Relevant Algorithms Research Articles

Retracted: A Relevant Customer Identification Algorithm Based on the Internet Financial Platform

Retracted: A Relevant Customer Identification Algorithm Based on the Internet Financial Platform

Deep learning and machine learning predictive models for neurological function after interventional embolization of intracranial aneurysms.

The objective of this study is to develop a model to predicts the postoperative Hunt-Hess grade in patients with intracranial aneurysms by integrating radiomics and deep learning technologies, using preoperative CTA imaging data. Thereby assisting clinical decision-making and improving the assessment and prognosis of postoperative neurological function. This retrospective study encompassed 101 patients who underwent aneurysm embolization surgery. 851 radiomic features were extracted from CTA images. 512 deep learning features are extracted from last layer of ResNet50 deep convolutional neural network model. The feature screening process pipeline encompassed intraclass correlation coefficient analysis, principal component analysis, U test, spearman correlation analysis, minimum redundancy maximum relevance algorithm and Lasso regression, to identify features most correlated with postoperative Hunt-Hess grading. In the model construction phase, three distinct models were constructed: radiomics feature-based model (RSM), deep learning feature-based model (DLM), and deep learning-radiomics feature fusion model (DLRSCM). The study also calculated the radiomics score and combined it with clinical data to construct a Nomogram for predictive modeling. DLM, RSM and DLRSCM model was constructed by 9 base algorithms and 1 ensemble learning algorithm - Stacking ensemble model. Model performance was evaluated based on the area under the Receiver Operating Characteristic (ROC) curve (AUC), Matthews Correlation Coefficient (MCC), calibration curves, and decision curves analysis. 5 significant radiomic feature and 4 significant deep learning features were obtained through the feature selection process. These features were utilized for model construction. Bootstrap resampling method was used for internal validation of the models. In terms of model evaluation, the DLM model, the stacking ensemble algorithm results achieved an AUC of 0.959 and MCC of 0.815. In the RSM model, the stacking ensemble model AUC was 0.935 and MCC was 0.793. The stacking ensemble model in DLRSCM outperformed others, with an AUC of 0.968 and MCC of 0.820. Results indicated that the ANN performed optimally among all base models, while the stacked ensemble learning model exhibited the highest predictive performance. This study demonstrates that the combination of radiomics and deep learning is an effective approach to predict the postoperative Hunt-Hess grade in patients with intracranial aneurysms. This holds significant value in the early identification of postoperative neurological complications and in enhancing clinical decision-making.

P300 An Expert Delphi Consensus: Early Identification of Patients at Risk of Crohn's Disease in Perianal Fistulas and Abscesses (PREFAB) and Identification and Management of Isolated Perianal Crohn's Disease (ipCD) - Part A, PREFAB

Abstract Background In ± 10% of all Crohn’s disease (CD) patients a perianal abscess (PAA) or fistula (PAF) is the manifesting symptom. The delay in diagnosis of CD, from fistula or abscess onset, is very long and associated with worse outcomes. The aim of this Delphi study was to reach consensus on a clinical decision tool to help select patients with raised suspicion of CD in PAA/PAF patients, to identify underlying CD earlier and reduce the diagnostic delay. Methods A panel of international experts in the field of proctology and/or Inflammatory Bowel Disease (IBD), consisting of surgeons, gastroenterologists and radiologists were invited to participate in the PREFAB part of this Delphi study. The first round was an electronic survey and the second round a virtual consensus meeting. In the first round, participants were asked to anonymously provide their opinion probing 1) the relevance and use of clinical characteristics (“Red Flags”) suggestive of underlying CD, 2) the use of faecal calprotectin (FCP) as an adjunct for screening for CD and 3) on the diagnostic work-up for IBD in PAA/PAF patients with raised clinical suspicion. In the second round, statements were paired/revised based on the feedback from the first round and presented in a final set of statements. Consensus was predefined as ≥70% agreement. Results A total of 30 experts participated in the first round and 25 in the second round (83.3%). Final consensus was reached for 29% of all statements in the first Delphi round and, after adjustments/merging, for all 11 statements (100%) in the second round. Ninety-seven percent of participants agreed that shortening of the delay in diagnosis by using a clinical decision tool could improve outcomes in patients with perianal disease as a first symptom. Consensus was reached on the red flags to be included in the clinical decision tool, on screening of all patients with any perianal fistula (regardless of the complexity, biological behaviour and co-existent perianal symptoms) to identify CD in an early (sub)clinical phase, on referral criteria and on the diagnostic workup and follow up of patients with raised suspicion. Conclusion A clinical decision tool for early identification of CD in patients with PAA/PAF as a first symptom could shorten delay in diagnosis and improve outcomes. Global consensus for this tool, including a practical and relevant algorithm for finding or excluding CD in patients with PAA/PAF as a manifesting sign, was reached within two rounds. This clinical decision tool will be validated in 2024 in a large, prospective multicenter study.

