Classification Function Research Articles

Social entertainment robot based on neural network algorithm in personalized music course simulation

This study focuses on the simulation of personalized music courses using social entertainment robots based on neural network algorithms, and introduces the current research status of music processing algorithms based on neural networks. A detailed explanation was provided on the training methods for neural network algorithms, including BP neural network training and music generation algorithms. Research was conducted on activation functions, music output detection, and loss function detection. A virtual music course system framework has been designed to provide students with personalized learning experiences. In a social entertainment robot based on neural network algorithms, a complete control node system was reconstructed, aiming to achieve interactive and entertainment effects of music courses through sensing technology. The control node system includes multiple key components, and the robot is able to perceive and recognize the actions, facial expressions, and sounds of students. Social entertainment robots have intelligent classification and recommendation functions, and this algorithm can intelligently classify and recommend music resources based on students’ music preferences, levels, and needs. By analyzing the personal preferences and learning progress of students, robots can provide customized music course content and learning materials for each student, improving learning effectiveness and personal interests.

Dynamic Security Rule Optimization Based on Deep Learning and Adaptive Algorithms

Compared with traditional computer network security identification techniques, deep learning algorithms are based on their own distributed network structure for storage, processing, classification, comparison, and automatic identification functions. The distribution and dynamism of cloud databases increase the difficulty of route prediction and recognition in the cloud, affecting the efficiency of cloud computing. In response to the above issues, the author proposes a dynamic path optimization process for cloud databases based on adaptive immune grouping polymorphic ant colony algorithm. By setting up two states of ant colony, reconnaissance ant and search ant, and introducing an adaptive polymorphic ant colony competition strategy, the defect of general ant colony algorithms being prone to falling into local optima is improved; On this basis, an artificial immune algorithm with fast global search capability is further integrated to improve the search ant path optimization process, improving search speed and accuracy. Simulation experiments show that the IPANT algorithm outperforms the other three algorithms, maintaining a throughput of 1000 kbps and relatively stable; The data of OSPF, SPF, and FR are not significantly different, significantly lower than IPANT. The immune polymorphic ant colony algorithm (IPANT) has the lowest time delay for packet routing and performs better than the other three algorithms, with FR and SPF having higher latency. It has been proven that this algorithm can better solve convergence speed and global optimization problems, and can quickly and reasonably find the database to be accessed in the cloud.

Robust Fisher-regularized extreme learning machine with asymmetric Welsch-induced loss function for classification

Robust Fisher-regularized extreme learning machine with asymmetric Welsch-induced loss function for classification

Optimal evolutionary framework-based activation function for image classification

Typically, supervised Machine Learning (ML)-based image classifiers leverage algorithms derived from either Artificial Neural Networks (ANNs) or optimal separating hyperplane (OSH)-based algorithms. However, despite recent progress has been made to enhance ANNs’ classification performance via the Rectified Linear Unit (ReLU)-based activation functions (AFs), there is currently no AF that scales across and benefit both ANNs and OSH-based classifiers. Moreover, the lack of globally optimal AFs leads to a high variance in image classification-related results. Thus, this study seeks to overcome this limitation by implementing a next-generation evolutionary framework (‘ActiGen’) to generate a novel and more reliable AF, which can scale to two families of AFs for two classifiers. The proposed evolutionary knowledge-based framework leverages a Multi-Objective (MO) optimisation method based on Genetic Algorithms (GA), or ‘MOGA’, to improve the generalisation of such classifiers. This evolutionary framework and its generated AF are validated using nine open-access datasets: seven image-based datasets, consisting of 22,136 images in total, and two large (561 features for 10,929 instances, 124 features for 1,700 instances) tabular datasets. These diverse datasets include both binary and multi-class classification, such as images of breast masses, those acquired via cardiac computed tomography, photos of famous people from the Internet, images of handwritten digits and those drawn on a graphics tablet, human faces with different lighting, details, and expressions, smartphone-related data captured during various activities and postural transitions, and clinical data on complications of myocardial infarction. Findings demonstrate that the proposed evolutionary optimisation framework (‘ActiGen-MOGA’) was able to generate a novel scalable AF, which led to achieve the highest classification performance and the fastest convergence across six out of nine datasets. In the best classification task, the ActiGen-MOGA-based AF led to a classification performance of 80 % and 78 % higher than the polynomial and Rectified Linear Unit (ReLU) AFs respectively.

