Explore the vast range of titles available.

Background: Monitoring of antibiotic prescription practices in hospitals is essential to assess and facilitate appropriate use. This is relevant to halt the progression of antimicrobial resistance. Methods: Assessment of antibiotic prescribing patterns and completeness of antibiotic prescriptions among out-patients in 2021 was conducted at the University Hospital of Kwame Nkrumah University of Science and Technology in the Ashanti region of Ghana. We reviewed electronic medical records (EMR) of 49,660 patients who had 110,280 encounters in the year. Results: The patient encounters yielded 350,149 prescriptions. Every month, 33–36% of patient encounters resulted in antibiotic prescription, higher than the World Health Organization’s (WHO) recommended optimum of 27%. Almost half of the antibiotics prescribed belonged to WHO’s Watch group. Amoxicillin–clavulanic acid (50%), azithromycin (29%), ciprofloxacin (28%), metronidazole (21%), and cefuroxime (20%) were the most prescribed antibiotics. Antibiotic prescribing parameters (indication, name of drug, duration, dose, route, and frequency) were documented in almost all prescriptions. Conclusions: Extending antimicrobial stewardship to the out-patient settings by developing standard treatment guidelines, an out-patient specific drug formulary, and antibiograms can promote rational antibiotic use at the hospital. The EMR system of the hospital is a valuable tool for monitoring prescriptions that can be leveraged for future audits.

Background and objectives: The inappropriate use of antibiotics in hospitals can potentially lead to the development and spread of antibiotic resistance, increased mortality, and high economic burden. The objective of the study was to assess current patterns of antibiotic use in leading hospitals of Pakistan. Moreover, the information collected can support in policy-making and hospital interventions aiming to improve antibiotic prescription and use. Methodology and materials: A point prevalence survey was carried out with data abstracted principally from patient medical records from 14 tertiary care hospitals. Data were collected through the standardized online tool KOBO application for smart phones and laptops. For data analysis, SPSS Software was used. The association of risk factors with antimicrobial use was calculated using inferential statistics. Results: Among the surveyed patients, the prevalence of antibiotic use was 75% on average in the selected hospitals. The most common classes of antibiotics prescribed were third-generation cephalosporin (38.5%). Furthermore, 59% of the patients were prescribed one while 32% of the patients were prescribed two antibiotics. Whereas the most common indication for antibiotic use was surgical prophylaxis (33%). There is no antimicrobial guideline or policy for 61.9% of antimicrobials in the respected hospitals. Conclusions: It was observed in the survey that there is an urgent need to review the excessive use of empiric antimicrobials and surgical prophylaxis. Programs should be initiated to address this issue, which includes developing antibiotic guidelines and formularies especially for empiric use as well as implementing antimicrobial stewardship activities.

With the widespread use of over-the-top (OTT) media, such as YouTube and Netflix, network markets are changing and innovating rapidly, making it essential for network providers to quickly and efficiently analyze OTT traffic with respect to pricing plans and infrastructure investments. This study proposes a time-aware deep-learning method of analyzing OTT traffic to classify users for this purpose. With traditional deep learning, classification accuracy can be improved over conventional methods, but it takes a considerable amount of time. Therefore, we propose a novel framework to better exploit accuracy, which is the strength of deep learning, while dramatically reducing classification time. This framework uses a two-step classification process. Because only ambiguous data need to be subjected to deep-learning classification, vast numbers of unambiguous data can be filtered out. This reduces the workload and ensures higher accuracy. The resultant method provides a simple method for customizing pricing plans and load balancing by classifying OTT users more accurately.

In this paper, we investigate how to modify the naive Bayes classifier in order to perform classification that is restricted to be independent with respect to a given sensitive attribute. Such independency restrictions occur naturally when the decision process leading to the labels in the data-set was biased; e.g., due to gender or racial discrimination. This setting is motivated by many cases in which there exist laws that disallow a decision that is partly based on discrimination. Naive application of machine learning techniques would result in huge fines for companies. We present three approaches for making the naive Bayes classifier discrimination-free: (i) modifying the probability of the decision being positive, (ii) training one model for every sensitive attribute value and balancing them, and (iii) adding a latent variable to the Bayesian model that represents the unbiased label and optimizing the model parameters for likelihood using expectation maximization. We present experiments for the three approaches on both artificial and real-life data.

