We present S-Path-RAG, a semantic-aware shortest-path Retrieval-Augmented Generation framework designed to improve multi-hop question answering over large knowledge graphs. S-Path-RAG departs from one-shot, text-heavy retrieval by enumerating bounded-length, semantically weighted candidate paths using a hybrid weighted $k$-shortest, beam, and constrained random-walk strategy, learning a differentiable path scorer together with a contrastive path encoder and lightweight verifier, and injecting a compact soft mixture of selected path latents into a language model via cross-attention. The system runs inside an iterative Neural-Socratic Graph Dialogue loop in which concise diagnostic messages produced by the language model are mapped to targeted graph edits or seed expansions, enabling adaptive retrieval when the model expresses uncertainty. This combination yields a retrieval mechanism that is both token-efficient and topology-aware while preserving interpretable path-level traces for diagnostics and intervention. We validate S-Path-RAG on standard multi-hop KGQA benchmarks and through ablations and diagnostic analyses. The results demonstrate consistent improvements in answer accuracy, evidence coverage, and end-to-end efficiency compared to strong graph- and LLM-based baselines. We further analyze trade-offs between semantic weighting, verifier filtering, and iterative updates, and report practical recommendations for deployment under constrained compute and token budgets.

Accurately forecasting power consumption in data center servers requires addressing the temporal distribution shift caused by dynamic resource demands. However, existing methods rely on global normalization, which cannot capture short-term localized shift, leading to unsatisfactory performance when forecasting non-stationary time series. To address this challenge, we propose a novel bi-level optimization framework for forecasting non-stationary long-term power consumption, namedNoSPF. The framework employs hierarchical optimization to separately model the stationary and local non-stationary components of power consumption time series, offering a flexible, model-agnostic paradigm for time-series forecasting. Using Discrete Wavelet Transform (DWT) for multi-scale time–frequency analysis,NoSPFdecomposes the series into non-stationary components driven by short-term fluctuations and stationary components that capture long-term trends. Furthermore,NoSPFintegrates a lightweight Multi-Layer Perceptron (MLP) to predict the local non-stationary components, enhancing the framework’s forecasting accuracy by providing more precise approximations of the future power distribution. Extensive experiments on real-world server datasets demonstrate the superior performance and effectiveness ofNoSPF.

In resource-constrained edge computing, the execution efficiency of workflow applications is significantly affected by bandwidth contention, especially during data transmissions between dependent tasks. However, existing workflow scheduling studies often struggle to optimize transmission delay effectively, whereas bandwidth slicing offers promising potential by leveraging the dynamic nature of bandwidth resources. To address this issue, we propose Workflow Scheduling with Dynamic Bandwidth Slicing (WSDBS), a novel scheduling algorithm that integrates bandwidth slicing into the workflow execution process. By introducing a dual-prediction strategy, WSDBS estimates the availability of both computational and bandwidth resources on servers, facilitating efficient task scheduling decisions under transmission uncertainty. Moreover, a novel transmission urgency metric is developed, which is derived from both link load and transmission criticality. This metric guides bandwidth slicing for the dynamic allocation of server-side bandwidth resources, ultimately alleviating contention among concurrent transmissions. Extensive experiments based on real-world Alibaba cluster traces show that WSDBS consistently improves scheduling efficiency, reducing the average makespan by 10.87%-14.44% over state-of-the-art baselines. These results validate its effectiveness in alleviating bandwidth contention and improving scheduling performance in edge computing environments.

Weight perception is essential for delivering compelling and realistic object interaction in VR systems. While existing pseudo-haptic techniques have enabled users to perceive virtual object weight without physical actuation, their application has primarily been limited to near-field, direct hand interaction. As VR systems continue to advance in fidelity and versatility, interactions with objects beyond arm's reach are becoming increasingly common. However, how weight perception can be introduced into such remote interactions, and whether it can enhance user immersion and experience, remains underexplored. To bridge this gap, we present RayFlex, a pseudo-haptic technique that conveys object weight in raycasting-based interaction through visual displacement and dynamic ray deformation. The technique was evaluated in two user studies, which examined its effectiveness in supporting weight discrimination and its impact on user experience across different interaction contexts. Results indicate that RayFlex leads to significant improvements in perceived realism, presence, and satisfaction, while maintaining usability. From the results, we derived two implications and an application guideline that can help design future VR systems.

In the field of the Internet of Medical Things (IoMT), efficiently deploying deep learning models for breast cancer pathological diagnosis is both important and challenging. Traditional cloud computing stores data on a central server, which poses a risk of data leakage. Although conventional federated learning can protect data privacy, it still faces limitations in computing resources. In addition, most existing pathological image segmentation methods use single-resolution features, making it difficult to jointly represent both global context and local lesion details. To address these challenges, we introduce a Breast Cancer Pathological Diagnosis Cloud-Edge-Collaborative Platform based on split federated learning. We developed a lightweight edge device that is built upon the Jetson Nano deep learning computing platform. The model uses heatmap-guided region search to efficiently discover and examine ROIs over histopathological images, which simulates the expert observation process by pathologists. The platform enables collaborative learning of global and local information without leaking out sensitive patients’ data. Our method ensures data security and privacy while significantly improving model accuracy. Experimental results show that our model achieves Dice scores of 81.31% on the public Breast Cancer Semantic Segmentation (BCSS) dataset and 87.03% on clinical data collected from the Chinese People’s Liberation Army (PLA) General Hospital, improving the Dice score from 79.36% to 81.31% on BCSS, and from 86.18% to 87.03% on our clinical dataset. These results outperform existing models, advancing a meaningful step towards the distributed pathological diagnosis of breast cancer.