Deep learning model for enhanced power loss prediction in the frequency domain for magnetic materials

AbstractThis paper outlines the methodology for predicting power loss in magnetic materials. It starts by introducing the concept of core loss and the complexity of modelling it. Steinmetz's equation is presented to calculate power loss based on frequency and magnetic flux density, but its limitations are highlighted. As an alternative, a neural network‐based method is introduced. The proposed methodology adopts a long short‐term memory network, expressing the core loss as a function of magnetic flux density, frequency, temperature, and wave classification. Fast Fourier transform was implemented to reduce the data points of the sampled flux density waveform while preserving its characteristics. Analyzing in the frequency domain enabled streamlining the training of the model. The input features were arranged as required, and the network architecture was designed with appropriate layers and optimal activation functions. Through extensive training using the datasets, the model assimilated intricate relationships between input variables and known power loss. Evaluation and validation metrics were subsequently employed to gauge the performance of the trained network. This innovative methodology aims to significantly augment the precision of power loss predictions, providing valuable insights into the nuanced behaviour of magnetic materials.

#2334 Elevated serum admission creatinine in acute heart failure hospitalizations: lifting the veil to understand outcome differences

Abstract Background and Aims Organ crosstalk in heart failure and kidney dysfunction is prevalent and is associated to a high event rate. In this setting elevated serum creatinine (sCr) at acute decompensated heart failure (ADHF) admissions is a strong risk factor for poor outcomes. Current accepted cardiorenal syndrome (CRS) classification is based on timing and initial organ dysfunction and fails at capturing its nature and prognosis. We sought to describe patient (P) profile and outcomes in P hospitalized for ADHF with high admission sCr in function of current CRS classification and assess for stratifying predictors. Method We conducted an observational, analytical and cross-sectional study, with prospective and consecutive data acquisition. We included P over 18 years-old, who were admitted to the Intensive Care Telemetry Unit in our Hospital between July, 2011 and Dec, 2023 due to ADHF. P who received a heart transplant during index admission or those on chronic dialysis were excluded. Clinical, epidemiological and biochemical variables were registered. Renal dysfunction was defined under admission sCr >1.4 mg/dl and divided according to Ronco et al classification, as initially cardiac (type 2) or renal (type 4) origin. We constructed multivariate logistic and survival regression models, with selection of relevant predictors by resampling and regularization of a Lasso regression model. Statistical significance was considered with a 5% of alpha error level. Analysis was conducted using R (R Core Team, 2023). Results A total of 1868 consecutive P were included. P population consisted in type 2: 124 and type 4: 258 P, respectively. We found no differences in gender (p = 0.47), while type 4 were older (75 ± 12 vs 71 ± 12; p < 0.01). Ventricular dysfunction was more severe in type 2 (left ventricular ejection fraction 34 ± 17 vs 41 ± 16%; p < 0.001). While ischaemic heart disease was more prevalent in type 4 (32.8 vs 23.5%; p = 0.001), idiopathic cardiomyopathy was more frequent in type 2 (18.8 vs 7.2%; p = 0.001). Clinical phenotypes differed between groups, being clinical hypoperfusion more frequent in type 2 (32.4 vs 12.3%; p < 0.001). Admission sCr was higher in type 4 (1.9 [0.8] vs 1.6 [0.58] mg/dl; p < 0.001), as was 1st day natriuresis (81 vs 67 mEq; p = 0.01). Both groups experienced a similar diuretic resistance rate (21.2 vs 17.1%; p = 0.34), while type 4 received rescue ultrafiltration more frequently (11.3 vs 3.5%; p < 0.01). Type 2 received a higher inotropic support rate (47.1 vs 19.5%; p < 0.01) and had a longer length of stay (9 vs 7 days; p = 0.01). In-hospital (11.8 vs 11.3%; p = 0.99); 6-month and 1-year mortality rates (66.8 vs 72.8% and 55.1 vs 64.6% respectively; p = 0.3). In a similar way, 1-month (94.8 vs 94.7%), 6-month (72 vs 74.7%) and 1-year readmission rates (59.2 vs 60.9%; p = 0.6) were similar in both groups. A multiple logistic regression model (AUCROC 0.82) identified admission sCr (OR 2.6; CI 95% 2.2-3.2; p < 0.001), inotropic need (OR 1.87; CI 95% 1.4-2.6; p < 0.001), anasarca phenotype (OR 1.63; CI 95% 1.06-2.5; p = 0.02), moderate to severe right ventricular dysfunction (OR 1.4; CI 95% 1.03-1.9; p = 0.03) as independent predictors for a descending sCr behavior during hospitalization, while the lower T3 levels were associated with less probability for a descent of sCr (OR 0.21; CI 95% 0.1-0.46; p < 0.001). Conclusion Organ crosstalk in ADHF hospitalizations was associated with a high event rate. One-year mortality and readmission rates were similar in both groups. Renal support therapies were more frequent under lower kidney reserve conditions. Ventricular dysfunction and systemic congestion phenotype are relevant to a better understanding for potential reversibility at the admission time and may promote actions for outcome enhancing.

