PSO. Moreover, RF and XGBoost hybrid with GWO also showed an improvement and promising results. This study has showed that nature-inspired optimization algorithms (PSO and GWO) are effective methods in addressing imbalanced dataset.&lt;/span&gt;

Machine learning-based decision tool for selecting patients with idiopathic acute pancreatitis for endosonography to exclude a biliary aetiology.

Biliary microlithiasis/sludge is detected in approximately 30% of patients with idiopathic acute pancreatitis (IAP). As recurrent biliary pancreatitis can be prevented, the underlying aetiology of IAP should be established. To develop a machine learning (ML) based decision tool for the use of endosonography (EUS) in pancreatitis patients to detect sludge and microlithiasis. We retrospectively used routinely recorded clinical and laboratory parameters of 218 consecutive patients with confirmed AP admitted to our tertiary care hospital between 2015 and 2020. Patients who did not receive EUS as part of the diagnostic work-up and whose pancreatitis episode could be adequately explained by other causes than biliary sludge and microlithiasis were excluded. We trained supervised ML classifiers using H2O.ai automatically selecting the best suitable predictor model to predict microlithiasis/sludge. The predictor model was further validated in two independent retrospective cohorts from two tertiary care centers (117 patients). Twenty-eight categorized patients' variables recorded at admission were identified to compute the predictor model with an accuracy of 0.84 [95% confidence interval (CI): 0.791-0.9185], positive predictive value of 0.84, and negative predictive value of 0.80 in the identification cohort (218 patients). In the validation cohort, the robustness of the prediction model was confirmed with an accuracy of 0.76 (95%CI: 0.673-0.8347), positive predictive value of 0.76, and negative predictive value of 0.78 (117 patients). We present a robust and validated ML-based predictor model consisting of routinely recorded parameters at admission that can predict biliary sludge and microlithiasis as the cause of AP.

Comparison of Machine Learning Pixel-Based Classifiers for Detecting Archaeological Ceramics

Recent improvements in low-altitude remote sensors and image processing analysis can be utilised to support archaeological research. Over the last decade, the increased use of remote sensing sensors and their products for archaeological science and cultural heritage studies has been reported in the literature. Therefore, different spatial and spectral analysis datasets have been applied to recognise archaeological remains or map environmental changes over time. Recently, more thorough object detection approaches have been adopted by researchers for the automated detection of surface ceramics. In this study, we applied several supervised machine learning classifiers using red-green-blue (RGB) and multispectral high-resolution drone imageries over a simulated archaeological area to evaluate their performance towards semi-automatic surface ceramic detection. The overall results indicated that low-altitude remote sensing sensors and advanced image processing techniques can be innovative in archaeological research. Nevertheless, the study results also pointed out existing research limitations in the detection of surface ceramics, which affect the detection accuracy. The development of a novel, robust methodology aimed to address the “accuracy paradox” of imbalanced data samples for optimising archaeological surface ceramic detection. At the same time, this study attempted to fill a gap in the literature by blending AI methodologies for non-uniformly distributed classes. Indeed, detecting surface ceramics using RGB or multi-spectral drone imageries should be reconsidered as an ‘imbalanced data distribution’ problem. To address this paradox, novel approaches need to be developed.

Composite Panel Damage Classification Based on Guided Waves and Machine Learning: An Experimental Approach

Ultrasonic guided waves (UGW) are widely used in structural health monitoring (SHM) systems due to the sensitivity of their propagation mechanisms to local material changes, i.e., those induced by damage. Post-processing of the signals gathered by piezoelectric sensors, typically used for both the excitation and the sensing of UGW, is a fundamental step to extract all the peculiar features that can be related to both damage location and severity. This research probes the efficacy of machine learning (ML) models in discerning damage location (R-Classification) and size (S-Classification). Seven supervised ML classifiers were examined: Ensemble-Subspace K-Nearest Neighbors (KNN), Ensemble-Bagged Trees, KNN-Fine, Ensemble-Boosted Trees, Support Vector Machine (SVM), Linear Discriminant, and SVM-Quadratic. The experimental dataset comprised measurements from varied reversible damage configurations on a composite panel, represented by wooden cuboids of single and three different sizes. Signal noise was minimized by performing a low-pass filter, and sequence forward selection-aided feature selection. The optimized ensemble classifier proved to be the most precise for R-Classification (95.83% accuracy), while Ensemble-Subspace KNN excelled in S-Classification (98.1% accuracy). This method offers accurate, efficient damage diagnosis and classification in composite structures, promising potential applications in aerospace, automotive, and civil engineering sectors.

