Feature distillation and accumulated selection for automated fraudulent publisher classification from user click data of online advertising

PurposeThe problem of choosing the utmost useful features from hundreds of features from time-series user click data arises in online advertising toward fraudulent publisher's classification. Selecting feature subsets is a key issue in such classification tasks. Practically, the use of filter approaches is common; however, they neglect the correlations amid features. Conversely, wrapper approaches could not be applied due to their complexities. Moreover, in particular, existing feature selection methods could not handle such data, which is one of the major causes of instability of feature selection.Design/methodology/approachTo overcome such issues, a majority voting-based hybrid feature selection method, namely feature distillation and accumulated selection (FDAS), is proposed to investigate the optimal subset of relevant features for analyzing the publisher's fraudulent conduct. FDAS works in two phases: (1) feature distillation, where significant features from standard filter and wrapper feature selection methods are obtained using majority voting; (2) accumulated selection, where we enumerated an accumulated evaluation of relevant feature subset to search for an optimal feature subset using effective machine learning (ML) models.FindingsEmpirical results prove enhanced classification performance with proposed features in average precision, recall, f1-score and AUC in publisher identification and classification.Originality/valueThe FDAS is evaluated on FDMA2012 user-click data and nine other benchmark datasets to gauge its generalizing characteristics, first, considering original features, second, with relevant feature subsets selected by feature selection (FS) methods, third, with optimal feature subset obtained by the proposed approach. ANOVA significance test is conducted to demonstrate significant differences between independent features.

Techniques to detect terrorists/extremists on the dark web: a review

PurposeWith the proliferation of terrorist/extremist websites on the World Wide Web, it has become progressively more crucial to detect and analyze the content on these websites. Accordingly, the volume of previous research focused on identifying the techniques and activities of terrorist/extremist groups, as revealed by their sites on the so-called dark web, has also grown.Design/methodology/approachThis study presents a review of the techniques used to detect and process the content of terrorist/extremist sites on the dark web. Forty of the most relevant data sources were examined, and various techniques were identified among them.FindingsBased on this review, it was found that methods of feature selection and feature extraction can be used as topic modeling with content analysis and text clustering.Originality/valueAt the end of the review, present the current state-of-the- art and certain open issues associated with Arabic dark Web content analysis.

Artificial intelligence technologies for more flexible recommendation in uniforms

PurposeThis research aims to collect human body variables via 2D images captured by digital cameras. Based on those human variables, the forecast and recommendation of the Digital Camouflage Uniforms (DCU) for Taiwan's military personnel are made.Design/methodology/approachA total of 375 subjects are recruited (male: 253; female: 122). In this study, OpenPose converts the photographed 2D images into four body variables, which are compared with those of a tape measure and 3D scanning simultaneously. Then, the recommendation model of the DCU is built by the decision tree. Meanwhile, the Euclidean distance of each size of the DCU in the manufacturing specification is calculated as the best three recommendations.FindingsThe recommended size established by the decision tree is only 0.62 and 0.63. However, for the recommendation result of the best three options, the DCU Fitting Score can be as high as 0.8 or more. The results of OpenPose and 3D scanning have the highest correlation coefficient even though the method of measuring body size is different. This result confirms that OpenPose has significant measurement validity. That is, inexpensive equipment can be used to obtain reasonable results.Originality/valueIn general, the method proposed in this study is suitable for applications in e-commerce and the apparel industry in a long-distance, non-contact and non-pre-labeled manner when the world is facing Covid-19. In particular, it can reduce the measurement troubles of ordinary users when purchasing clothing online.

Dynamic Distributed and Parallel Machine Learning algorithms for big data mining processing

PurposeThis work can be used as a building block in other settings such as GPU, Map-Reduce, Spark or any other. Also, DDPML can be deployed on other distributed systems such as P2P networks, clusters, clouds computing or other technologies.Design/methodology/approachIn the age of Big Data, all companies want to benefit from large amounts of data. These data can help them understand their internal and external environment and anticipate associated phenomena, as the data turn into knowledge that can be used for prediction later. Thus, this knowledge becomes a great asset in companies' hands. This is precisely the objective of data mining. But with the production of a large amount of data and knowledge at a faster pace, the authors are now talking about Big Data mining. For this reason, the authors’ proposed works mainly aim at solving the problem of volume, veracity, validity and velocity when classifying Big Data using distributed and parallel processing techniques. So, the problem that the authors are raising in this work is how the authors can make machine learning algorithms work in a distributed and parallel way at the same time without losing the accuracy of classification results. To solve this problem, the authors propose a system called Dynamic Distributed and Parallel Machine Learning (DDPML) algorithms. To build it, the authors divided their work into two parts. In the first, the authors propose a distributed architecture that is controlled by Map-Reduce algorithm which in turn depends on random sampling technique. So, the distributed architecture that the authors designed is specially directed to handle big data processing that operates in a coherent and efficient manner with the sampling strategy proposed in this work. This architecture also helps the authors to actually verify the classification results obtained using the representative learning base (RLB). In the second part, the authors have extracted the representative learning base by sampling at two levels using the stratified random sampling method. This sampling method is also applied to extract the shared learning base (SLB) and the partial learning base for the first level (PLBL1) and the partial learning base for the second level (PLBL2). The experimental results show the efficiency of our solution that the authors provided without significant loss of the classification results. Thus, in practical terms, the system DDPML is generally dedicated to big data mining processing, and works effectively in distributed systems with a simple structure, such as client-server networks.FindingsThe authors got very satisfactory classification results.Originality/valueDDPML system is specially designed to smoothly handle big data mining classification.