ABSTRACT Learning path recommendation is crucial for guiding learners through a series of courses in a logical sequence based on their previous learning experiences. This is particularly important for improving learning outcomes in massive open online courses (MOOCs) for diverse learners. Because both the historical learning courses and recommended learning paths can be represented as sequential patterns (SPs); it is reasonable to approach this problem through SP mining (SPM). In addition to support, we incorporate three factors, that is, course learning days, grades and engagement, to model frequent high‐utility SPs (FHUSPs). When recommending a learning path, FHUSPs that align with the target user's learning history and are common among successful learners, while rare among less successful ones, are prioritised. If there are insufficient matching FHUSPs, we address this by recommending additional courses based on the joint competency and complementarity of learners similar to the target learner. Experimental results on a real‐world dataset demonstrate that our method provides highly accurate and relevant recommendations.

Currently, with the rapid advancement of aerospace technology, scientists are increasingly capable of exploring other planets, which has brought greater attention to the challenge of crater detection, particularly the detection of smaller craters. This letter presents a novel single-stage target detector based on YOLOv9, aimed at improving the detection of craters, especially small ones. First, we propose a novel feature extraction module that enhances the detection capabilities of the network. By integrating deformable convolution into the original feature extraction module, we improve the capability of convolutional neural network (CNN) to catch long-range dependencies and spatial relationships, thereby enhancing the detection accuracy for craters of various sizes. Second, we introduce a new pooling structure called averaged spatial pyramid pooling (ASPP). This structure uses a parallel configuration of average and maximum pooling techniques to enrich the overall feature extraction process, thereby improving the detection capability for small craters. To confirm the efficacy of our proposed approach, we performed comprehensive experiments using a large public Mars crater dataset. The results indicate that our approach significantly surpasses most current mainstream one-stage object detection algorithms in both precision and recall.

Satellite remote sensing, with its wide coverage, long time series, and high revisit frequency, has become an indispensable tool for CO2monitoring. However, mainstream XCO2datasets derived from satellite observations typically feature low spatial resolution, which limits the practical applications of these valuable satellite data. To overcome this limitation, we propose a multifactor Deep Forest Regression Stepwise Downscaling (DFRSD) Framework using the existing global monthly and Gap-Free column-averaged dry-air mole fraction of CO₂ (GF-XCO2) resolution of 0.1°. This approach introduces an intermediate resolution level (0.05°) between the initial resolution (0.1°) and target resolution (1km). Specifically, column-averaged dry-air mole fraction of CO₂ (XCO2) at resolution of 0.1° is first downscaled to that of 0.05°, and then further refined to 1 km resolution. The Deep Forest Regression (DFR) model is used to establish relationships between XCO2 and auxiliary variables at each downscaling step, including near-surface air pollutant, Normalized Difference Vegetation Index (NDVI), Temperature (TMP), and Precipitation (PRE). We successfully enhance the XCO2spatial resolution of 0.1° to that of 1 km, generating high-resolution monthly spatial resolution products of 1 km for the period from January 2016 to December 2020. To validate the accuracy of the reconstructed high-resolution XCO2, we compare it with field measurements from ground monitoring stations. The results demonstrate a strong agreement between the downscaled XCO2and field observations, with R² of 0.912 and RMSE of 1.10ppm, highlighting its reliability and precision. The results demonstrate the effectiveness of using SRF and Multifactor methods for XCO₂ downscaling. The proposed method not only significantly enhances spatial resolution but also preserves spatial distribution integrity, which satisfies the precision requirements of various research applications.

Achieving real-time, remote control and precise localization of early gastric cancer (EGC) lesions in endoscopic capsules is a significant obstacle in biomedical imaging development. This work outlines an innovative integrated system that uses an effective combination of Zynq UltraScale+ and Gizwits IoT to overcome this challenge. Our work employs a fully convolutional neural network underpinning grid-search clustering-driven multi-modals gating information local patch learning (GS-MGIF-LPLs). This system, designed as an adaptive location computing acceleration platform (ACAP), elegantly marries a double threshold fast search strategy with patch-based FCNN, fueling efficient training, testing, and performance metrics with an accuracy of 99.53%, precision coefficient of 86.06%, and an IoU of 84.26%. Upon benchmarking against four EGC types, our GS-MGIF-LPLs model demonstrates exceptional superiority against five established methods, providing a significant stride in computational efficiency and diagnostic advancements for gastrointestinal diseases.