Total soluble solids, dry matter content prediction and maturity stage classification of durian fruit using long-wavelength NIR reflectance

The DM and TSS of durian pulp moving on a conveyor belt were measured for their rapid and non-destructive qualities based on quantitative and qualitative measurements. The calibration set and prediction set equaled 209 and 69 pulps, respectively. The quantitative test compared the performance of PLS regression for DM and TSS prediction developed from full wavelength (860–1754 nm) and a few significant variables using SPA, GA, and VIP methods. The qualitative test identified the possibility of maturity stage classification by comparing three supervised machine learning classifiers, namely SVM, random forest (RF), and LDA. Effective models for DM and TSS prediction were developed from second derivatives spectra combined with the GA method, exhibiting r2, SEP, and RPD of 0.85, 4.50%, and 2.64, respectively for DM, and 0.66, 5.15%, and 1.60, respectively, for TSS. The model classifying samples into two distinct groups, namely "reject" and "pass," utilizing the LDA algorithm, exhibited an impressive accuracy rate of 94.20%, making it a suitable choice for quality assurance purposes. This result indicates that the few effective variables were more efficient than full wavelength and improved model accuracy with greater model stability. Enhancing the classification model could involve data sample balancing in each group, leading to further improvements.

Potential of GJA8 gene variants in predicting age-related cataract: A comparison of supervised machine learning methods.

Cataracts are the problems associated with the crystallins proteins of the eye lens. Any perturbation in the conformity of these proteins results in a cataract. Age-related cataract is the most common type among all cataracts as it accounts for almost 80% of cases of senile blindness worldwide. This research study was performed to predict the role of single nucleotide polymorphisms (SNPs) of the GJA8 gene with age-related cataracts in 718 subjects (400 age-related cataract patients and 318 healthy individuals). A comparison of supervised machine learning classification algorithm including logistic regression (LR), random forest (RF) and Artificial Neural Network (ANN) were presented to predict the age-related cataracts. The results indicated that LR is the best for predicting age-related cataracts. This successfully developed model after accounting different genetic and demographic factors to predict cataracts will help in effective disease management and decision-making medical practitioner and experts.

An Ensemble Approach to Cyberbullying Detection and Prevention on Social Media

Over the past decade, digital communication has reached a massive scale globally. Unfortunately, cyberbullying has also seen a significant increase which commensurate with the growth of digital technology, and perpetrators hiding behind the cloak of relative internet anonymity. Studies have shown that cyberbullying leaves a lasting psychological scar on its victims and often have devastating outcome. This has necessitated the development of measures to curb cyberbullying. This study presents one of such measure in the form of an ensemble model for cyberbullying detection. The proposed model features a majority voting ensemble approach to cyberbullying detection using three (3) supervised machine learning classifiers: SVM, NB and K-NN, as base learners. The malignant comment dataset, sourced from Kaggle.com. was used for model building at a split ratio of 70: 30 to achieve maximum model training and evaluation respectively. Evaluation result was based on standard metrics. The proposed ensemble model performed best of all the models implemented, with an accuracy of 95%. It was also observed to be the most consistent classifier across all the metrics considered. This showcased the efficacy of the ensemble model in cyberbullying comments detection.

Supervised machine learning classification of psychosis biotypes based on brain structure: findings from the Bipolar-Schizophrenia network for intermediate phenotypes (B-SNIP)

Traditional diagnostic formulations of psychotic disorders have low correspondence with underlying disease neurobiology. This has led to a growing interest in using brain-based biomarkers to capture biologically-informed psychosis constructs. Building upon our prior work on the B-SNIP Psychosis Biotypes, we aimed to examine whether structural MRI (an independent biomarker not used in the Biotype development) can effectively classify the Biotypes. Whole brain voxel-wise grey matter density (GMD) maps from T1-weighted images were used to train and test (using repeated randomized train/test splits) binary L2-penalized logistic regression models to discriminate psychosis cases (n = 557) from healthy controls (CON, n = 251). A total of six models were evaluated across two psychosis categorization schemes: (i) three Biotypes (B1, B2, B3) and (ii) three DSM diagnoses (schizophrenia (SZ), schizoaffective (SAD) and bipolar (BD) disorders). Above-chance classification accuracies were observed in all Biotype (B1 = 0.70, B2 = 0.65, and B3 = 0.56) and diagnosis (SZ = 0.64, SAD = 0.64, and BD = 0.59) models. However, the only model that showed evidence of specificity was B1, i.e., the model was able to discriminate B1 vs. CON and did not misclassify other psychosis cases (B2 or B3) as B1 at rates above nominal chance. The GMD-based classifier evidence for B1 showed a negative association with an estimate of premorbid general intellectual ability, regardless of group membership, i.e. psychosis or CON. Our findings indicate that, complimentary to clinical diagnoses, the B-SNIP Psychosis Biotypes may offer a promising approach to capture specific aspects of psychosis neurobiology.