Sex Differences in the Trajectories of Cognitive Decline and Affected Cognitive Domains Among Older Adults With Controlled and Uncontrolled Glycemia.

We aimed to analyze the trajectories of cognitive decline as a function of the presence of type 2 diabetes and glycemic control in analyzes stratified by sex in an 8-year follow-up period. A total of 1752 men and 2232 women aged ≥50 years who participated in the English Longitudinal Study of Ageing (ELSA), conducted from 2004 to 2012, were analyzed. The outcomes of interest were performance on the cognitive domains of memory, executive function, and temporal orientation as well as the global cognition score. Cognitive performance was standardized in z-scores in strata based on schooling and age. The participants were classified as without diabetes, with controlled glycemia, and with uncontrolled glycemia, according to medical diagnosis, glucose-lowering medications use and HbA1c levels. Generalized linear mixed models controlled by sociodemographic, behavioral, and health-related characteristics were used for the trajectory analyses. No differences in z-scores were found for global cognition or cognitive domains based on diabetes classification in men and women at baseline. More than 8 years of follow up, women with uncontrolled glycemia had a greater decline in z-scores for global cognition (-0.037 SD/year [95% CI: -0.073; -0.001]) and executive function (-0.049 SD/year [95% CI: -0.092; -0.007]) compared with those without diabetes. No significant difference in trajectories of global cognition or any cognitive domain was found in men as a function of diabetes classification. Women with uncontrolled glycemia are at greater risk of a decline in global cognition and executive function than those without diabetes.

Time dimension feature extraction and classification of high-dimensional large data streams based on unsupervised learning

In order to solve the problem of low accuracy of time dimension feature extraction and classification of high-dimensional large data streams, this paper proposes a time dimension feature extraction and classification algorithm of high-dimensional large data streams based on unsupervised learning. Analyze the trend of high-dimensional data flow changes under machine learning, and achieve dimensionality reduction of high-dimensional large traffic time dimensional data through local save projection. Analyze the spatial relationship between feature attributes and feature space, segment and fit high-dimensional big data streams and time dimensional feature data streams, further segment time dimensional sequences using sliding windows, and complete feature extraction through discrete dyadic wavelet transform. According to the clustering algorithm, cluster the time dimension feature data stream, calculate the cosine similarity of the feature data, model the time dimension feature stream of training samples, use the feature classification function to minimize the classification loss, and use unsupervised learning to achieve the final classification task. The test results show that this method can improve the temporal feature extraction and classification accuracy streams.

MCAD: Multi-classification anomaly detection with relational knowledge distillation

With the wide application of deep learning in anomaly detection (AD), industrial vision AD has achieved remarkable success. However, current AD usually focuses on anomaly localization and rarely investigates anomaly classification. Furthermore, anomaly classification is currently requested for quality management and anomaly reason analysis. Therefore, it is essential to classify anomalies while improving the accuracy of AD. This paper designs a novel multi-classification AD (MCAD) framework to achieve high-accuracy AD with an anomaly classification function. In detail, the proposal model based on relational knowledge distillation consists of two components. The first one employs a teacher–student AD model, utilizing a relational knowledge distillation approach to transfer the interrelationships of images. The teacher–student critical layer feature activation values are used in the knowledge transfer process to achieve anomaly detection. The second component realizes anomaly multi-classification using the lightweight convolutional neural network. Our proposal has achieved 98.95, 96.04, and 92.94% AUROC AD results on MNIST, FashionMNIST, and CIFAR10 datasets. Meanwhile, we earn 97.58 and 98.10% AUROC for AD and localization in the MVTecAD dataset. The average classification accuracy of anomaly classification has reached 76.37% in fifteen categories of the MVTec-AD dataset. In particular, the classification accuracy of the leather category has gained 95.24%. The results on the MVTec-AD dataset show that MCAD achieves excellent detection, localization, and classification results.