Medical image classification has beenan important application for Deep Learning techniques for over a decade, and since the emergence of Explainable AI (XAI), researchers have started using XAI to validate the results produced by these black box models. In the research field, it has become clear that accuracy and efficiency are not the only crucial factors for developing medical deep learning models; the authenticity of results and the accountability of the model and its creator also matter greatly. The novelty of this paper lies in its specific application of extended preprocessing techniques-namely, the removal of background and irrelevant parts-to medical images for improving the performance of deep learning models in classification tasks. While the concept of preprocessing images has been explored by many researchers, applying such targeted preprocessing steps to medical images, combined with the use of XAI to validate and illustrate the benefits, is a novel approach. This paper highlights the unique requirements of medical image data and proposes an innovative method to enhance model accuracy and reliability in medical diagnostics by removing background and redundant features from the images.

Trajectory prediction of a moving object has imperative significance in both research and practical applications, ranging from target tracking, security surveillance and autonomous vehicle driving. For improving the efficacy of such prediction, a novel approach of data stream learning coupled with Kalman-filter and evolving correlated horizons feature selection (KF-ECH-FS) is proposed. KF has traditionally been used as a control-feedback-loop mechanism that corrects the errors from past trials, to predict the next-step position in trajectory prediction. In our fusion model here, KF and its windowed version are being used as predictor variables in a multivariate time series forecasting process. The predictor variables which serve as additional features aid in improving the trajectory prediction when only the relevant features are being selected in the incremental learning manner by multi-variate data stream analytics. Our proposed ECH-FS solves the problem of model overfitting when many features after expansion by time-series windowing are evaluated and selected along the learning process. A simple and efficient feature selection heuristics, Auto-encoder is used, along with data stream learning by Gate Recurrent Unit. The results, through an experimentation over a sample case of camera surveillance of accident prevention, show that our proposed KF-ECH-FS is superior to either KF or windowing alone in 1-step horizon trajectory prediction.

Environmental perception provides the necessary information for unmanned ground vehicles to recognize and interact with surrounding objects. Velodyne light detection and ranging (LiDAR) is widely used for this purpose due to its significant advantages such as high precision and being uninfluenced by varying illuminations. However, the unstructured distribution of LiDAR point clouds always affects the performance of feature extraction and object recognition. Moreover, the numbers of parameters in most deep learning models of object recognition are very large and the training process costs lots of computation consumption. This paper proposes a broad learning system (BLS) variant with a unified space autoencoder (USAE) as a lightweight model to recognize 3D objects. When the proposed method was evaluated on the LiDAR point cloud dataset and ModelNet10 dataset, the experimental results indicated that the recognition accuracy of our USAE-BLS model was similar to that of state-of-the-art 3D object recognition models. Moreover, the USAE-BLS has a much smaller model size and shorter training time than that of the deep learning models.

Analysis of irregularities in Covid-19 data could open a new window to learn more about the unprecedented problems of the current global pandemic. Of many, radiographs and clinical records are reliable sources for viral infection investigation and treatment planning. Clinical records help track the Covid-19 pandemic. In this paper, we present a Spike Neural Network (SNN) with supervised synaptic learning to detect abnormalities in Chest X-rays (CXRs) In other words, the proposed SNN can distinguish Covid-19 positive cases from healthy ones. In our decision-making procedure, we introduce clinical practice so Explainable AI (XAI) is possible to carry out. In addition, Support Vector Machine (SVM) with local interpretable model-agnostic explanation (LIME) provides reliable analysis of abnormalities in Covid-19 clinical data.

Internet of Things has been one of the catalysts in revolutionizing conventional healthcare services. With the growing society, traditional healthcare systems reach their capacity in providing sufficient and high-quality services. The world is facing the aging population and the inherent need for assisted-living environments for senior citizens. There is also a commitment by national healthcare organizations to increase support for personalized, integrated care to prevent and manage chronic conditions. Many applications related to In-Home Health Monitoring have been introduced over the last few decades, thanks to the advances in mobile and Internet of Things technologies and services. Such advances include improvements in optimized network architecture, indoor networks coverage, increased device reliability and performance, ultra-low device cost, low device power consumption, and improved device and network security and privacy. Current studies of in-home health monitoring systems presented many benefits including improved safety, quality of life and reduction in hospitalization and cost. However, many challenges of such a paradigm shift still exist, that need to be addressed to support scale-up and wide uptake of such systems, including technology acceptance and adoption by patients, healthcare providers and policymakers. The aim of this paper is three folds: First, review of key factors that drove the adoption and growth of the IoT-based in-home remote monitoring; Second, present the latest advances of IoT based in-home remote monitoring system architecture and key building blocks; Third, discuss future outlook and our recommendations of the in-home remote monitoring applications going forward.

