Dataset Of Users Research Articles

Fast Parallelizable Cassava Plant Disease Detection using Ensemble Learning with Fine Tuned AmoebaNet and ResNeXt-101

The study focuses on applying advanced machine learning approaches, namely deep learning models and ensemble learning, to develop an automated system for Cassava Brown Streak Disease (CBSD) detection using robotic systems. We leverage an extensively used dataset to evaluate our proposed methodologies. The dataset comprises thousands of labeled images illustrating various stages of CBSD infection. Our experiment is furnished with high-resolution cameras and sophisticated image-processing algorithms. We finetune AmoebaNet and ResNeXt-101 32x16d models using the dataset to distinguish between healthy and diseased cassava plants. Ensemble learning is then applied to consolidate the prediction outputs from both models, consequently enhancing overall diagnostic accuracy. The implemented system showcases exceptional performance, delivering high precision and recall rates in recognizing CBSD cases. Through automation, our solution significantly diminishes the reliance on human expertise, streamlines the diagnostic procedure, and extends the reach of CBSD detection applications. This pioneering research marks a significant stride forward in fortifying food security and promoting sustainable agriculture in regions affected by CBSD. Further details about the dataset usage will be discussed in the Methodology section.

Predicting Reputation in the Sharing Economy with Twitter Social Data

In recent years, the sharing economy has become popular, with outstanding examples such as Airbnb, Uber, or BlaBlaCar, to name a few. In the sharing economy, users provide goods and services in a peer-to-peer scheme and expose themselves to material and personal risks. Thus, an essential component of its success is its capability to build trust among strangers. This goal is achieved usually by creating reputation systems where users rate each other after each transaction. Nevertheless, these systems present challenges such as the lack of information about new users or the reliability of peer ratings. However, users leave their digital footprints on many social networks. These social footprints are used for inferring personal information (e.g., personality and consumer habits) and social behaviors (e.g., flu propagation). This article proposes to advance the state of the art on reputation systems by researching how digital footprints coming from social networks can be used to predict future behaviors on sharing economy platforms. In particular, we have focused on predicting the reputation of users in the second-hand market Wallapop based solely on their users’ Twitter profiles. The main contributions of this research are twofold: (a) a reputation prediction model based on social data; and (b) an anonymized dataset of paired users in the sharing economy site Wallapop and Twitter, which has been collected using the user self-mentioning strategy.

Convolutional Neural Networks Based Algorithm for Speech Separation

Vocal Separation is the separation of a set of source signals from a set of mixed signals, without extensive information about the source signals or the mixing process. Audio data is available in large amounts across the internet. Usage of machine learning and datasets constructed around one context can be used to design a system capable of performing vocal separation on audio data. Our algorithm uses a convolutional neural network, down sampling the input and then up sampling to a desired output. A custom dataset comprised of mixed conversations is used for training, with the expected output being the two vocal sources correctly separated.

Research on Data Mining and Investment Recommendation of Individual Users Based on Financial Time Series Analysis

With the continuous development of financial information technology, traditional data mining technology cannot effectively deal with large-scale user data sets, nor is it suitable to actively discover various potential rules from a large number of data and predict future trends. Time series are the specific values of statistical indicators on different time scales. Data sequences arranged in chronological order exist in our lives and scientific research. Financial time series is a special kind of time series, which has the commonness of time series, chaos, non-stationary and non-linear characteristics. Financial time series analysis judges the future trend of change through the analysis of historical time series. Through in-depth analysis of massive financial data, mining its potential valuable information, it can be used for individual or financial institutions in various financial activities, such as investment decision-making, market forecasting, risk management, customer requirement analysis provides scientific evidence.

ContextPCA: Predicting Context-Aware Smartphone Apps Usage Based On Machine Learning Techniques

This paper mainly formulates the problem of predicting context-aware smartphone apps usage based on machine learning techniques. In the real world, people use various kinds of smartphone apps differently in different contexts that include both the user-centric context and device-centric context. In the area of artificial intelligence and machine learning, decision tree model is one of the most popular approaches for predicting context-aware smartphone usage. However, real-life smartphone apps usage data may contain higher dimensions of contexts, which may cause several issues such as increases model complexity, may arise over-fitting problem, and consequently decreases the prediction accuracy of the context-aware model. In order to address these issues, in this paper, we present an effective principal component analysis (PCA) based context-aware smartphone apps prediction model, “ContextPCA” using decision tree machine learning classification technique. PCA is an unsupervised machine learning technique that can be used to separate symmetric and asymmetric components, and has been adopted in our “ContextPCA” model, in order to reduce the context dimensions of the original data set. The experimental results on smartphone apps usage datasets show that “ContextPCA” model effectively predicts context-aware smartphone apps in terms of precision, recall, f-score and ROC values in various test cases.

Electricity Usage Efficiency and Electricity Demand Modeling in the Case of Germany and the UK

In this article, monthly and yearly electricity consumption predictions for the German power market were calculated using the multiple variable regression model. This model accounts for several factors that are often neglected when forecasting electricity demand in practice, in particular the role of the higher efficiency of electricity usage from year to year. The analysis performed in this paper helps to explain why no growth in power consumption has been observed in Germany during the last decade. It shows that the electricity efficiency usage dataset is a relevant input for the model, which mitigates the combined impact of other factors on the final electricity consumption. The electricity demand forecasting model presented in this article was built in the year 2013 with forecasts for the future years’ electricity demand in Germany provided until 2020. These forecasts and related findings are also evaluated in this article.

