Explore the vast range of titles available.

Localisation of objects and people in indoor environments has been widely studied due to security issues and because of the benefits that a localisation system can provide. Indoor positioning systems (IPSs) based on more than one technology can improve localisation performance by leveraging the advantages of distinct technologies. This study proposes a multi-sensor IPS able to estimate the three-dimensional (3D) location of stationary objects using off-the-shelf equipment. By using radio-frequency identification (RFID) technology, machine-learning models based on support vector regression (SVR) and artificial neural networks (ANNs) are proposed. A k-means technique is also applied to improve accuracy. A computer vision (CV) subsystem detects visual markers in the scenario to enhance RFID localisation. To combine the RFID and CV subsystems, a fusion method based on the region of interest is proposed. We have implemented the authors’ system and evaluated it using real experiments. On bi-dimensional scenarios, localisation error is between 9 and 29 cm in the range of 1 and 2.2 m. In a machine-learning approach comparison, ANN performed 31% better than SVR approach. Regarding 3D scenarios, localisation errors in dense environments are 80.7 and 73.7 cm for ANN and SVR models, respectively.

The rapid escalation of applying Machine Learning (ML) in various domains has led to paying more attention to the quality of ML components. There is then a growth of techniques and tools aiming at improving the quality of ML components and integrating them into the ML-based system safely. Although most of these tools use bugs’ lifecycle, there is no standard benchmark of bugs to assess their performance, compare them and discuss their advantages and weaknesses. In this study, we firstly investigate the reproducibility and verifiability of the bugs in ML-based systems and show the most important factors in each one. Then, we explore the challenges of generating a benchmark of bugs in ML-based software systems and provide a bug benchmark namely defect4ML that satisfies all criteria of standard benchmark, i.e. relevance, reproducibility, fairness, verifiability, and usability. This faultload benchmark contains 100 bugs reported by ML developers in GitHub and Stack Overflow, using two of the most popular ML frameworks: TensorFlow and Keras. defect4ML also addresses important challenges in Software Reliability Engineering of ML-based software systems, like: 1) fast changes in frameworks, by providing various bugs for different versions of frameworks, 2) code portability, by delivering similar bugs in different ML frameworks, 3) bug reproducibility, by providing fully reproducible bugs with complete information about required dependencies and data, and 4) lack of detailed information on bugs, by presenting links to the bugs’ origins. defect4ML can be of interest to ML-based systems practitioners and researchers to assess their testing tools and techniques.

Abstract Background Collagenous colitis (CC) is diagnosed histologically and is characterised by a thickened subepithelial collagen band together with inflammatory and epithelial changes. Although routine haematoxylin and eosin (H&E) staining is sufficient for diagnosis in most cases, visual assessment of the collagen band can be challenging in borderline or heterogeneous specimens. Additional stains may be required in diagnostically difficult situations. The aim To develop a machine-learning–based algorithm for detecting subepithelial collagen band-associated regions in routine H&E-stained colonic biopsy images as a decision-support tool for histopathological assessment. Methods H&E-stained colonic biopsy specimens from 36 patients with histologically confirmed CC were imaged at 20 × magnification (1392 × 1040 pixels). Images were segmented into 1,000 superpixels using the Simple Linear Iterative Clustering (SLIC) algorithm. Superpixels overlapping with expert-provided rough annotations of the collagen band were labelled and characterised using normalised RGB histograms. A feed-forward neural network classifier (three hidden layers, 10 neurons per layer) was trained to distinguish collagen band–associated from non-collagen regions. Class imbalance was addressed by data augmentation of minority-class superpixels. Post-processing with connected-component size filtering was applied to enforce spatial continuity. Superpixel-level performance was evaluated quantitatively, and image-level outputs were assessed using expert acceptability scoring. Results The classifier achieved a superpixel-wise accuracy of 0.928 (sensitivity 0.898, specificity 0.953). Size-based post-processing substantially reduced isolated false-positive detections. At the image level, the final algorithm achieved an acceptability accuracy of 0.846 according to expert evaluation. The model successfully highlighted subepithelial collagen band–associated regions consistent with expert annotations but did not model additional diagnostic features required for complete CC diagnosis. Conclusion Our superpixel-based neural network highlights collagen-rich regions in H&E-stained colonic biopsies, offering decision support for pathologists. As diagnosis of collagenous colitis requires broader histopathological and clinical context, this method is intended as a decision-support tool rather than a stand-alone diagnostic solution.