Under the pressure of the growing millennial population and senior citizens are aging, which is one of the top societal priorities in many countries, the provision of healthcare needs to evolve and improve. A roadmap paved by World Health Organization (WHO) in March 2019 called Global Strategy on Digital Health 2020-2024, specified a grand vision of promoting healthy lives and well-beings for everyone, everywhere, at all ages[1]. WHO urges all nations to work hand in hand in developing and delivering Digital Health initiatives supported by robust government strategies that amalgamate financial, organizational, human and technological resources[2]. In particular, there are some niches areas in the strategies such as the adoption of distributed sensors and assisted living emerging in recent years. To this end, a lot of efforts both from the research community and industrial providers are anticipated to put forth in the coming decade, in implementing the concept of assisted living using hardware devices into meaningful solutions for fulfilling the growing needs of assisted living.

Abstract In this paper, we propose a new method for improving human activity recognition (HAR) datasets in order to increase their classification accuracy when trained with a certain classifier like a Neural Network. In this paper a novel training/testing process for building/testing a classification model for human activity recognition (HAR) is proposed. Traditionally, HAR is done by a classifier that learns what activities a person is doing by training with skeletal data obtained from a motion sensor such as Microsoft Kinect or accelerometer sensors. These skeletal data are the spatial coordinates (x, y, z) of different parts of the human body. In addition to the spatial features that describe current positions in the skeletal data, new features called Shadow Features are used to improve the supervised learning efficiency and accuracy of Neural Network classifiers. Shadow Features are inferred from the dynamics of body movements, thereby modelling the underlying momentum of the performed activities. They provide extra dimensions of information for characterizing activities in the classification process and thus significantly improving the accuracy. These Shadow Features are generated based on the existing features obtained from sensor datasets. In this paper we show that the accuracy of a neural network classifier can be significantly improved by the addition of Shadow Features and we also show that the generation of Shadow Features can be achieved with little time cost, on the fly, with the NVIDIA GPU technology and the CUDA programming model, this way we can improve the Neural Network accuracy at almost no time cost. GPUs are particularly suitable for generating Shadow Features, since they possess multiple cores which can be taken advantage of, in order to generate Shadow Features for multiple data columns in parallel, therefore reducing a lot of processing time, especially when dealing with huge datasets.

Numerous supply-chain combines with internet of things (IoT) applications have been proposed, and many methods and algorithms enhance the convenience of supply chains. However, new businesses still find it challenging to enter a supply chain, because unauthorised IoT devices of different companies illegally access resources. As security is paramount in a supply chain, IoT management has become very difficult. Public resources allocation and waste management also pose a problem. To solve the above problems, we proposed a new IoT management framework that embraces blockchain technology to help companies to form a supply chain effectively. This framework consists of an access control system, a backup peer mechanism and an internal data isolation and transmission approach. The access control system has a registrar module and an inspection module. The registrar module is mainly responsible for information registration with a registration policy, which has to be followed by all the companies in the supply chain. Besides, it provides a revocation and updating function. The inspection module focuses on judging misbehaviour and monitors the actions of the subjects; when any misoperation occurs, the system will correspondingly penalise violators. So that all related actions and information are verified and stored into blockchain, the IoT access control and safety of IoT admission are enhanced. Furthermore, in a blockchain system, if one single peer in the network breaks down, then the whole system may stop, because consensus cannot be reached. The data of the broken peer may be lost if it does not commit yet. The backup peer mechanism allows the primary peer and the backup peer to connect to an inspecting server for acquiring real-time data. The internal data isolation and transmission modules transmit and stores private data without creating a new subchannel. The proposed method is taken full account of the stability of the network and the fault tolerance to guarantee the robust of the system. To obtain unbiases results, experiments are conducted in two different blockchain environment. The results show our proposed method are promising IoT blockchain system for the supply chain.

Every healthcare unit is experiencing tremendous pressure to improve its practice quality across several dimensions. Multiple bodies of literature support the importance of establishing community of practice (CoP) to enrich the professional practice and add the expert context on the patient cases. The CoP emphasizes the importance of qualitative social learning and connectivity as preferred sources of knowledge updates to guide the practice rather than using the mere direct quantitative evidence. Social learning and connectivity in CoP is a complex sociotechnical process that takes an abstract idea through a cycle of participation and reification to derive more thickened context and refined knowledge that will help largely the accuracy of decision making. This process is not a straightforward one requiring the use of suitable hyper structure for representing the contextual evolving knowledge as well as a flexible infrastructure to enable CoP learning from experts, agents and connected services as well as other sources of data and knowledge. This article focuses on using the notion of workflow as the hyper structure and Node-RED as the platform that can facilitate CoP learning and connectivity. The focus is on using the CoP Node-RED workflows in healthcare settings to provide basic collaboration and connectivity as well as extensions to facilitate higher participation, learning, and connectivity to arrive at reification of the practice experience. With Node-RED workflows ideas can be represented as flows and sub flows where it can be shared with other CoP members as JSON hyper structure for further improvement, analytics and decision making.

The following topics are dealt with: learning (artificial intelligence); feature extraction; medical signal processing; convolutional neural nets; deep learning (artificial intelligence); neural nets; neurophysiology; electroencephalography; mobile robots; brain-computer interfaces.

