Journal of Big Data最新文献

Background

Models for structuring big-data and data-analytics projects typically start with a definition of the project’s goals and the business value they are expected to create. The literature identifies proper project definition as crucial for a project’s success, and also recognizes that the translation of business objectives into data-analytic problems is a difficult task. Unfortunately, common project structures, such as CRISP-DM, provide little guidance for this crucial stage when compared to subsequent project stages such as data preparation and modeling.

Contribution

This paper contributes structure to the project-definition stage of data-analytic projects by proposing the Data-Analytic Problem Structure (DAPS). The diagrammatic technique facilitates the collaborative development of a consistent and precise definition of a data-analytic problem, and the articulation of how it contributes to the organization’s goals. In addition, the technique helps to identify important assumptions, and to break down large ambitions in manageable subprojects.

Methods

The semi-formal specification technique took other models for problem structuring — common in fields such as operations research and business analytics — as a point of departure. The proposed technique was applied in 47 real data-analytic projects and refined based on the results, following a design-science approach.

Threats on computer networks have been increasing rapidly, and irresponsible parties are always trying to exploit vulnerabilities in the network to do various dangerous things. One way to exploit vulnerabilities in a computer network is by employing malware. Botnets are a type of malware that infects and attacks targets in groups. Botnets develop quickly; the characteristics of initially sporadic attacks have grown into periodic and simultaneous. This rapid development has proved that the botnet is advanced and requires more attention and proper handling. Many studies have introduced detection models for botnet attack activity on computer networks. Apart from detecting the presence of botnet attacks, those studies have attempted to explore the characteristics of botnets, such as attack intensity, relationships between activities, and time segment analysis. However, there has been no research that explicitly detects those characteristics. On the other hand, each botnet characteristic requires different handling, while recognizing the characteristics of the botnet can help network administrators make appropriate decisions. Based on these reasons, this research builds a detection model that can recognize botnet characteristics using sequential traffic mining and similarity analysis. The proposed method consists of two main processes. The first is training to build a knowledge base, and the second is testing to detect botnet activity and attack characteristics. It involves dynamic thresholds to improve the model sensitivity in recognizing attack characteristics through similarity analysis. The novelty includes developing and combining analytical techniques of sequential traffic mining, similarity analysis, and dynamic threshold to detect and recognize the characteristics of botnet attacks explicitly on actual behavior in network traffic. Extensive experiments have been conducted for the evaluation using three different datasets whose results show better performance than others.

Football is the most popular game in the world and has significant influence on various aspects including politics, economy and culture. The experience of the football developed nation has shown that the steady growth of youth football is crucial for elevating a nation's overall football proficiency. It is essential to develop techniques and create strategies that adapt to their individual physical features to resolve the football players’ problem of lacking exercise in various topics. In this manuscript, Computer aided technology depending on the Graph Sample and Aggregate Attention Network Optimized for Soccer Teaching and Training (CAT-GSAAN-STT) is proposed to improve the efficiency of Soccer teaching and training effectively. The proposed method contains four stages, like data collection, data preprocessing, prediction and optimization. Initially the input data are collected by Microsoft Kinect V2 smart camera. Then the collected data are preprocessed by using Improving graph collaborative filtering. After preprocessing the data is given for motion recognition layer here prediction is done using Graph Sample and Aggregate Attention Network (GSAAN) for improving the effectiveness of Soccer Teaching and Training. To enhance the accuracy of the system, the GSAAN are optimized by using Artificial Rabbits Optimization. The proposed CAT-GSAAN-STT method is executed in Python and the efficiency of the proposed technique is examined with different metrics, like accuracy, computation time, learning activity analysis, student performance ratio and teaching evaluation analysis. The simulation outcomes proves that the proposed technique attains provides28.33%, 31.60%, 25.63% higherRecognition accuracy and33.67%, 38.12% and 27.34%lesser evaluation time while compared with existing techniques like computer aided teaching system based upon artificial intelligence in football teaching with training (STT-IOT-CATS), Computer Aided Teaching System for Football Teaching and Training Based on Video Image (CAT-STT-VI) and method for enhancing the football coaching quality using artificial intelligence and meta verse-empowered in mobile internet environment (SI-STQ-AI-MIE) respectively.