How do Machine Learning Algorithms Effectively Classify Toxic Comments? An Empirical Analysis

Toxic comments on social media platforms, news portals, and online forums are impolite, insulting, or unreasonable that usually make other users leave a conversation. Due to the significant number of comments, it is impractical to moderate them manually. Therefore, online service providers use the automatic detection of toxicity using Machine Learning (ML) algorithms. However, the model's toxicity identification performance relies on the best combination of classifier and feature extraction techniques. In this empirical study, we set up a comparison environment for toxic comment classification using 15 frequently used supervised ML classifiers with the four most prominent feature extraction schemes. We considered the publicly available Jigsaw dataset on toxic comments written by human users. We tested, analyzed and compared with every pair of investigated classifiers and finally reported a conclusion. We used the accuracy and area under the ROC curve as the evaluation metrics. We revealed that Logistic Regression and AdaBoost are the best toxic comment classifiers. The average accuracy of Logistic Regression and AdaBoost is 0.895 and 0.893, respectively, where both achieved the same area under the ROC curve score (i.e., 0.828). Therefore, the primary takeaway of this study is that the Logistic Regression and Adaboost leveraging BoW, TF-IDF, or Hashing features can perform sufficiently for toxic comment classification.

Effect of COVID-19 vaccination and booster on maternal–fetal outcomes: a retrospective cohort study

COVID-19 in pregnant people increases the risk for poor maternal-fetal outcomes. However, COVID-19 vaccination hesitancy remains due to concerns over the vaccine's potential effects on maternal-fetal outcomes. Here we examine the impact of COVID-19 vaccination and boosters on maternal SARS-CoV-2 infections and birth outcomes. This was a retrospective multicentre cohort study on the impact of COVID-19 vaccination on maternal-fetal outcomes for people who delivered (n=106 428) at Providence St Joseph Health across seven western US states from Jan 26, 2021 to Oct 26, 2022. Cohorts were defined by vaccination status at delivery: vaccinated (n=35 926; two or more doses of mRNA-1273 Moderna or BNT162b2 Pfizer-BioNTech), unvaccinated (n=55 878), unvaccinated propensity score matched (n=16 771), boosted (n=10 927; three or more doses), vaccinated unboosted (n=13 243; two doses only), and vaccinated unboosted with propensity score matching (n=4414). We built supervised machine learning classification models, which we used to determine which people were more likely to be vaccinated or boosted at delivery. The primary outcome was maternal SARS-CoV-2 infection. COVID-19 vaccination status at delivery, COVID-19-related health care, preterm birth, stillbirth, and very low birthweight were evaluated as secondary outcomes. Vaccinated people were more likely to conceive later in the pandemic, have commercial insurance, be older, live in areas with lower household composition vulnerability, and have a higher BMI than unvaccinated people. Boosted people were more likely to have more days since receiving the second COVID-19 vaccine dose, conceive earlier in the pandemic, have commercial insurance, be older, and live in areas with lower household composition vulnerability than vaccinated unboosted people. Vaccinated pregnant people had lower rates of COVID-19 during pregnancy (4·0%) compared with unvaccinated matched people (5·3%; p<0·0001). COVID-19 rates were even lower in boosted people (3·2%) compared with vaccinated unboosted matched people (5·6%; p<0·0001). Vaccinated people were also less likely to have a preterm birth (7·9%; p<0·0001), stillbirth (0·3%; p<0·0002), or very low birthweight neonate (1·0%; p<0·0001) compared with unvaccinated matched people (preterm birth 9·4%; stillbirth 0·6%; very low birthweight 1·5%). Boosted people were less likely to have a stillbirth (0·3%; p<0·025) and have no differences in rates of preterm birth (7·6%; p=0·090) or very low birthweight neonates (0·8%; p=0·092) compared with vaccinated unboosted matched people (stillbirth 0·5%; preterm birth 8·4%; very low birthweight 1·1%). COVID-19 vaccination protects against adverse maternal-fetal outcomes, with booster doses conferring additional protection. Pregnant people should be high priority for vaccination and stay up to date with their COVID-19 vaccination schedule. National Institute for Child Health & Human Development and the William O and K Carole Ellison Foundation.