Social media conveys a reachable platform for users to share information. The inescapable practice of social media has produced remarkable volumes of social data. Social media gathers the data in both structured-unstructured and formal-informal ways as users are not concerned with the exact grammatical structure and spelling when interacting with each other by means of various social networking websites (Twitter, Facebook, YouTube, LinkedIn, etc.). People are increasingly involved in and dependent on social media networks for data, news and opinions of other handlers on a variety of topics. The strong dependence on social media network sites contributes to enormous data generation characterized by three issues: scale, noise, and variety. Such problems also hinder social network data to be evaluated manually, resulting in the correct use of statistical analytical methods. Mining social media data can extract significant patterns that can be advantageous for consumers, users, and business. Pattern mining offers a wide variety of methods to detect valuable knowledge from huge datasets, such as patterns, trends, and rules. In this work, data was collected comprised of users’ opinions and sentiments and then processed using a significant number of pattern mining methods. The results were then further analyzed to attain meaningful information. The aim of this paper is to deliver a summary and a set of strategies for utilizing the ubiquitous pattern mining approaches, and to recognize the challenges and future research guidelines of dealing out social media data.

Building energy demand prediction (BEDP) concerns sensing the environment using the Internet of Things (IoT), making seamless decisions and responding and controlling certain devices automatically, intelligently, and quickly. Typically, the BEDP application can be empowered by fog computing where the sensed data are processed at the edge nodes rather than in a central cloud. The challenge is that in this decentralized IoT environment, the machine learning algorithm implemented at the fog node must learn a model from the incoming data accurately and fast. Which type of incremental learning algorithms, combined with traditional or swarm types of stochastic feature selection methods, are more suitable for BEDP? In this article, this topic is investigated in detail by introducing a new incremental learning model, the swarm decision table (SDT) in comparison with the classical decision tree. The simulation experiments using an empirical energy consumption data set that represent a typical IoT-connected BEDP scenario are tested, and the SDT shows superior results in terms of accuracy and time, demonstrating it as a suitable machine learning candidate in a fog computing environment.

In this paper, we proposed an approach to clustering based on bio-inspired behaviour and distributed energy efficient model. The motivation to propose this clustering approach is due to the challenge of performance in terms of finding an efficient way to send data packets to base stations and to maintain the lifetime performance of wireless sensor networks. The bioinspired approach adopted the behaviour of a bird called Kestrel. This behaviour is expressed using mathematical formulation and then translated into an algorithm. The bio-inspired algorithm is combined with the distributed energy efficient model for clustering to ensure efficient energy optimization. The proposed clustering approach, referred to as DEEC-KSA, is evaluated through simulation and compared with benchmarked clustering algorithms. The result of simulation showed that the performance of DEEC-KSA is efficient among the comparative clustering algorithms for energy optimization in terms of stability period, network lifetime and network throughput. Additionally, the proposed DEEC-KSA has the optimal time (in seconds) to send packets to base station successfully.

Abstract Tourist reviews on social media websites reflect the tourist's opinions concerning various aspects of a tourist place or service (e.g., “comfortable room” and “terrible service” in hotel reviews). Extracting these aspects from reviews is a challenging task in opinion mining. Therefore, aspect‐based opinion mining has emerged as a new area of social review mining. Existing approaches in this area focus on extracting explicit aspects and classification of opinions around these aspects. However, the implicit and coreferential aspects during aspect extraction are often neglected, and the classification of multiaspect opinions is relatively less emphasized in prior art. In this paper, we propose a model, namely, “enhanced multiaspect‐based opinion classification” that addresses existing challenges by automatically extracting both explicit and implicit aspects and classifying the multiaspect opinions. In this model, first, a probabilistic co‐occurrence‐based method is proposed that utilizes the co‐occurrence between aspects and sentiment words to identify the coreferential aspects and merge them into groups. Second, an implicit aspect extraction method is proposed that associates the sentiment words with suitable aspects to build an aspect‐sentiment hierarchy. Third, a multiaspect opinion classification approach is proposed that employs multilabel classification algorithms to classify opinions into different polarity classes. The effectiveness of the proposed model is evaluated by conducting experiments on benchmark and real‐world datasets. The experimental results revealed the supremacy of multilabel classifiers by achieving 90% accuracy per label on classification when extracting 87% domain‐relevant aspects. A state‐of‐the‐art performance comparison is conducted that also verifies the advantages of the proposed model.

Although deep learning methods have been applied to the selection of features in the classification problem, current methods of learning parameters to be used in the classification approach can vary in terms of accuracy at each time interval, resulting in potentially inaccurate classification.To address this challenge, this study proposes an approach to learning these parameters by using two different aspects of Kestrel bird behavior to adjust the learning rate until the optimal value of the parameter is found: random encircling from a hovering position and learning through imitation from the well-adapted behaviour of other Kestrels.Additionally, deep learning method (that is, recurrent neural network with long short term memory network) was applied to select features and the accuracy of classification.A benchmark dataset (with continuous data attributes) was chosen to test the proposed search algorithm.The results showed that KSA is comparable to BAT, ACO and PSO as the test statistics (that is, Wilcoxon signed rank test) show no statistically significant differences between the mean of classification accuracy at level of significance of 0.05.However, KSA, when compared with WSA-MP, shows a statistically significant difference between the mean of classification accuracy.