A robust framework for shoulder implant X-ray image classification

PurposeMedical images are increasingly popular; therefore, the analysis of these images based on deep learning helps diagnose diseases become more and more essential and necessary. Recently, the shoulder implant X-ray image classification (SIXIC) dataset that includes X-ray images of implanted shoulder prostheses produced by four manufacturers was released. The implant's model detection helps to select the correct equipment and procedures in the upcoming surgery.Design/methodology/approachThis study proposes a robust model named X-Net to improve the predictability for shoulder implants X-ray image classification in the SIXIC dataset. The X-Net model utilizes the Squeeze and Excitation (SE) block integrated into Residual Network (ResNet) module. The SE module aims to weigh each feature map extracted from ResNet, which aids in improving the performance. The feature extraction process of X-Net model is performed by both modules: ResNet and SE modules. The final feature is obtained by incorporating the extracted features from the above steps, which brings more important characteristics of X-ray images in the input dataset. Next, X-Net uses this fine-grained feature to classify the input images into four classes (Cofield, Depuy, Zimmer and Tornier) in the SIXIC dataset.FindingsExperiments are conducted to show the proposed approach's effectiveness compared with other state-of-the-art methods for SIXIC. The experimental results indicate that the approach outperforms the various experimental methods in terms of several performance metrics. In addition, the proposed approach provides the new state of the art results in all performance metrics, such as accuracy, precision, recall, F1-score and area under the curve (AUC), for the experimental dataset.Originality/valueThe proposed method with high predictive performance can be used to assist in the treatment of injured shoulder joints.

A novel semi-supervised self-training method based on resampling for Twitter fake account identification

PurposeTwitter fake accounts refer to bot accounts created by third-party organizations to influence public opinion, commercial propaganda or impersonate others. The effective identification of bot accounts is conducive to accurately judge the disseminated information for the public. However, in actual fake account identification, it is expensive and inefficient to manually label Twitter accounts, and the labeled data are usually unbalanced in classes. To this end, the authors propose a novel framework to solve these problems.Design/methodology/approachIn the proposed framework, the authors introduce the concept of semi-supervised self-training learning and apply it to the real Twitter account data set from Kaggle. Specifically, the authors first train the classifier in the initial small amount of labeled account data, then use the trained classifier to automatically label large-scale unlabeled account data. Next, iteratively select high confidence instances from unlabeled data to expand the labeled data. Finally, an expanded Twitter account training set is obtained. It is worth mentioning that the resampling technique is integrated into the self-training process, and the data class is balanced at the initial stage of the self-training iteration.FindingsThe proposed framework effectively improves labeling efficiency and reduces the influence of class imbalance. It shows excellent identification results on 6 different base classifiers, especially for the initial small-scale labeled Twitter accounts.Originality/valueThis paper provides novel insights in identifying Twitter fake accounts. First, the authors take the lead in introducing a self-training method to automatically label Twitter accounts from the semi-supervised background. Second, the resampling technique is integrated into the self-training process to effectively reduce the influence of class imbalance on the identification effect.

Data repair of density-based data cleaning approach using conditional functional dependencies

PurposeData quality is a major challenge in data management. For organizations, the cleanliness of data is a significant problem that affects many business activities. Errors in data occur for different reasons, such as violation of business rules. However, because of the huge amount of data, manual cleaning alone is infeasible. Methods are required to repair and clean the dirty data through automatic detection, which are data quality issues to address. The purpose of this work is to extend the density-based data cleaning approach using conditional functional dependencies to achieve better data repair.Design/methodology/approachA set of conditional functional dependencies is introduced as an input to the density-based data cleaning algorithm. The algorithm repairs inconsistent data using this set.FindingsThis new approach was evaluated through experiments on real-world as well as synthetic datasets. The repair quality was determined using the F-measure. The results showed that the quality and scalability of the density-based data cleaning approach improved when conditional functional dependencies were introduced.Originality/valueConditional functional dependencies capture semantic errors among data values. This work demonstrates that the density-based data cleaning approach can be improved in terms of repairing inconsistent data by using conditional functional dependencies.