Knowledge-based system and expectation-maximization to discovering causes of imperfect labels in vehicular networks clustering

This paper focuses on the development of a knowledge-based system for automatically diagnosing issues in Vehicular ad hoc networks (VANETs). VANETs enable communication between vehicles and infrastructure, enhancing road safety and efficiency through timely information exchange. The proposed system aims to efficiently maintain and ensure the continuity of network service by leveraging innovative pattern recognition methods tailored to VANETs. The automatic diagnosis problem in VANETs involves estimating the operating class of network components based on sensor observations. This entails associating sensor measurements with specific operating modes. By implementing condition-based preventive maintenance procedures, potential component failures can be detected early, mitigating network disruptions. Various approaches, such as expert systems, fault trees, network state models, and statistical learning through pattern recognition, can be employed to address this problem. This paper primarily focuses on the statistical learning approach, where a classification or regression function is learned from a set of examples to assign operation modes to new measurements. It discusses relevant metrics and preprocessing techniques to simplify the decision-making process. The diagnostic system's results are determined based on the formulation of the classification or regression problem. The learning base is constructed, and an appropriate classification method is selected to develop and validate the automatic diagnosis system. While non-parametric models like support vector machines are commonly used, this article emphasizes the significance of considering assumptions and leveraging additional information to enhance performance. It proposes a more specific formalization of the problem, integrating the unique characteristics of VANETs. The contributions of this article revolve around the theory of belief functions, a generative approach, and the utilization of parametric models defined using graphical models. Experimental studies conducted on artificial datasets have demonstrated the benefits of the semi-supervised approach within the context of VANET networks.

Application of neural network algorithm in computer data information processing system

In order to solve the problem that the data processing frequency of the traditional centralized data processing system is low, which leads to the poor feedback effect of massive data, this paper proposes a computer data information processing system based on neural network algorithm. Based on the original system hardware, the system replaces the data processor and increases the total number of the processor, so as to realize the distributed synchronous processing of massive data. In the aspect of software design, the interaction between system units and modules is formed through the protocol to improve the data communication mode of the system; Calculate the Euclidean distance and set the distributed processing mode of the system; According to the definition of cloud computing, the classification function is used to determine the constraints, establish the processing frequency equation, and realize the rapid processing of massive data. The experimental results show that the processing frequency of the system designed in this paper is more than 50 % higher than that of the traditional system, which solves the problem of low processing frequency caused by weak analysis ability and slow response speed in the traditional system. Conclusioncompared with the traditional system, the system designed in this paper has faster processing frequency for massive data and better feedback effect to users.

Optimizing erectile function evaluation pattern in postoperative pelvic fracture-urethral injury patients: introducing PFUI pEF PROM.

To establish a psychometric validated pelvic fracture-urethral injury postoperative erectile function patient reported outcome measure (PFUI pEF PROM). We also aim to explore PFUI patients' potential classification and suitable postoperative erectile function assessment pattern. A total of 93 PFUI patients who treated by excisionandprimaryanastomotic (EPA) urethroplasty from January 2020 to August 2022 were enrolled to this study. Patients who had intercourse completed the IIEF-5, those who had sexual stimulation other than intercourse finished PFUI pEF PROM, and those who performed both completed the IIEF-5, and PFUI pEF PROM. Erection Hardness Score (EHS) was completed by all patients. This PROM was performed psychometric validation and used to find PFUI patients potential classification through latent class analysis. Then, we determined the cut-off value though receiver-operating characteristic (ROC) curve and performed univariate analysis subgroups feature. The PFUI pEF PROM demonstrated high reliability and validity with a Cronbach's alpha of 0.928. It correlated significantly with IIEF-5 (r = 0.550, p < 0.001) and EHS (r = 0.909, p < 0.001). The latent class analysis identified three patient subgroups, with 14.5 as the subgroup cut-off value. Urethral stricture length, IIEF-5, and EHS score were identified as influence factors for maximal erection potential. An assessment pattern combining IIEF-5, EHS, and the PFUI pEF PROM covered 92.5% of patients. This PROM effectively addresses the current limitation in assessing erectile function in PFUI patients. This study provides a promising tool for stratified assessment, prediction erection recovery, and treatment guidance in the PFUI Erectile dysfunction field.

A biochemical feedback signal for hypothermia treatment for neonatal hypoxic-ischemic encephalopathy: focusing on central nervous system proteins in biofluids.