This concept paper reviews relevant literature in order first to identify the most common features needed for an effective health care Real Time Location System (RTLS) and secondly to evaluate common RTLS technologies against this identified feature criteria in order to determine the most feasible technology. The most feasible set of technologies arising from this evaluation is considered for further research and development RTLS in healthcare.

Apriori Association Rule Mining (ARM) is a popular data mining technique for deriving association rules from frequent itemsets, and it has a long history. Despite of its popularity, its performance suffers from a bottleneck in scalability. Many attempts were made in the past, including changing the frequent item database structure to sophisticated parallel execution. In this paper an alternative strategy is proposed which centred on segmenting the database in lieu of using the full database. The segmentation is by ensemble method which sifts and selects the most effective clustering algorithm; the resultant segmented data cluster is subsequently used for ARM. Using only a fraction of data transactions which supposedly has a high concentration of expressive data, ARM produces higher quality association rules at shorter time. The proposed ARM model is tested using three cases - bank, homicide and lung cancer. The results confirm the usefulness of this new model - higher quality rules are gained

The objective of this study is to solve the problem of user data not being precisely received from sensors because of sensing region limitations in invoked reality (IR) space, distortion of colors or patterns by lighting, and blocking or overlapping of a user by other users. The sensing scope range is thus expanded using multiple sensors in the IR space. Moreover, user feature data are accurately identified by user sensing. Specifically, multiple sensors are employed when not all of user data are sensed because they overlap with data of other users. In the proposed approach, all clients share the user feature data from multiple sensors. Accordingly, each client recognizes that the user is the same individual on the basis of the shared data. Furthermore, the identification accuracy is improved by identifying the user features based on colors and patterns that are less affected by lighting. Therefore, accurate identification of the user feature data is enabled, even under lighting changes. The proposed system was implemented based on system performance analysis standards. The practicality and system performance in identifying the same person using the proposed method were verified through an experiment.

Due to the popularity of web services on the Internet, it is important to have a clear view of their utilization behaviors. Despite asynchronous service invocations and distributed executions can provide better user experience, our views are blurred by out-of-order and fragmented service logs. Researchers have been trying various methods to reveal emerging service orchestration patterns, but nearly all of them have taken deterministic approaches. Hence, they do not natively cater for incomplete data and noises. In this paper, we propose to address these problems by using topic models aiming to reveal service orchestration patterns from sparse service logs. Probabilistic approaches do not only tolerate data defects, but their associated approximation methods also overcome combinatorial explosion. We first investigate the implications of sparsity on topic models. Secondly, we propose an extended time-series form of susceptible-infectious-recovered model to monitor the dynamics of emerging service orchestrations. We quantify their emerging-potential by estimated effective-reproduction-number, which is obtained incrementally by Bayesian parameter estimations. Guided by our proposed emerging-potential measure, one can profile and categorize emerging service orchestration patterns, and generate automated alerts on upcoming consumption peaks. In practice, our model enables service providers to better allocate their resources to meet demands dynamically. While our findings affirm that biterm topic model can be applied to service logs with short and sparse log entries, the effectiveness of our proposed monitoring solutions is also shown by experiments.

Big Data though it is a hype up-springing many technical challenges that confront both academic research communities and commercial IT deployment, the root sources of Big Data are founded on data streams and the curse of dimensionality. It is generally known that data which are sourced from data streams accumulate continuously making traditional batch-based model induction algorithms infeasible for real-time data mining. Feature selection has been popularly used to lighten the processing load in inducing a data mining model. However, when it comes to mining over high dimensional data the search space from which an optimal feature subset is derived grows exponentially in size, leading to an intractable demand in computation. In order to tackle this problem which is mainly based on the high-dimensionality and streaming format of data feeds in Big Data, a novel lightweight feature selection is proposed. The feature selection is designed particularly for mining streaming data on the fly, by using accelerated particle swarm optimization (APSO) type of swarm search that achieves enhanced analytical accuracy within reasonable processing time. In this paper, a collection of Big Data with exceptionally large degree of dimensionality are put under test of our new feature selection algorithm for performance evaluation.

In software engineering, the concept of code reuse is very common. Code reuse is the concept of copying and pasting the code in multiple places in the same software or different software without modification. This practice may reduce software maintainability and give rise to serious maintenance problems. In the last few decades numerous code clone detection techniques and tools have been proposed for capturing duplicated redundant code. Each of these techniques attempts to find out the duplicated code, which is also known as software clone. These techniques include Kclone, CP-Miner, CC-Finder, CReN etc. The objective of those researches is the exploration of various clone detection techniques and tools proposed so far. In this study, we propose an efficient clone detection technique which is used to detect clones in various programming languages. We have endeavored to improve performance and overcome the key problem of detecting clones in only one language. The proposed technique has been evaluated using two-dimensional array which has exhibited a faster method of storing and identification of clones in source code. We are also working on some of its future directions including the removal of the clones detected from the source code.