A new digital image correlation method for discontinuous measurement in fracture analysis

The Digital Image Correlation (DIC) method is commonly used in fracture analysis, but it has limitations in measuring discontinuities, which is crucial in fracture analysis. To address this issue, this paper proposes the Adaptive Discontinuous Digital Image Correlation (AD-DIC) method, which is explicitly designed to measure discontinuous displacement fields effectively and accurately in fracture analysis. The AD-DIC method both overcomes the limitations of previous methods by incorporating an adaptive subset tracking technique, which adaptively adjusts the binarization threshold, and improves the traditional Reliability Guided-Digital Image Correlation algorithm to reconstruct the displacement field near cracks. According to high-precision discontinuous displacement fields measured by AD-DIC method, we can accurately identify cracks, measure crack opening, and calculate inclination angles. To validate the effectiveness of the AD-DIC method, theoretical images of type I and mixed-mode I/II cracks are analyzed, as well as four actual experiments. The results show that the AD-DIC method achieves a high level of accuracy in calculating crack opening, with an error of less than 0.1 pixels. The actual experiments presented in this study demonstrate the wide-ranging applications of the AD-DIC method in fracture analysis, including identifying cracks that are difficult to detect with the naked eye, measuring crack opening and inclination angle, and tracking crack propagation processes over time. This study provides an effective tool for brittle fracture analysis, and the relevant algorithms can serve as a reference for discontinuous measurement.

An artificial intelligence approach to predict infants’ health status at birth

BackgroundMachine learning could be used for prognosis/diagnosis of maternal and neonates’ diseases by analyzing the data sets and profiles obtained from a pregnant mother. PurposeWe aimed to develop a prediction model based on machine learning algorithms to determine important maternal characteristics and neonates’ anthropometric profiles as the predictors of neonates’ health status. MethodsThis study was conducted among 1280 pregnant women referred to healthcare centers to receive antenatal care. We evaluated several machine learning methods, including support vector machine (SVM), Ensemble, K-Nearest Neighbor (KNN), Naïve Bayes (NB), and Decision tree classifiers, to predict newborn health state. ResultsThe minimum redundancy-maximum relevance (MRMR) algorithm revealed that variables, including head circumference of neonates, pregnancy intention, and drug consumption history during pregnancy, were top-scored features for classifying normal and unhealthy infants. Among the different classification methods, the SVM classifier had the best performance. The average values of accuracy, precision, recall, F1-score, and area under the receiver operating characteristic curve (AUC) in the test group were 75%, 75%, 76%, 76%, and 65%, respectively, for SVM model. ConclusionMachine learning methods can efficiently forecast the neonate's health status among pregnant women. This study proposed a new approach toward the integration of maternal data and neonate profiles to facilitate the prediction of neonates’ health status.

Research on Student Behavior Analysis and Grade Prediction System Based on Student Behavior Characteristics

In the era of big data, the traditional governance model of student behavior management gradually shows the disadvantage of “post positioning”. Therefore, the research uses relevant data mining algorithms to extract and analyze students’ behavior characteristics, and constructs a SPC system model for students’ behavior analysis and performance prediction. At the same time, the experiment verifies its effectiveness. The experimental results show that in the factor analysis, the overall variance of the first seven indicators is 69.942\%, which shows that there is a significant correlation between students’ learning behavior and academic performance. In the model performance analysis of SPC system, the accuracy of the five algorithms is kept at 40% - 60%, while SVM has higher stability than other methods. In addition, the prediction accuracy and response rate of SPC model reached 86.90% and 81.57% respectively. When the sequence length was increased from 1 to 20, the accuracy of SPC model exceeded 70%; However, when the feature dimension exceeds 50, the model representation ability will decline. Therefore, appropriate feature dimensions are needed to predict student performance. To sum up, the SPC model built by the research is effective in analyzing student behavior and predicting students’ performance, and practical in actual student management.