Hypothermia has been widely used to treat moderate to severe neonatal hypoxic-ischemic encephalopathy (HIE), yet evaluating the effects of hypothermia relies on clinical neurology, neuroimaging, amplitude-integrated electroencephalography, and follow-up data on patient outcomes. Biomarkers of brain injury have been considered for estimating the effects of hypothermia. Proteins specific to the central nervous system (CNS) are components of nervous tissue, and once the CNS is damaged, these proteins are released into biofluids (cerebrospinal fluid, blood, urine, tears, saliva), and they can be used as markers of brain damage. Clinical reports have shown that CNS-specific marker proteins (CNSPs) were early expressed in biofluids after brain damage and formed unique biochemical profiles. As a result, these markers may serve as an indicator for screening brain injury in infants, monitoring disease progression, identifying damage region of brain, and assessing the efficacy of neuroprotective measures. In clinical work, we have found that there are few reports on using CNSPs as biological signals in hypothermia for neonatal HIE. The aim of this article is to review the classification, origin, biochemical composition, and physiological function of CNSPs with changes in their expression levels after hypothermia for neonatal HIE. Hopefully, this review will improve the awareness of CNSPs among pediatricians, and encourage future studies exploring the mechanisms behind the effects of hypothermia on these CNSPs, in order to reduce the adverse outcome of neonatal HIE.

Lauric acid content in intramuscular fat is a reliable indicator of black soldier fly larvae meal consumption in Muscovy ducks

The present research investigated if intramuscular fatty acid (FA) profile could distinguish meat from ducks fed with black soldier fly larvae meal (BSFLM) during fattening. By using stepwise linear discriminant analysis on FA profiles of 96 meat samples, lauric acid (C12:0) was found to be the best predictor, accurately differentiating samples with only two misclassifications. The Fisher classification functions indicated that breast samples with lauric acid content above 0.222 % in intramuscular fat would be classified as BSFLM-fed. The Fisher classification functions were also effective in other poultry species fed BSFLM, with only two misclassified samples out of 42 samples from a dataset compiled from published papers. Misclassifications were linked to an unexpected lauric acid content in the intramuscular fat. In conclusion, given that BSFLM is the only relevant source of lauric acid in the diet, this fatty acid could identify meat from BSFLM-fed ducks and possibly other poultry species, warranting further exploration of other FA as potential identifiers of BSFLM-fed poultry.

The Inflammatory State of Follicular Fluid Combined with Negative Emotion Indicators can Predict Pregnancy Outcomes in Patients with PCOS.

Polycystic ovary syndrome (PCOS) is a complex endocrine disorder syndrome with an incidence of 6% to 10% in women of reproductive age. Women with PCOS not only exhibit abnormal follicular development and fertility disorders, but also have a greater tendency to develop anxiety and depression. Our aim was to evaluate the ability of inflammatory factors in follicular fluid to predict embryonic developmental potential and pregnancy outcome and to construct a machine learning model that can predict IVF pregnancy outcomes based on indicators such as basic sex hormones, embryonic morphology, the follicular microenvironment, and negative emotion. In this study, inflammatory factors (CRP, IL-6, and TNF-α) in follicular fluid samples obtained from 225 PCOS and 225 non-PCOS women were detected via ELISA. For patients with PCOS, the levels of CRP and IL-6 in the follicular fluid in the pregnant group were significantly lower than those in the nonpregnant group. For non-patients with PCOS, only the level of IL-6 in the follicular fluid was significantly lower in the pregnant group than in the nonpregnant group. In addition, for both PCOS and non-patients with PCOS, compared with those in the pregnant group, patients in the nonpregnant group showed more pronounced signs of anxiety and depression. Finally, the factors that were significantly different between the two subgroups (pregnancy and nonpregnancy) of patients with or without PCOS were identified by an independent sample t test first and further analysed by multilayer perceptron (MLP) and random forest (RF) models to distinguish the two clinical pregnancy outcomes according to the classification function. The accuracy of the RF model in predicting pregnancy outcomes in patients with or without PCOS was 95.6% and 91.1%, respectively. The RF model is more suitable than the MLP model for predicting pregnancy outcomes in IVF patients. This study not only identified inflammatory factors that can affect embryonic development and assessed the anxiety and depression tendencies of PCOS patients, but also constructed an AI model that predict pregnancy outcomes through machine learning methods, which is a beneficial clinical tool.