Efficiently treating cardiac patients before the onset of a heart attack relies on the precise prediction of heart disease. Identifying and detecting the risk factors for heart disease such as diabetes mellitus, Coronary Artery Disease (CAD), hyperlipidemia, hypertension, smoking, familial CAD history, obesity, and medications is critical for developing effective preventative and management measures. Although Electronic Health Records (EHRs) have emerged as valuable resources for identifying these risk factors, their unstructured format poses challenges for cardiologists in retrieving relevant information. This research proposed employing transfer learning techniques to automatically extract heart disease risk factors from EHRs. Leveraging transfer learning, a deep learning technique has demonstrated a significant performance in various clinical natural language processing (NLP) applications, particularly in heart disease risk prediction. This study explored the application of transformer-based language models, specifically utilizing pre-trained architectures like BERT (Bidirectional Encoder Representations from Transformers), RoBERTa, BioClinicalBERT, XLNet, and BioBERT for heart disease detection and extraction of related risk factors from clinical notes, using the i2b2 dataset. These transformer models are pre-trained on an extensive corpus of medical literature and clinical records to gain a deep understanding of contextualized language representations. Adapted models are then fine-tuned using annotated datasets specific to heart disease, such as the i2b2 dataset, enabling them to learn patterns and relationships within the domain. These models have demonstrated superior performance in extracting semantic information from EHRs, automating high-performance heart disease risk factor identification, and performing downstream NLP tasks within the clinical domain. This study proposed fine-tuned five widely used transformer-based models, namely BERT, RoBERTa, BioClinicalBERT, XLNet, and BioBERT, using the 2014 i2b2 clinical NLP challenge dataset. The fine-tuned models surpass conventional approaches in predicting the presence of heart disease risk factors with impressive accuracy. The RoBERTa model has achieved the highest performance, with micro F1-scores of 94.27%, while the BERT, BioClinicalBERT, XLNet, and BioBERT models have provided competitive performances with micro F1-scores of 93.73%, 94.03%, 93.97%, and 93.99%, respectively. Finally, a simple ensemble of the five transformer-based models has been proposed, which outperformed the most existing methods in heart disease risk fan, achieving a micro F1-Score of 94.26%. This study demonstrated the efficacy of transfer learning using transformer-based models in enhancing risk prediction and facilitating early intervention for heart disease prevention.

Abstract

RNA Sequencing (RNA-Seq) has been considered a revolutionary technique in gene profiling and quantification. It offers a comprehensive view of the transcriptome, making it a more expansive technique in comparison with micro-array. Genes that discriminate malignancy and normal can be deduced using quantitative gene expression. However, this data is a high-dimensional dense matrix; each sample has a dimension of more than 20,000 genes. Dealing with this data poses challenges. This paper proposes RBNRO-DE (Relief Binary NRO based on Differential Evolution) for handling the gene selection strategy on (rnaseqv2 illuminahiseq rnaseqv2 un edu Level 3 RSEM genes normalized) with more than 20,000 genes to pick the best informative genes and assess them through 22 cancer datasets. The k-nearest Neighbor (k-NN) and Support Vector Machine (SVM) are applied to assess the quality of the selected genes. Binary versions of the most common meta-heuristic algorithms have been compared with the proposed RBNRO-DE algorithm. In most of the 22 cancer datasets, the RBNRO-DE algorithm based on k-NN and SVM classifiers achieved optimal convergence and classification accuracy up to 100% integrated with a feature reduction size down to 98%, which is very evident when compared to its counterparts, according to Wilcoxon’s rank-sum test (5% significance level).

Bloom filters, since their introduction over 50 years ago, have become a pillar to handle membership queries in small space, with relevant application in Big Data Mining and Stream Processing. Further improvements have been recently proposed with the use of Machine Learning techniques: learned Bloom filters. Those latter make considerably more complicated the proper parameter setting of this multi-criteria data structure, in particular in regard to the choice of one of its key components (the classifier) and accounting for the classification complexity of the input dataset. Given this State of the Art, our contributions are as follows. (1) A novel methodology, supported by software, for designing, analyzing and implementing learned Bloom filters that account for their own multi-criteria nature, in particular concerning classifier type choice and data classification complexity. Extensive experiments show the validity of the proposed methodology and, being our software public, we offer a valid tool to the practitioners interested in using learned Bloom filters. (2) Further contributions to the advancement of the State of the Art that are of great practical relevance are the following: (a) the classifier inference time should not be taken as a proxy for the filter reject time; (b) of the many classifiers we have considered, only two offer good performance; this result is in agreement with and further strengthens early findings in the literature; (c) Sandwiched Bloom filter, which is already known as being one of the references of this area, is further shown here to have the remarkable property of robustness to data complexity and classifier performance variability.

In the context of high-dimensional credit card fraud data, researchers and practitioners commonly utilize feature selection techniques to enhance the performance of fraud detection models. This study presents a comparison in model performance using the most important features selected by SHAP (SHapley Additive exPlanations) values and the model’s built-in feature importance list. Both methods rank features and choose the most significant ones for model assessment. To evaluate the effectiveness of these feature selection techniques, classification models are built using five classifiers: XGBoost, Decision Tree, CatBoost, Extremely Randomized Trees, and Random Forest. The Area under the Precision-Recall Curve (AUPRC) serves as the evaluation metric. All experiments are executed on the Kaggle Credit Card Fraud Detection Dataset. The experimental outcomes and statistical tests indicate that feature selection methods based on importance values outperform those based on SHAP values across classifiers and various feature subset sizes. For models trained on larger datasets, it is recommended to use the model’s built-in feature importance list as the primary feature selection method over SHAP. This suggestion is based on the rationale that computing SHAP feature importance is a distinct activity, while models naturally provide built-in feature importance as part of the training process, requiring no additional effort. Consequently, opting for the model’s built-in feature importance list can offer a more efficient and practical approach for larger datasets and more intricate models.