Vehicle surface defect detection faces three core challenges: classification–localization inconsistency for boundary-sensitive defects, insufficient multi-scale feature response across defect sizes, and cross-scenario generalization degradation caused by domain shift among production lines. This paper proposes YOLOv12-VSD, an improved detection algorithm addressing these issues through coordinated modifications at three levels. An IoU-aware classification loss aligns classification confidence with localization quality. A reparameterized convolution module at the P4 feature level (P4-RepC3) enriches intermediate-layer directional feature diversity without increasing inference cost. A multi-scale spatial pyramid pooling–fast structure at the P5 feature level (P5-SPPF) expands the effective receptive field for large-area defects. A three-stage transfer learning framework comprising source-domain pretraining, target-domain adaptation, and low-learning-rate refinement is further designed to reduce domain shift with limited annotations. Experiments show that YOLOv12-VSD achieves a mean Average Precision at IoU threshold 0.50 (mAP@50) of 0.715, the highest among six comparison models, with only 6.1M parameters and 17.1 giga floating-point operations per second (GFLOPs). After three-stage transfer, mAP@50 improves from 0.531 to 0.652, with training duration reduced by 64%.

This paper introduces TigCLaF , a novel cross-lingual large language model framework for sentiment-aware text classification in low-resource Tigrigna. The framework integrates tokenizer extension with adaptation, continual pretraining on unlabeled Tigrigna data, and LoRA-based parameter-efficient fine-tuning to enable effective cross-lingual adaptation from high-resource languages. Leveraging recent advances in multilingual pre-trained language models and large language models, we investigate zero-shot, few-shot, and full fine-tuning strategies for sentiment detection, incorporating transformer models XLM-RoBERTa and AfriBERTa, as well as instruction-tuned LLaMA models. Our approach integrates a Tigrigna sentiment lexicon into transformer-based embeddings via feature fusion, thereby enhancing preservation of the sentiment signal during cross-lingual transfer. The proposed framework is evaluated on a newly curated dataset of 30,000 Tigrigna instances, supported by auxiliary English and Amharic sentiment corpora for transfer learning. Experimental results show that sentiment-aware feature integration improves classification accuracy and Macro-F1 up to 7% over baseline multilingual models without sentiment augmentation. Furthermore, parameter-efficient fine-tuning LoRA achieved competitive accuracy while reducing model size and inference latency, making it suitable for computationally constrained settings. The systematic error analysis highlights the roles of script-specific preprocessing, idiomatic expressions, and nuances of cultural sentiment in classification performance. The proposed framework demonstrates the viability of combining LLM-based cross-lingual transfer with masked sentiment-aware enhancements for practical, resource-efficient NLP applications in low-resource language contexts.

We develop a new approach to solving classification problems, which is based on the theory of coherent measures of risk and risk sharing ideas. We introduce the notion of a risk-averse classifier and a family of risk-averse classification problems. We show that risk-averse classifiers are associated with minimal points of the possible classification errors, where the minimality is understood with respect to a suitable stochastic order. The new approach allows for measuring risk by distinct risk functional for each class. We analyze the structure of the new classification problem and establish its theoretical relation to known risk-neutral design problems. In particular, we show that the risk-sharing classification problem is equivalent to an implicitly defined optimization problem with unequal weights for each data point. Additionally, we derive a confidence interval for the total risk of a risk-averse classifier. We implement our methodology in a binary classification scenario on several different data sets. We formulate specific risk-averse support vector machines in order to demonstrate the proposed approach and carry out numerical comparison with classifiers which are obtained using the Huber loss function and other loss functions known in the literature.

Background: Surveillance of antimicrobial consumption (AMC) is important to address inappropriate use. AMC data for countries in the European Union (EU) and European Economic Area (EEA) and Eastern European and Central Asian countries were compared to provide future guidance. Methods: Analyses of 2014–2018 data from 30 EU/EEA countries of the European Surveillance of Antibiotic Consumption network (ESAC-Net) and 15 countries of the WHO Regional Office for Europe (WHO Europe) AMC Network were conducted using the Anatomical Therapeutic Chemical (ATC) classification and Defined Daily Dose (DDD) methodology. Total consumption (DDD per 1000 inhabitants per day) of antibacterials for systemic use (ATC group J01), relative use (percentages), trends over time, alignment with the WHO Access, Watch, Reserve (AWaRe) classification, concordance with the WHO global indicator (60% of total consumption should be Access agents), and composition of the drug utilization 75% (DU75%) were calculated. Findings: In 2018, total consumption of antibacterials for systemic use (ATC J01) ranged from 8.9 to 34.1 DDD per 1000 inhabitants per day (population-weighted mean for ESAC-Net 20.0, WHO Europe AMC Network 19.6, ESAC-Net Study Group, and WHO Europe AMC Network Study Group). ESAC-Net countries consumed more penicillins (J01C; 8.7 versus 6.3 DDD per 1000 inhabitants per day), more tetracyclines (J01A; 2.2 versus 1.2), less cephalosporins (J01D; 2.3 versus 3.8) and less quinolones (J01M; 1.7 versus 3.4) than WHO Europe AMC Network countries. Between 2014 and 2018, there were statistically significant reductions in total consumption in eight ESAC-Net countries. In 2018, the relative population-weighted mean consumption of Access agents was 57.9% for ESAC-Net and 47.4% for the WHO Europe AMC Network. For each year during 2014–2018, 14 ESAC-Net and one WHO Europe AMC Network countries met the WHO global monitoring target of 60% of total consumption being Access agents. DU75% analyses showed differences in the choices of agents in the two networks. Interpretation: Although total consumption of antibacterials for systemic use was similar in the two networks, the composition of agents varied substantially. The greater consumption of Watch group agents in WHO Europe AMC Network countries suggests opportunities for improved prescribing. Significant decreases in consumption in several ESAC-Net countries illustrate the value of sustained actions to address antimicrobial resistance.