MRI radiomics-based decision support tool for a personalized classification of cervical disc degeneration: a two-center study.

Objectives: To develop and validate an MRI radiomics-based decision support tool for the automated grading of cervical disc degeneration. Methods: The retrospective study included 2,610 cervical disc samples of 435 patients from two hospitals. The cervical magnetic resonance imaging (MRI) analysis of patients confirmed cervical disc degeneration grades using the Pfirrmann grading system. A training set (1,830 samples of 305 patients) and an independent test set (780 samples of 130 patients) were divided for the construction and validation of the machine learning model, respectively. We provided a fine-tuned MedSAM model for automated cervical disc segmentation. Then, we extracted 924 radiomic features from each segmented disc in T1 and T2 MRI modalities. All features were processed and selected using minimum redundancy maximum relevance (mRMR) and multiple machine learning algorithms. Meanwhile, the radiomics models of various machine learning algorithms and MRI images were constructed and compared. Finally, the combined radiomics model was constructed in the training set and validated in the test set. Radiomic feature mapping was provided for auxiliary diagnosis. Results: Of the 2,610 cervical disc samples, 794 (30.4%) were classified as low grade and 1,816 (69.6%) were classified as high grade. The fine-tuned MedSAM model achieved good segmentation performance, with the mean Dice coefficient of 0.93. Higher-order texture features contributed to the dominant force in the diagnostic task (80%). Among various machine learning models, random forest performed better than the other algorithms (p < 0.01), and the T2 MRI radiomics model showed better results than T1 MRI in the diagnostic performance (p < 0.05). The final combined radiomics model had an area under the receiver operating characteristic curve (AUC) of 0.95, an accuracy of 89.51%, a precision of 87.07%, a recall of 98.83%, and an F1 score of 0.93 in the test set, which were all better than those of other models (p < 0.05). Conclusion: The radiomics-based decision support tool using T1 and T2 MRI modalities can be used for cervical disc degeneration grading, facilitating individualized management.

Advances in the tea plants phenotyping using hyperspectral imaging technology.

Rapid detection of plant phenotypic traits is crucial for plant breeding and cultivation. Traditional measurement methods are carried out by rich-experienced agronomists, which are time-consuming and labor-intensive. However, with the increasing demand for rapid and high-throughput testing in tea plants traits, digital breeding and smart cultivation of tea plants rely heavily on precise plant phenotypic trait measurement techniques, among which hyperspectral imaging (HSI) technology stands out for its ability to provide real-time and rich-information. In this paper, we provide a comprehensive overview of the principles of hyperspectral imaging technology, the processing methods of cubic data, and relevant algorithms in tea plant phenomics, reviewing the progress of applying hyperspectral imaging technology to obtain information on tea plant phenotypes, growth conditions, and quality indicators under environmental stress. Lastly, we discuss the challenges faced by HSI technology in the detection of tea plant phenotypic traits from different perspectives, propose possible solutions, and envision the potential development prospects of HSI technology in the digital breeding and smart cultivation of tea plants. This review aims to provide theoretical and technical support for the application of HSI technology in detecting tea plant phenotypic information, further promoting the trend of developing high quality and high yield tea leaves.

Quantitative analysis and development of alopecia areata classification frameworks

Alopecia areata is an autoimmune disorder resulting in rapid and unpredictable hair loss on the scalp or body as the immune system mistakenly attacks human hair follicles. In the United States alone, about 6.7 million people experience a form of Alopecia. Early identification of the condition has shown notable potential in improving treatment outcomes and reducing complications. To diagnose Alopecia, researchers have proposed the use of deep learning (DL) techniques to classify images of hair as healthy or alopecia-affected, which has shown high potential. However, the research implementing relevant DL algorithms in the field of hair loss detection and estimation is limited. This paper presents a comparative analysis of our two newly optimized Convolutional neural networks (CNN) with other existing models. For training, we considered datasets comprised of images of healthy hair and alopecia-affected hair. Due to data unavailability, we gathered images of alopecia-affected hair from two distinct datasets: one from Figaro1k and one independently created dataset. After training the algorithms, we performed a contrastive assessment to determine the most effective one based on relevant criteria. We hypothesized that the initial performance of the base neural network would be closely connected to the subsequent accuracy of the algorithm when training it for a new task. As expected, the modified Inception-Resnet-v2 model achieved the greatest performance, with a validation accuracy and loss of 97.94% and 10.4%, respectively. The experimental results indicated that the proposed algorithm serves as an effective framework for Alopecia Areata classification.