Background Collagenous colitis (CC) is diagnosed histologically and is characterised by a thickened subepithelial collagen band together with inflammatory and epithelial changes. Although routine haematoxylin and eosin (H&E) staining is sufficient for diagnosis in most cases, visual assessment of the collagen band can be challenging in borderline or heterogeneous specimens. Additional stains may be required in diagnostically difficult situations. The aim To develop a machine-learning–based algorithm for detecting subepithelial collagen band-associated regions in routine H&E-stained colonic biopsy images as a decision-support tool for histopathological assessment. Methods H&E-stained colonic biopsy specimens from 36 patients with histologically confirmed CC were imaged at 20 × magnification (1392 × 1040 pixels). Images were segmented into 1,000 superpixels using the Simple Linear Iterative Clustering (SLIC) algorithm. Superpixels overlapping with expert-provided rough annotations of the collagen band were labelled and characterised using normalised RGB histograms. A feed-forward neural network classifier (three hidden layers, 10 neurons per layer) was trained to distinguish collagen band–associated from non-collagen regions. Class imbalance was addressed by data augmentation of minority-class superpixels. Post-processing with connected-component size filtering was applied to enforce spatial continuity. Superpixel-level performance was evaluated quantitatively, and image-level outputs were assessed using expert acceptability scoring. Results The classifier achieved a superpixel-wise accuracy of 0.928 (sensitivity 0.898, specificity 0.953). Size-based post-processing substantially reduced isolated false-positive detections. At the image level, the final algorithm achieved an acceptability accuracy of 0.846 according to expert evaluation. The model successfully highlighted subepithelial collagen band–associated regions consistent with expert annotations but did not model additional diagnostic features required for complete CC diagnosis. Conclusion Our superpixel-based neural network highlights collagen-rich regions in H&E-stained colonic biopsies, offering decision support for pathologists. As diagnosis of collagenous colitis requires broader histopathological and clinical context, this method is intended as a decision-support tool rather than a stand-alone diagnostic solution.

Context An increasing demand is observed in various domains to employ Machine Learning (ML) for solving complex problems. ML models are implemented as software components and deployed in Machine Learning Software Systems (MLSSs). Problem There is a strong need for ensuring the serving quality of MLSSs. False or poor decisions of such systems can lead to malfunction of other systems, significant financial losses, or even threats to human life. The quality assurance of MLSSs is considered a challenging task and currently is a hot research topic. Objective This paper aims to investigate the characteristics of real quality issues in MLSSs from the viewpoint of practitioners. This empirical study aims to identify a catalog of quality issues in MLSSs. Method We conduct a set of interviews with practitioners/experts, to gather insights about their experience and practices when dealing with quality issues. We validate the identified quality issues via a survey with ML practitioners. Results Based on the content of 37 interviews, we identified 18 recurring quality issues and 24 strategies to mitigate them. For each identified issue, we describe the causes and consequences according to the practitioners’ experience. Conclusion We believe the catalog of issues developed in this study will allow the community to develop efficient quality assurance tools for ML models and MLSSs. A replication package of our study is available on our public GitHub repository.

The forecasting of exchange rates has become a challenging area of research that has attracted many researchers over recent years. This work presents a sliding-window metaheuristic optimization-based forecast (SMOF) system for one-step ahead forecasting. The proposed system is a graphical user interface, which is developed in the MATLAB environment and functions as a stand-alone application. The system integrates the novel firefly algorithm (FA), metaheuristic (Meta) intelligence, and least squares support vector regression (LSSVR), namely MetaFA-LSSVR, with a sliding-window approach. The MetaFA automatically tunes the hyperparameters of the LSSVR to construct an optimal sliding-window LSSVR prediction model. The optimization effectiveness of the MetaFA is verified using ten benchmark functions. Two case studies on the daily Canadian dollar-USD exchange rate (CAN/USD) and the 4-h closing EUR-USD rates (EUR/USD) were used to confirm the performance of the system, in which the mean absolute percentage errors are 0.2532% and 0.169%, respectively. The forecast system has an 89.8–99.7% greater predictive accuracy than prior work when applied to the currency pair CAN/USD. With respect to the EUR/USD exchange rate, the error rates obtained using the proposed system were 20.8–23.9% better than those obtained by the baseline sliding-window LSSVR model. Therefore, the SMOF system is potentially useful for decision-makers in financial markets.