What are the main findings? * A velocity-aware multi-feature fusion method was proposed for classifying ground and low-altitude targets using measured X-band pulse-Doppler radar echoes. * By combining echo preprocessing and discriminative feature extraction, the proposed method achieved robust separation of humans, vehicles, and UAVs in cluttered outdoor environments. A velocity-aware multi-feature fusion method was proposed for classifying ground and low-altitude targets using measured X-band pulse-Doppler radar echoes. By combining echo preprocessing and discriminative feature extraction, the proposed method achieved robust separation of humans, vehicles, and UAVs in cluttered outdoor environments. What are the implications of the main findings? * The proposed framework provides a practical approach for radar target classification under complex outdoor clutter conditions. * The study supports the use of measured radar data and multi-feature fusion for ground surveillance and low-altitude target monitoring. The proposed framework provides a practical approach for radar target classification under complex outdoor clutter conditions. The study supports the use of measured radar data and multi-feature fusion for ground surveillance and low-altitude target monitoring. Classification of ground and low-altitude targets in radar remote sensing is challenging because environmental clutter and noise can significantly degrade the discriminability of target echoes, especially under complex outdoor observation conditions. To improve the classification performance for humans, vehicles, and unmanned aerial vehicles (UAVs), this paper proposes a velocity-aware multi-feature fusion method based on measured radar echo data. First, radar echoes are preprocessed using a wavelet-decomposition-based strategy to suppress clutter and noise while preserving useful target information. Then, multiple complementary features, including wavelet packet energy distribution, spectral entropy, spectral standard deviation, temporal standard deviation, amplitude dispersion coefficient, and relative radar cross-section (RCS), are extracted to characterize the target echoes from different perspectives. Considering the influence of target velocity on Doppler distribution and class separability, the measured data are further divided into different velocity intervals for stratified classification. Based on the fused feature vectors, a long short-term memory (LSTM) network is employed to model feature relationships and perform target classification. Experiments conducted on real measured radar echo data demonstrate that the proposed method achieves classification accuracies of 97.82% for UAVs, 96.00% for vehicles, and a mean interval-level accuracy of 96.94%, indicating its effectiveness for ground and low-altitude target classification in cluttered radar remote sensing environments.

Background Inappropriate antibiotic use is a major driver of antimicrobial resistance (AMR), particularly in low- and middle-income countries. To improve rational prescribing, the World Health Organization (WHO) developed core prescribing indicators and the AWaRe antibiotic classification. Evidence from Afghanistan remains scarce, especially in outpatient settings. Methods This cross-sectional prescription audit was conducted from June to December 2024 in six public and private hospitals in Kabul. Prescriptions were sampled from outpatient departments of internal medicine, pediatrics, surgery, and ENT. Of 623 prescriptions screened, 306 (49.1%) contained at least one antibiotic and were analyzed using WHO prescribing indicators and the AWaRe classification. Data were processed with SPSS v25. Results A total of 374 antibiotics were prescribed, with an average of 1.5 per prescription. The most common agents were ceftriaxone (19.9%), metronidazole (16.0%), ciprofloxacin (12.4%), and amoxicillin (11.4%). Generic prescribing accounted for 63.3%, while 42.7% of prescriptions included parenteral antibiotics. All antibiotics were from the National Essential Medicines List. AWaRe analysis showed Watch antibiotics (58.6%) predominated, while Access antibiotics (41.4%) were underused; no Reserve antibiotics were prescribed. No culture and sensitivity testing supported prescribing. Conclusion Antibiotic prescribing in Kabul outpatient settings substantially exceeded WHO standards, with high reliance on broad-spectrum Watch-group agents, injectables, and non-generic formulations. These findings underscore the urgent need for Antimicrobial Stewardship Programs, prescriber training, and improved diagnostic capacity to ensure rational antibiotic use and contain AMR in Afghanistan.