Nonlinear Dynamic Granger Causality Analysis Framework for Root-Cause Diagnosis of Quality-Related Faults in Manufacturing Processes

Root-cause diagnosis of quality-related faults plays a crucial role in ensuring stable product quality and high-efficient production for modern manufacturing processes. However, there exists complex nonlinear and dynamic sequential characteristics for process and quality data before and after faults, which contain important fault information. How to fully explore and use this information to locate the root-cause and identify the propagation path is a hot topic. Thus, a new nonlinear dynamic Granger causality analysis method is developed for root-cause diagnosis, which will provide timely and useful reference information to take reasonable measures for field engineers. First, an optimal variable division based on minimal redundancy maximal relevance algorithm is presented to get the quality-related variables. Then, a new attention-based random disturbance gated recurrent unit is designed for nonlinear dynamic Granger causality analysis, aiming at locating the root-cause and identifying the propagation path of quality-related faults. A typical manufacturing process, the hot strip mill process is used for verification. The results show the practicability of the proposed method, and its strong robustness to noise. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —Reasonable and feasible root-cause diagnosis technology will provide timely reference information for field operators to take maintenance measures. However, the nonlinear and dynamic characteristics of time-series data before and after quality-related faults have brought great challenges to the traditional root-cause diagnosis techniques. To address this issue, a practical nonlinear dynamic Granger causality analysis method is proposed, which is purely data driven without knowing complex mechanism knowledge. It will provide feasible solutions for safety monitoring of hot strip mill process as the representative of manufacturing processes.

Construction of a digital twin model for vertical storage tank deformation assessment using terrestrial laser scanning technology

The foundational settlement and deformation of vertical storage tanks are crucial factors influencing their safe operation. To enable rapid deformation assessment of storage tanks, this paper combines point cloud data acquired through terrestrial laser scanning with relevant data processing algorithms to construct a digital twin (DT) model. This achieves high-precision automated detection of tank deformation, facilitating the digital transformation of deformation assessment and offering an integrated detection strategy. First, Euclidean distance clustering is applied to the point cloud, and the point density within clusters is statistically analyzed using a Gaussian distribution. This results in a collection of point clusters within one standard deviation, effectively filtering out outliers and noise points, which facilitates the rapid global registration of the point cloud. Second, in order to quickly segment tank point clouds in the scene, back propagation neural network classification learning based on principal component analysis information is used. The point cloud model is combined with the fitting information of slices to generate a DT model, whose deformation can be evaluated through comparison with appropriate storage tank specifications, taking radial deformation, tank inclination, and foundation settlement as indicators.

Tactical features of the scene inspection in the investigation of the taking of a government official or law enforcement officer as a hostage

The criminal offence of taking a government official or a law enforcement officer hostage under Article 349 of the Criminal Code of Ukraine is becoming increasingly relevant in today’s realities. However, in the course of the study of scientific findings, the absence of methodological recommendations for investigating this crime has been established, which negatively affects the state of combating crime. In particular, the investigation methodology includes recommendations on the specifics of certain investigative (detective) actions, the primary of which is the inspection of the scene. Thus, there is a need to conduct research on the above issue.&#x0D; The purpose of the study is to determine the tactical features of conducting an inspection of the scene of a hostage-taking of a government official or law enforcement officer. To achieve this research goal, the analysis of scientific literature, investigative and judicial practice has been carried out.&#x0D; The article outlines some peculiarities of the tactics of examining the scene of an incident, and identifies the necessary material and technical and forensic support for such an investigative (detective) action. The article also provides examples of possible evidence to be considered at the scene of an event.&#x0D; Based on the results of the study, some tactical techniques for inspecting the scene of a seizure of a government official or a law enforcement officer have been identified, and the inspection has been classified depending on the location: 1) the place of seizure; 2) the place of movement of the seized person; 3) the place of detention of the hostage; 4) the place of detention of the perpetrator. For each place of inspection, the peculiarities and the relevant algorithm of actions by the pre-trial investigation authorities have been identified.&#x0D; It has been established that the purpose of conducting a crime scene investigation of the criminal offences under study is to: establish the location and release of a hostage; establish the location and detention of suspects; identify and seize tools, means of committing a crime and other things relevant to criminal proceedings.