Objective: Malignant pleural mesothelioma is a highly aggressive tumor of the serous membranes, which in humans results from exposure to asbestos and asbestiform fibers. The incidence of malignant mesothelioma is extremely high in some Turkish villages where there is a low-level environmental exposure to erionite, a fibrous zeolite. Therefore epidemiological studies are difficult to perform in Turkey. Methods: In this paper, a study on malignant pleural mesothelioma disease diagnosis was realized by using artificial immune system. Also, the artificial immune system result was compared with the result of the multi-layer neural network focusing on malignant pleural mesothelioma disease diagnosis and using same database. The malignant pleural mesothelioma disease dataset were prepared from a faculty of medicine’s database using patient’s hospital reports. Results: 97.74% accuracy performance is obtained by artificial immune system. The accuracy results of artificial immune system algorithm are much better than the accuracy results of multi-layer neural network algorithm. Conclusion: This system is capable of conducting the classification process with a good performance to help the expert while deciding the healthy and patient subjects. So, this structure can be helpful as learning based decision support system for contributing to the doctors in their diagnosis decisions. Key words: malignant pleural mesothelioma disease diagnosis, artificial immune system, machine learning based decision support system. Amaç: İnsanların beyin zarında bulunan, asbestos ve asbestiform liflerine maruz kalmakla oluşan kötü huylu plevral Mezotelyoma, oldukça saldırgan bir tümördür. Düşük seviyeli çevresel erionite fibrous zeolite’e maruz bırakılmış Türkiye’deki bazı kasabalarda Mezotelyoma görülme oranı oldukça yüksektir.Yöntemler: Bu çalışmada Mezotelyoma hastalığı teşhisi yapay bağışıklık sistemi kullanımı ile gerçekleştirilmiştir. Bununla beraber yapay bağışıklık sistemi sonuçları, aynı veri tabanını kullanan, Mezotelyoma hastalığının teşhisine odaklanmış çok katmanlı yapay sinir ağı sonuçları ile karşılaştırılmıştır. Mezotelyoma hastalığı veri seti, hastaların hastane raporlarını kullanan tıp fakültesi veri tabanından alınmıştır.Bulgular: Yapay bağışıklık sistemi tarafından hastalık teşhisi için %97,74 doğruluk oranında bir performans elde edilmiştir. Yapay bağışıklık sistemi algoritmasının doğruluk sonuçları çok katmanlı yapay sinir ağı algoritmasından çok daha iyi olduğu görülmüştür.Sonuç: Bu sistem uzmana, sağlıklı ve hasta kişiyi sınıflandırma sürecinde doğru teşhisi bulma yönünde iyi bir performans sağlar. Böylece bu yapı ile doğru teşhis sonucuna ulaşmada doktorlara bir karar destek sistemi olarak yardımcı olur

Background Antimicrobial resistance poses challenges for physicians, who must balance individual patient care and public health when prescribing antibiotics. Machine learning-based computerized decision support systems (ML-CDSS) are increasingly proposed to aid in this challenge. We aimed to assess physicians’ decision-making in a common bacterial vs. viral infection scenario, and the impact of an ML-CDSS on it. Methods We administered an online scenario-based survey to physicians ( N  = 211), mainly pediatricians (67.8%). The estimated response rate was 33–40%. Each participant encountered four sore throat scenarios, corresponding to one of four McIsaac scores. Participants estimated the probabilities of bacterial infections and determined treatment strategies. This sequence occurred both before and after simulated hypothetical ML-CDSS interventions, in the form of a probability of bacterial infection output. Results The average probability estimates of bacterial infection under the four McIsaac scenarios were monotonically increasing: (1) 25.6% (95% CI 22.8–28.4%), (2) 43.8% (40.6–46.7%), (3) 65.1% (62.2–67.0%), and (4) 69.1% (66.3–71.8%). Furthermore, empiric treatment was generally overprescribed: (1) 11.4% (2) 38.4% (3) 65.8% (4) 73.0%. These estimates and treatment percentages are higher than expected given the relevant scientific literature. The interventions had substantial effects on probability estimates and empiric prescription; e.g. reducing average estimates by up to 14% points and lowering odds of antibiotic prescription by a factor 0.42. Conclusions Overestimation of bacterial infections and subsequent antibiotic overprescription are common, particularly under conditions of clinical uncertainty. These tendencies can be mitigated through ML-CDSS interventions, as demonstrated in a scenario-based survey setting. Our findings provide initial support for the design of ML-CDSS tools and their integration into primary care, pending further validation in clinical trials. Additionally, they support policy initiatives aimed at clarifying default clinical actions in situations of diagnostic uncertainty. Clinical trial Not applicable.

The recent increase in infiltration of distributed resources has challenged the traditional operation of power systems. Simultaneously, devastating effects of recent natural disasters have questioned the resilience of power infrastructure for an electricity dependent community. In this study, a solution has been presented in the form of a resilient smart grid network which utilises distributed energy resources (DERs) and machine learning (ML) algorithms to improve the power availability during disastrous events. In addition to power electronics with load categorisation features, the presented system utilises ML tools to use the information from neighbouring units and external sources to make complicated logical decisions directed towards providing power to critical loads at all times. Furthermore, the provided model encourages consideration of ML tools as a part of smart grid design process together with power electronics and controls, rather than as an additional feature.

Senior citizens have increased plasma glucose and a higher risk of diabetes-related complications than young people. However, it is difficult to diagnose and manage elderly diabetics because there is no clear symptom according to current diagnostic criteria. They also dislike the invasive blood sample test. This study aimed to classify a difference in gait and physical fitness characteristics between senior citizens with and without diabetes for a non-invasive method and propose a machine-learning-based personal home-training system for training abnormal gait motions by oneself. We used a dataset for classification with 200 over 65-year-old elders who walked a flat and straight 15 m route in 3 different walking speed conditions using an inertial measurement unit and physical fitness test. Then, questionnaires for participants were included to identify life patterns. Through results, it was found that there were abnormalities in gait and physical fitness characteristics related to balance ability and walking speed. Using a single RGB camera, the developed training system for improving abnormalities enabled us to correct the exercise posture and speed in real-time. It was discussed that there are risks and errors in the training system based on human pose estimation for future works.