Departmental workflows within a digital business ecosystem are often executed concurrently and required to share limited number of resources. However, unexpected events from the business environment and delay in activities can cause temporal exceptions in these workflows. Predicting temporal exceptions is a complex task since a workflow can be implemented with various types of control flow patterns. In this paper, we describe a critical path based approach for predicting temporal exceptions in concurrent workflows which are required to share limited resources. Our approach allows predicting temporal exceptions in multiple attempts while workflows are being executed.

Infectious diseases that are caused by pathogenic microorganisms can spread fast and far, from one person to another, directly or indirectly. Prompt quarantining of the infected from the rest, coupled with contact tracing, has been an effective measure to encounter outbreaks. However, urban life and international travel make containment difficult. Furthermore, the length of incubation periods of some contagious diseases like SARS enable infected passengers to elude health screenings before first symptoms appear and thus to carry the disease further. Detecting and visualizing contact–tracing networks, and immediately identifying the routes of infection, are thus important. We apply information visualization and hidden cluster detection for finding cliques of potentially infected people during incubation. Preemptive control and early quarantine are hence possible by our method. Our prototype Infectious Disease Detection and Quarantine Management System (IDDQMS), which can identify and trace clusters of infection by mining patients’ history, is introduced in this paper.

Online recommenders are usually referred to those used in e-Commerce websites for suggesting a product or service out of many choices. The core technology implemented behind this type of recommenders includes content analysis, collaborative filtering and some hybrid variants. Since they all have certain strengths and limitations, combining them may be a promising solution provided there is a way of overcoming a large amount of input variables especially from combining different techniques. Genetic algorithm (GA) is an ideal optimization search function, for finding a best recommendation out of a large population of variables. In this paper we presented a GA-based approach for supporting combined modes of collaborative filtering. In particular, we show that how the input variables can be coded into GA chromosomes in various modes. Insights of how GA can be used in recommenders are derived through our experiments with the input data taken from Movielens and IMDB.

A double agents-based model in a make-to-order supply chain (MTOSC), called collaborative single machine earliest/tardiness model (CSET model) is proposed. It integrates Pareto-optimal method with just-in-time principle (JIT) into two intelligent agents for optimizing dynamic supply chain formation and scheduling respectively. This model mainly focuses on improving the sequence timing factor. JIT is able to shorten the waiting time while Pareto-optimality provides a mechanism that each participant won't suffer loss. Combining the two methods, an unprecedented efficiency on supply chain (SC) can be achieved. We have compared how our proposed method performed in various MTOSC formations. The experimental results indicate that CSET model yields time improvement on the SC workflows.

World-Wide-Web is a huge pool of valuable information for companies to know what their competitors are doing and what products and services they offer up-to-date. Companies can gather business intelligence from the Web for planning countermeasures strategies. Hence it is crucial to have the right tool to effectively gather such information from the Web. Many information retrieval and monitoring technologies have been developed. But they are more for generally tracking changes and downloading the whole websites for offline browsing. This paper is to shed some light on specifically the design of a Web monitoring system for gathering business information relevant to a company. The Watcher Agent is a server-based system that is built with two main parts, namely Price Watcher and Market Watcher. The system will assist company users in price information collection, news information filtering, and product ranking estimation, thus saving time and effort for them.

Supply chains have evolved to web-applications that tap on the power of internet to expand their networks online. Recently some research attention is focused on make-to-order supply chain formation where orders are scheduled to be optimally distributed among online manufacturers and suppliers for mutual benefits. A SET model was proposed in [1] using Pareto theories. The model is then extended into a collaborative manner throughout the whole supply chain by incorporating it with the Just-in-Time (JIT) principle, known as CSET. The CSET framework was proposed, and the advantage of time efficiency was shown in [2]. The core of the CSET model is based on intelligent agent technology. Specifically the model is supported by double-agent architecture with each type of agents who makes provisional plans of order distribution by Pareto optimality and JIT coordination respectively. This paper defines such double-agent mechanisms in details, as well as demonstrating its merits via simulation study.

In mobile commerce, short-message-service (SMS) is an important technique for delivering payment instruction. A payment model "SMS Credit" was proposed earlier [1]. Such payment service or similar relies on the transmission of SMS; it is needed to reduce the occurrence of packet losses and delay, to improve the quality of packet transmission services (QOS) in the network. This paper discusses how the payment service operates in a configurable radio resource environment via data mining. A Radio Resource Management and Prediction Server equipped with data mining algorithms will optimize the radio resources for both voice and data services in order to provide an optimized QOS. Specifically, data mining techniques are applied to define traffic policy and to calculate optimization result through traffic profile analysis.

Recent studies show that the number of mobile devices has already exceeded the number of personal computers. Contradictorily, the commerce conducted via the mobile devices (ldquoM-Commercerdquo) is far less than that of internet (ldquoE-Commercerdquo). This paper looks into the context of the mobile commerce from the dimensions of application scenarios and user acceptance. Spinning out from these discussions, it helps to provide a better understanding of how the mobile payment model is constructed and what m-commerce is all about. From the userpsilas perspective, the user acceptance analysis of the mobile users and the mobile commerce system the constraints of the system are discussed. With these two dimensions in mind, the application scenarios of the m-commerce system, that is, the situations where m-commerce can put forth to perform required task, are suggested. Our empirical results show that although people in Macau think using mobile phone is an innovative and convenient tool for making payment, the acceptance level is still not high.