Research on Physical Education Management System in Higher Education Institutions in the Context of Deep Learning

Abstract This paper focuses on the management of physical movement sports disciplines and combines relevant algorithms to design an artificial intelligence-based physical education management system framework. In terms of related algorithms, the first is for the research of character recognition algorithms. Considering the intra-class differences between individual movements, the LR-based feature extraction algorithm is proposed. Secondly, for the research of the action evaluation algorithm, the statistical value, DTW parameter action time series difference and correlation coefficient based on the statistical value and DTW parameter are used as the features, combined with SVM-based feature extraction algorithm for perceiving the action differences. Finally, with the help of deep neural networks for movement analysis, we can improve students’ movement techniques during training. In the course schedule supported by the system, the proportion of moderate-intensity exercise time is high, basically, more than 40%, which meets the WHO requirements for high-intensity exercise time in physical education classes. The post-course RPE feedback from students was spread out between 12 and 15, and the highest post-course RPE value was 14.82&lt;unk&gt;1.38, which was evaluated within a reasonable range of intervals. The teaching management system in this paper can meet the requirements of physical training and teaching, and provide quantifiable basis and management tools for the informatization and precise teaching of physical movement subjects.

Сценарии за развитие на сектор „Биологично земеделие“ в светлината на ОСП 2023 – 2027 година

The research aims to make a scenario analysis that will show what changes, perspectives and trends would be expected in the development of organic farming (OF) in our country, related to the change of main factors, mechanisms and means of support in the policy and with the objectives of the EU Green Deal. The study of the scenarios in the organic farming sector, their evaluation and the measurement of the probability that each of the options will happen is indicative of the future development of this sector, of the opportunities and prospects for it. It follows a quantitative approach to measure the possible changes in the organic farming sector. It is based on and is part of the work on the research project “Perspectives for Bulgarian agriculture and rural areas in the context of the CAP 2021 – 2027 and the EU Recovery Plan”, developed by a team led by associate professor Dr. B. Ivanov, IAI, Agricultural Academy – Sofia in 2021 – 2022. The research is based on the methodology developed for the purposes of the above-mentioned project, using the relevant algorithm, and the main applied methods are comparative assessment, probabilistic study with covariances, descriptive analysis, indicator method, expert assessment and statistical methods for determining trends (Ivanov et al., 2023). When developing the forecast scenarios, to track the possible consequences and effects of assuming different options for the development of OF in our country, two groups of indicators have been used – result and derivative ones. The average scenario score was used as an indicator of how likely the different scenarios were to occur. The specific scenarios and the average scores are determined within limits, according to the standard error and confidence interval according to the methodology of the above-mentioned project. Although the overall probability of the proposed three scenarios (pessimistic, normal and optimistic), measured by the probability coefficient, is unsatisfactory – 0.43, as well as the average scores for the pessimistic and moderate scenarios, the average score for the optimistic scenario is considered satisfactory. From the presented assumptions, estimates and probability scenarios, it can be concluded that the assumed changes in the indicators for the development of organic agriculture lead to relatively close probabilities for the occurrence of the optimistic and moderate scenarios. However, the optimistic scenario has the highest percentage – 18%, followed by the moderate scenario – 16%. I.e. the probability that organic agriculture will develop by 2027 according to the optimistic scenario is the highest. The realization of such a scenario by 2027 would be positive for Bulgarian agriculture in general and for OF in particular. This coincides with the trends that have been noticed recently – there is an increased interest of society in consuming organic products; at the same time, a number of EC- documents set goals for increasing the areas managed by organic farming methods. The increased public interest and concern for the protection of the environment and the health of people and animals has provoked a number of measures in both EU and national policy related to the support of the sector of OF.