Loss Relaxation Strategy for Noisy Facial Video-based Automatic Depression Recognition

Automatic depression analysis has been widely investigated on face videos that have been carefully collected and annotated in lab conditions. However, videos collected under real-world conditions may suffer from various types of noise due to challenging data acquisition conditions and lack of annotators. Although deep learning (DL) models frequently show excellent depression analysis performances on datasets collected in controlled lab conditions, such noise may degrade their generalization abilities for real-world depression analysis tasks. In this article, we uncovered that noisy facial data and annotations consistently change the distribution of training losses for facial depression DL models; i.e., noisy data–label pairs cause larger loss values compared to clean data–label pairs. Since different loss functions could be applied depending on the employed model and task, we propose a generic loss function relaxation strategy that can jointly reduce the negative impact of various noisy data and annotation problems occurring in both classification and regression loss functions for face video-based depression analysis, where the parameters of the proposed strategy can be automatically adapted during depression model training. The experimental results on 25 different artificially created noisy depression conditions (i.e., five noise types with five different noise levels) show that our loss relaxation strategy can clearly enhance both classification and regression loss functions, enabling the generation of superior face video-based depression analysis models under almost all noisy conditions. Our approach is robust to its main variable settings and can adaptively and automatically obtain its parameters during training.

Research on PSO‐SVM base wine grade recognition based on Max‐Relevance and Min‐Redundancy feature selection

SummaryIn response to the challenges inherent in Baijiu classification, characterized by ambiguity and limited methodologies, this study introduces a novel framework for comprehensive Baijiu brewing process management. By integrating mathematical models and machine learning algorithms, our aim is to standardize and enhance the accuracy of the Baijiu brewing process. Through a meticulous analysis of 13 input features using the MRMR algorithm, we identified key variables that significantly influence base wine grades. Subsequently, we compared the efficacy of support vector machine (SVM) algorithms under three kernel functions for base wine grading classification. Additionally, we developed an enhanced particle swarm‐support vector machine (IPSO‐SVM) algorithm, which incorporates nonlinearly decreasing inertia weights to enhance global and local search capabilities. Evaluation metrics, including accuracy and F1‐SCORE, were employed to assess the model's performance. Our experimental results demonstrate that IPSO‐SVM significantly improves convergence speed and accuracy, achieving a classification accuracy of 98.31% on the test set with an F1‐SCORE of 98.18%. Furthermore, experimental tests conducted at a prominent strong‐flavored wine enterprise in China validated the model's effectiveness, revealing no significant difference in grading quality between model and manual Baijiu classification, with a mean error below 1%. This study presents a pioneering approach to digitize and standardize Baijiu classification processes, laying theoretical foundations for the development of intelligent Baijiu classification systems.

Stock Trend Prediction and Analysis Based on Machine Learning

Stock trend prediction is a critically important activity in the financial sector. Investors aim to gain a better understanding of market dynamics, make wiser investment decisions, and manage investment risks more effectively by forecasting stock price trends. This is the focus of our upcoming research. This paper aims to utilize machine learning methods such as Support Vector Machines (SVM), based on spatial interval maximization strategies, Random Forest, based on multi-decision tree parallel learning strategies, and eXtreme Gradient Boosting (XGBoost), based on classification functions, to devise accurate classification function strategies for predicting and analyzing fluctuations in Dow Jones stock prices. Principal Component Analysis (PCA) will then be used to select some suitable features, identifying key factors with the greatest impact and minimal inter-correlation related to stock prices. Through a combination of various classification models, the research aims to forecast the future rise and fall of Dow Jones stocks. It has been observed that performing feature curation through PCA in advance enhances the predictive accuracy of different machine learning methods. Notably, the combination of PCA and XGBoost achieves the highest predictive rate. This approach provides a robust and scientifically grounded method for future stock price predictions and the formulation of trading strategies.

Fractional cross entropy-based loss function for classification of IoT services with semantic graph based on IFTTT recipes

Fractional cross entropy-based loss function for classification of IoT services with semantic graph based on IFTTT recipes

System reliability analysis with small failure probability based on relevant vector machine and Meta-IS idea

Structural failure is a complex system problem and traditional methods face difficulties in solving system reliability problems with small failure probability. This paper proposes an efficient sampling simulation method termed SDMI, which integrates metamodel-based importance sampling (Meta-IS) with spherical decomposition (SD). SDMI transforms the failure probability into a product of spherical augmented failure probability and a correction factor. The implementation of SDMI for each component further facilitates the proposition of a system probability classification function. The integration of the relevant vector machine (RVM) leads to the proposal of a more efficient methodology, termed RVM-SDMI. The efficacy and precision of the proposed methods have been validated through four numerical examples and two engineering examples. The results demonstrate that SDMI successfully addresses the convergence challenges inherent in Meta-IS. Moreover, RVM-SDMI exhibits remarkable accuracy and computational efficiency in the system reliability analysis with small failure probabilities.