In the view of escalating global threat in security, it is imperative to have an automated detection system that can pick up suspicious patterns of human movement in physical environments. It can give a forewarning before a planned attack happens or an ultimate security is breached. In the past, significant research on the intrusion detection was established, but limited to virtual environments like computer networks and operating systems. In this paper, we proposed a general security model for detecting suspicious patterns in physical environment. Suspicious patterns are subtle and we showed that they can be detected via an experiment.

Knowledge incorporation is one challenge in e-Commerce automated negotiation. In this paper, we describe a model of B2B negotiation using knowledge. We classify the types of knowledge namely general knowledge and negotiation knowledge, in the negotiation process. A methodology that uses Knowledge Bead (KB) and meta-KB as knowledge representation that would be suitable for the design of automated negotiation systems is discussed. An experimental prototype demonstrates that by incorporating knowledge into automated negotiation yields improved results.

The technology of using computer Agents for B2B trading in Cyberworld is getting prevalent nowadays. Much of the research work has been done on the models, architectures and service provision in the past. However, agent negotiation remains as a challenge in making the whole trading process fully automated, due to its fuzzy and complex nature. The lack of interoperability and knowledge-reuse that limits to case-based, pose certain drawback. An object-oriented ontology-based Knowledge Bead (KB) method for knowledge representation has been proposed (S. Fong et al., 2002). It was designed as a foundation to enable agent negotiation in e-trading environment in a systematic way. This paper continues the research work of KB on formulating its theorems and methodology. Some typical negotiation paradigms using appropriate strategies based on the KB's methodology are presented as well. In particular, KB's taxonomies and their use in the whole negotiation process are discussed. The main advantage of this approach is the ability to describe the deal that is under negotiation in an object-oriented format which in turn allows reasoning, optimizing, knowledge reuse and management.

Online negotiation for today's B2B e-commerce plays a promising role in assisting traders to best fulfill their business deals. But existing uncertain constraints and sophisticated strategies presented in e-trading environment make negotiation a rather complex process. As business intelligence and efficient protocols are essential to successful negotiation, we defined an object-oriented ontology knowledge bead (KB) for knowledge representation and enabling agent negotiation in e-trading environment [S. Fong et al. (2002)]. We continue the research work of KB on formulating its theorems and methodology. Some typical negotiation paradigms using appropriate strategies based on the KB's methodology are presented. In particular, KB's taxonomies and their use in the negotiation process are discussed.

This paper discusses about how business intelligence on a website could be obtained from users' access records instead of web logs of "hits". Users' access records are captured by implementing an Access-Control (AC) architectural model on the website. This model requires users to register their profiles in an exchange of a password; and thereafter they have to login before gaining access to certain resources on the website. The links to the resources on the website have been modified such that a record of information about the access would be recorded in the database when clicked. This way, datamining can be performed on a relatively clean set of access records about the users. Hence, a good deal of business intelligence about the users' behaviors, preferences and about the popularities of the resources (products) on the website can be gained. In this paper, we also discussed how the business intelligence acquired, in turn, can be used to provide e-CRM for the users.

The challenge in building an ATM switch, especially in the area of QoS control, lies in the design of the port controllers. By combining port reconfiguration, cell scheduling and connection admission control, a programmable QoS controller at the output port of the non-blocking, output-buffered reconfigurable ATM switch is proposed. In such a switch, the output port is the only point of contention that can cause possible QoS degradations due to excessive queueing delays. Cell scheduling is used as an essential mechanism for attaining meaningful differentiated QoS behaviour. The QoS controller combines per-connection buffer management with a table-driven cell scheduler to achieve a broad range of QoS behaviour. This paper reports the simulation results of our study on (i) the effects of different traffic characteristics, and (ii) the effects of slot algorithms in the cell scheduler, on the QoS performance.

The framework of the performance analysis for a buffered asynchronous transfer mode (ATM) switch usually consists of modeling the input traffic arrivals, the switching mechanism, and the cell departure process. The overall accuracy of the performance results relies on how accurately the cell departure process, especially for shared buffer switches is modelled. Unlike output buffer switches where there are at most one cell that can leave, multiple cells may depart from the shared buffer for shared buffer switches. Modeling the cell departure process is hence more complex for shared buffer switches. It is of practical interest and is challenging to find the appropriate probabilistic model to describe the cell departure process for shared buffer switches. This paper compares and verifies the accuracy of three models, including a new one called "Urn Model" proposed by the authors. These models are put under test in a performance evaluation of a shared buffer ATM switch, by using a discrete-time Markov chain. The numerical results are compared to the simulation, and they show that the Urn Model is a good compromise between accuracy and efficiency. This finding is significant because it helps to speed up running an analytical model of a large network while providing satisfactory accuracy.