Research on the Integration Teaching Mode of Traditional Music Elements and College Piano under the Background of Information Technology

Abstract With the deep integration of information technology and music education, how to integrate traditional music to innovate college piano teaching mode has become a problem that must be solved for the development of the times. Based on the scaffolding teaching method, this paper integrates traditional music elements with college piano in-depth and realizes the innovation of college piano integration teaching mode by adopting online pre-study before class, offline teaching during class and timely reflection and evaluation after class. In the teaching process, traditional music signals are represented by time-frequency domain features so as to realize the digital pre-processing of traditional music signals, and traditional music fundamental tone contours are transformed by using the fundamental tone feature extraction method. Then, based on the principles of physics, we explored the string vibration equation of the piano-generating principle and constructed a mathematical model for the piano music signal. To verify the effectiveness of this paper’s method in piano teaching in colleges and universities, relevant algorithms and teaching practices are used. The results show that the GNSDR of this paper’s method on vocals is improved by 6.58% compared with Unet, and its SDR is 4.25dB and 5.21dB when separating the two music source signals of Piano and Vocals, which is improved by 0.36dB and 0.14dB compared with SHN-4, respectively. The college piano scaffolding teaching model that integrates traditional music elements has enhanced sight-reading ability by 116.44%. Under the background of information technology, it is necessary to fully combine the relevant technology to extract the characteristics of traditional music elements so as to help teachers better carry out the design of piano music teaching programs and improve the quality of colleges and universities to cultivate piano talents who inherit traditional music.

Shorter quantum circuits via single-qubit gate approximation

We give a novel procedure for approximating general single-qubit unitaries from a finite universal gate set by reducing the problem to a novel magnitude approximation problem, achieving an immediate improvement in sequence length by a factor of 7/9. Extending the works \cite{Hastings2017} and \cite{Campbell2017}, we show that taking probabilistic mixtures of channels to solve fallback \cite{BRS2015} and magnitude approximation problems saves factor of two in approximation costs. In particular, over the Clifford+T gate set we achieve an average non-Clifford gate count of 0.23log2&amp;#x2061;(1/&amp;#x03B5;)+2.13 and T-count 0.56log2&amp;#x2061;(1/&amp;#x03B5;)+5.3 with mixed fallback approximations for diamond norm accuracy &amp;#x03B5;.This paper provides a holistic overview of gate approximation, in addition to these new insights. We give an end-to-end procedure for gate approximation for general gate sets related to some quaternion algebras, providing pedagogical examples using common fault-tolerant gate sets (V, Clifford+T and Clifford+T). We also provide detailed numerical results for Clifford+T and Clifford+T gate sets. In an effort to keep the paper self-contained, we include an overview of the relevant algorithms for integer point enumeration and relative norm equation solving. We provide a number of further applications of the magnitude approximation problems, as well as improved algorithms for exact synthesis, in the Appendices.

A feasibility study to explore the use of real‐time caregiver reports for people living with dementia at home to label multi‐modal sensor data

AbstractBackgroundMulti‐modal sensor systems, consisting of sensors throughout the home, are able to collect data on the person living with dementia’s (PlwD’s) activities. However, caregivers may have challenges interpreting the data, leading to a lack of actionable information. It’s important to get real‐time reports from caregivers to label the sensor data to build relevant machine learning algorithms; yet, this has not been done with PlwD. The aim of this feasibility study was to explore if the use of a multi‐modal sensor system in conjunction with real‐time caregiver reports can provide valuable insights into the daily lives of PlwD.MethodWe developed the MAISON (Multimodal AI‐based Sensor platform for Older Individuals) system and deployed it into the homes of dyads (PlwD and their family caregiver) for 8‐weeks. MAISON includes: a smartphone, a smartwatch to continuously collect heart rate, step, and accelerometer data to infer walking, and running, as well as a smart mattress that can collect sleep time and quality. A digital diary was added to the MAISON system that automatically sends the caregiver 2‐3 pre‐set questions every day about their loved one. Caregivers could also report any significant occurrences throughout the day, as they see fit, through the app.ResultTwo dyads were recruited to date. Only about 15% of data was missing from the sensor system by the end of the study. Caregivers were providing their reports with high adherence. The caregivers’ reports were used to label the sensor data successfully, and upon analysis, we found that caregivers’ reports were reflected in the sensor data and the labels derived from their reports were accurate. Results indicate this study methodology is an innovative approach to utilize the family caregiver’s reports.ConclusionThis is the first longitudinal study to demonstrate the feasibility of a multi‐modal sensor system used by caregivers and PwlD. Findings show that caregivers' reports can be useful sources of data for labeling sensor data. The conjunction of caregiver reports and sensor data will provide valuable information to build accurate machine learning models about behaviors of PlwD which is a foundational step to support dyads at home using technology.