Abstract

Modern deep learning training procedures rely on model regularization techniques such as data augmentation methods, which generate training samples that increase the diversity of data and richness of label information. A popular recent method, mixup, uses convex combinations of pairs of original samples to generate new samples. However, as we show in our experiments, mixup can produce undesirable synthetic samples, where the data is sampled off the manifold and can contain incorrect labels. We propose (zeta ) -mixup, a generalization of mixup with provably and demonstrably desirable properties that allows convex combinations of ({T} ge 2) samples, leading to more realistic and diverse outputs that incorporate information from ({T}) original samples by using a p-series interpolant. We show that, compared to mixup, (zeta ) -mixup better preserves the intrinsic dimensionality of the original datasets, which is a desirable property for training generalizable models. Furthermore, we show that our implementation of (zeta ) -mixup is faster than mixup, and extensive evaluation on controlled synthetic and 26 diverse real-world natural and medical image classification datasets shows that (zeta ) -mixup outperforms mixup, CutMix, and traditional data augmentation techniques. The code will be released at https://github.com/kakumarabhishek/zeta-mixup.

Streaming adaptations of manifold learning based dimensionality reduction methods, such as Isomap, are based on the assumption that a small initial batch of observations is enough for exact learning of the manifold, while remaining streaming data instances can be cheaply mapped to this manifold. However, there are no theoretical results to show that this core assumption is valid. Moreover, such methods typically assume that the underlying data distribution is stationary and are not equipped to detect, or handle, sudden changes or gradual drifts in the distribution that may occur when the data is streaming. We present theoretical results to show that the quality of a manifold asymptotically converges as the size of data increases. We then show that a Gaussian Process Regression (GPR) model, that uses a manifold-specific kernel function and is trained on an initial batch of sufficient size, can closely approximate the state-of-art streaming Isomap algorithms, and the predictive variance obtained from the GPR prediction can be employed as an effective detector of changes in the underlying data distribution. Results on several synthetic and real data sets show that the resulting algorithm can effectively learn lower dimensional representation of high dimensional data in a streaming setting, while identifying shifts in the generative distribution. For instance, key findings on a Gas sensor array data set show that our method can detect changes in the underlying data stream, triggered due to real-world factors, such as introduction of a new gas in the system, while efficiently mapping data on a low-dimensional manifold.

Online platforms and social networking have increased in the contemporary years. They are now a major news source worldwide, leading to the online proliferation of Fake News (FNs). These FNs are alarming because they fundamentally reshape public opinion, which may cause customers to leave these online platforms, threatening the reputations of several organizations and industries. This rapid dissemination of FNs makes it imperative for automated systems to detect them, encouraging many researchers to propose various systems to classify news articles and detect FNs automatically. In this paper, a Fake News Detection (FND) methodology is presented based on an effective IBAVO-AO algorithm, which stands for hybridization of African Vultures Optimization (AVO) and Aquila Optimization (AO) algorithms, with an extreme gradient boosting Tree (Xgb-Tree) classifier. The suggested methodology involves three main phases: Initially, the unstructured FNs dataset is analyzed, and the essential features are extracted by tokenizing, encoding, and padding the input news words into a sequence of integers utilizing the GLOVE approach. Then, the extracted features are filtered using the effective Relief algorithm to select only the appropriate ones. Finally, the recovered features are used to classify the news items using the suggested IBAVO-AO algorithm based on the Xgb-Tree classifier. Hence, the suggested methodology is distinguished from prior models in that it performs automatic data pre-processing, optimization, and classification tasks. The proposed methodology is carried out on the ISOT-FNs dataset, containing more than 44 thousand multiple news articles divided into truthful and fake. We validated the proposed methodology’s reliability by examining numerous evaluation metrics involving accuracy, fitness values, the number of selected features, Kappa, Precision, Recall, F1-score, Specificity, Sensitivity, ROC_AUC, and MCC. Then, the proposed methodology is compared against the most common meta-heuristic optimization algorithms utilizing the ISOT-FNs. The experimental results reveal that the suggested methodology achieved optimal classification accuracy and F1-score and successfully categorized more than 92.5% of news articles compared to its peers. This study will assist researchers in expanding their understanding of meta-heuristic optimization algorithms applications for FND.

Graphical Abstract