A graph-based approach for representing, integrating and analysing neuroscience data: the case of the murine basal ganglia

PurposeNeuroscience data are spread across a variety of sources, typically provisioned through ad-hoc and non-standard approaches and formats and often have no connection to the related data sources. These make it difficult for researchers to understand, integrate and reuse brain-related data. The aim of this study is to show that a graph-based approach offers an effective mean for representing, analysing and accessing brain-related data, which is highly interconnected, evolving over time and often needed in combination.Design/methodology/approachThe authors present an approach for organising brain-related data in a graph model. The approach is exemplified in the case of a unique data set of quantitative neuroanatomical data about the murine basal ganglia––a group of nuclei in the brain essential for processing information related to movement. Specifically, the murine basal ganglia data set is modelled as a graph, integrated with relevant data from third-party repositories, published through a Web-based user interface and API, analysed from exploratory and confirmatory perspectives using popular graph algorithms to extract new insights.FindingsThe evaluation of the graph model and the results of the graph data analysis and usability study of the user interface suggest that graph-based data management in the neuroscience domain is a promising approach, since it enables integration of various disparate data sources and improves understanding and usability of data.Originality/valueThe study provides a practical and generic approach for representing, integrating, analysing and provisioning brain-related data and a set of software tools to support the proposed approach.

An improved density-based approach to risk assessment on railway investment

PurposeDensity-based spatial clustering of applications with noise (DBSCAN) is the most commonly used density-based clustering algorithm, while it cannot be directly applied to the railway investment risk assessment. To overcome the shortcomings of calculation method and parameter limits of DBSCAN, this paper proposes a new algorithm called Improved Multiple Density-based Spatial clustering of Applications with Noise (IM-DBSCAN) based on the DBSCAN and rough set theory.Design/methodology/approachFirst, the authors develop an improved affinity propagation (AP) algorithm, which is then combined with the DBSCAN (hereinafter referred to as AP-DBSCAN for short) to improve the parameter setting and efficiency of the DBSCAN. Second, the IM-DBSCAN algorithm, which consists of the AP-DBSCAN and a modified rough set, is designed to investigate the railway investment risk. Finally, the IM-DBSCAN algorithm is tested on the China–Laos railway's investment risk assessment, and its performance is compared with other related algorithms.FindingsThe IM-DBSCAN algorithm is implemented on China–Laos railway's investment risk assessment and compares with other related algorithms. The clustering results validate that the AP-DBSCAN algorithm is feasible and efficient in terms of clustering accuracy and operating time. In addition, the experimental results also indicate that the IM-DBSCAN algorithm can be used as an effective method for the prospective risk assessment in railway investment.Originality/valueThis study proposes IM-DBSCAN algorithm that consists of the AP-DBSCAN and a modified rough set to study the railway investment risk. Different from the existing clustering algorithms, AP-DBSCAN put forward the density calculation method to simplify the process of optimizing DBSCAN parameters. Instead of using Euclidean distance approach, the cutoff distance method is introduced to improve the similarity measure for optimizing the parameters. The developed AP-DBSCAN is used to classify the China–Laos railway's investment risk indicators more accurately. Combined with a modified rough set, the IM-DBSCAN algorithm is proposed to analyze the railway investment risk assessment. The contributions of this study can be summarized as follows: (1) Based on AP, DBSCAN, an integrated methodology AP-DBSCAN, which considers improving the parameter setting and efficiency, is proposed to classify railway risk indicators. (2) As AP-DBSCAN is a risk classification model rather than a risk calculation model, an IM-DBSCAN algorithm that consists of the AP-DBSCAN and a modified rough set is proposed to assess the railway investment risk. (3) Taking the China–Laos railway as a real-life case study, the effectiveness and superiority of the proposed IM-DBSCAN algorithm are verified through a set of experiments compared with other state-of-the-art algorithms.

Decision tree-based classification in coastal area integrating polarimetric SAR and optical data

PurposeAs one of the world's most productive ecosystems, ecological land plays an important role in regional and global environments. Utilizing advanced optical and synthetic aperture radar (SAR) data for land cover/land use research becomes increasingly popular. This research aims to investigate the complementarity of fully polarimetric SAR and optical imaging for ecological land classification in the eastern coastal area of China.Design/methodology/approachFour polarimetric decomposition methods, namely, H/Alpha, Yamaguchi3, VanZyl3 and Krogager, were applied to Advanced Land Observing Satellite (ALOS) SAR image for scattering parameter extraction. These parameters were merged with ALOS optical parameters for subsequent classification using the object-based quick, unbiased, efficient statistical tree decision tree method.FindingsThe experimental results indicate that an improved classification performance was obtained in the decision level when merging the two data sources. In fact, unlike classification using only optical images, the proposed approach allowed to distinguish ecological land with similar spectrum but different scattering. Moreover, unlike classification using only polarimetric information, the integration of polarimetric and optical data allows to accurately distinguish reed from artemisia and sand from salt field and therefore achieve a detailed classification of the coastal area characteristics.Originality/valueThis research proposed an integrated classification method for coastal ecological land with polarimetric SAR and optical data. The object-based and decision-level fusion enables effective ecological land classification in coastal area was verified.