Sparse low-redundancy multilabel feature selection based on dynamic local structure preservation and triple graphs exploration

Much semantic information is involved in multilabel data due to more than one label associated with each instance. The redundant features and noise challenge knowledge mining in multilabel data. Constructing a learning model with discriminative features is essential for multilabel learning. Sparse graph-based methods simultaneously consider the topological structure, complex relations between features and labels, and the significance of features. However, three challenges exist. First, they either consider local label correlation or local label relevance and are complementary in the feature selection process. Second, existing methods use low-quality static graphs to explore local label correlations that result in degraded performance. Finally, only some methods deal with redundant features. A ridge regression-based sparse multilabel learning is proposed in this study to address these problems. The global and local label correlation are explored by preserving the instance-level graph structure to obtain a robust low-dimensional pseudo-label matrix to construct a high-quality dynamic label-level graph. Meanwhile, it preserves the feature-level graph structure to select low-redundant features. In addition, a new ℓ2,1/2−2-norm is designed to maintain the high-row sparsity of the model. The above items are embedded into a unified multilabel learning framework. A simple and effective optimization solution is finally designed and compared with eight relevant algorithms on twelve public benchmark data sets. The results demonstrate that the algorithm can improve classification performance.

Refining one-class representation: A unified transformer for unsupervised time-series anomaly detection

The deep unsupervised time-series anomaly detector depends on the one-class representation, which is more effective by only formulating the normal samples. However, normal samples are always mixed with anomalies in the unlabeled training dataset. The learned one-class representation may be biased and violates the one-class setting. To address this problem, we refine the one-class representation and propose a unified AMFormer (Active Masked transFormer) framework, which integrates Transformer with the masked operation mechanism and cost-sensitive learning theory. Specifically, we first develop a network-driven masked operation with Hadamard product transformation to damage the initial input samples. The encoder and decoder representations rebuild the processed incomplete samples, which avoids identical shortcuts and further enhances robustness. Secondly, we exploit the active MSE (Mean Squared Error) loss function to purify the training samples. The different weights are dynamically added to different kinds of samples according to their rebuilding-errors-based pseudo labels. The pseudo anomalies with more significant rebuilding errors are removed by putting lower weights. Finally, extensive experiments are conducted on four benchmark datasets. The experimental results demonstrate that our AMFormer outperforms the nine relevant benchmark algorithms, boosting the mean f1-score from 0.851 to 0.937.

An Open-Source Deep Learning Algorithm for Efficient and Fully Automatic Analysis of the Choroid in Optical Coherence Tomography.

To develop an open-source, fully automatic deep learning algorithm, DeepGPET, for choroid region segmentation in optical coherence tomography (OCT) data. We used a dataset of 715 OCT B-scans (82 subjects, 115 eyes) from three clinical studies related to systemic disease. Ground-truth segmentations were generated using a clinically validated, semiautomatic choroid segmentation method, Gaussian Process Edge Tracing (GPET). We finetuned a U-Net with the MobileNetV3 backbone pretrained on ImageNet. Standard segmentation agreement metrics, as well as derived measures of choroidal thickness and area, were used to evaluate DeepGPET, alongside qualitative evaluation from a clinical ophthalmologist. DeepGPET achieved excellent agreement with GPET on data from three clinical studies (AUC = 0.9994, Dice = 0.9664; Pearson correlation = 0.8908 for choroidal thickness and 0.9082 for choroidal area), while reducing the mean processing time per image on a standard laptop CPU from 34.49 ± 15.09 seconds using GPET to 1.25 ± 0.10 seconds using DeepGPET. Both methods performed similarly according to a clinical ophthalmologist who qualitatively judged a subset of segmentations by GPET and DeepGPET, based on smoothness and accuracy of segmentations. DeepGPET, a fully automatic, open-source algorithm for choroidal segmentation, will enable researchers to efficiently extract choroidal measurements, even for large datasets. As no manual interventions are required, DeepGPET is less subjective than semiautomatic methods and could be deployed in clinical practice without requiring a trained operator. DeepGPET addresses the lack of open-source, fully automatic, and clinically relevant choroid segmentation algorithms, and its subsequent public release will facilitate future choroidal research in both ophthalmology and wider systemic health.