Towards De-Anonymization of Google Play Search Rank Fraud

Search rank fraud, the fraudulent promotion of products hosted on peer-review sites, is driven by expert workers recruited online, often from crowdsourcing sites. In this paper we introduce the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">fraud de-anonymization</i> problem, that goes beyond fraud detection, to unmask the human masterminds responsible for posting search rank fraud in peer-review sites. We collect and study data from crowdsourced search rank fraud jobs, and survey the capabilities and behaviors of 58 search rank fraud workers recruited from 6 crowdsourcing sites. We collect a gold standard dataset of Google Play user accounts attributed to 23 crowdsourced workers and analyze their fraudulent behaviors in the wild. We propose <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Dolos</small> , a fraud de-anonymization system that leverages traits and behaviors we extract from our studies, to attribute detected fraud to crowdsourcing site workers, thus to real identities and bank accounts. We introduce MCDense, a min-cut dense component detection algorithm to uncover groups of user accounts controlled by <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">different</i> workers, and use stylometry and supervised learning to attribute them to crowdsourcing site profiles. <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Dolos</small> correctly identified the owners of 95 percent of fraud worker-controlled communities, and uncovered fraud workers who promoted as many as 97.5 percent of fraud apps we collected from Google Play. When evaluated on 13,087 apps (820,760 reviews), which we monitored over more than 6 months, <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Dolos</small> identified 1,056 apps with suspicious reviewer groups. We report orthogonal evidence of their fraud, including fraud duplicates and fraud re-posts. <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Dolos</small> significantly outperformed adapted dense subgraph detection and loopy belief propagation competitors, on two new coverage scores that measure the quality of detected community partitions.

Sequential Recommendations on Board-Game Platforms

The game market is an increasingly large industry. The board-game market, which is the most traditional in the game market, continues to show a steady growth. It is very important for both publishers and players to predict the propensity of users in this huge market and to recommend new games. Despite its importance, no study has been performed on board-game recommendation systems. In this study, we propose a method to build a deep-learning-based recommendation system using large-scale user data of an online community related to board games. Our study showed that new games can be effectively recommended for board-game users based on user big data accumulated for a long time. This is the first study to propose a personalized recommendation system for users in the board-game market and to introduce a provision of new large datasets for board-game users. The proposed dataset shares symmetric characteristics with other datasets and has shown its ability to be applied to various recommendation systems through experiments. Therefore, the dataset and recommendation system proposed in this study are expected to be applied for various studies in the field.

Azimuthal Anisotropy of the North American Upper Mantle Based on Full Waveform Inversion

AbstractA new azimuthal anisotropy model for the North American and Caribbean Plates, namely, , is constructed based on full waveform inversion and records from the USArray and other temporary/permanent networks deployed in the study region. A total of 180 earthquakes and 4,516 seismographic stations are employed in the inversion to simultaneously constrain radially and azimuthally anisotropic model parameters: , , , and , within the crust and mantle. Thirty‐two preconditioned conjugate gradient iterations have been utilized to minimize frequency‐dependent phase discrepancies between observed and predicted seismograms for three‐component short‐period (15–40 s) body waves and long‐period (25–100 s) surface waves. Model exhibits complicated variations in anisotropic fabrics underneath the western and eastern United States, especially at depths shallower than 100 km. For instance, the fast axis orientations in model suggest the presence of trench‐perpendicular mantle flows underneath the Cascadia Subduction Zone and also follow the strikes of the Snake River Plain, the Ouachita Orogenic Front, and the Grenville and Appalachian Orogenic Belts. The amplitudes of azimuthal anisotropy reduce to around 1% at depths greater than 200 km, and the orientations are subparallel to the global plate motion directions to the east of the Rocky Mountain, except for large discrepancies in central and eastern Canada. At a depth of 700 km, the fast axes change along the trajectory of the Farallon slab underneath the Great Lakes region and Gulf of Mexico, which might indicate the development of 2‐D poloidal‐mode mantle flows perpendicular to the strike of the sinking slab within the uppermost lower mantle. Comparisons between model with a western U.S. model from ambient noise tomography and SKS splitting measurements demonstrate a relatively good agreement for the fast axis orientations, considering the usage of different data sets and imaging techniques. However, the absolute magnitude of azimuthal anisotropy in model might be underestimated, especially at greater depths, given the poor agreement on the amplitudes of predicted and observed SKS splitting times. At the current stage, the agreement among different azimuthal anisotropy models at global and continental scales is still poor even for the United States with a dense station coverage.

Assigning credit to scientific datasets using article citation networks

A citation is a well-established mechanism for connecting scientific artifacts. Citation networks are used by citation analysis for a variety of reasons, prominently to give credit to scientists’ work. However, because of current citation practices, scientists tend to cite only publications, leaving out other types of artifacts such as datasets. Datasets then do not get appropriate credit even though they are increasingly reused and experimented with. We develop a network flow measure, called DataRank , aimed at solving this gap. DataRank assigns a relative value to each node in the network based on how citations flow through the graph, differentiating publication and dataset flow rates. We evaluate the quality of DataRank by estimating its accuracy at predicting the usage of real datasets: web visits to GenBank and downloads of Figshare datasets. We show that DataRank is better at predicting this usage compared to alternatives while offering additional interpretable outcomes. We discuss improvements to citation behavior and algorithms to properly track and assign credit to datasets.

Electricity Theft Pinpointing Through Correlation Analysis of Master and Individual Meter Readings

Electricity theft costs utility companies billions of dollars worldwide annually. The electricity consumption data recorded by consumers' smart meters, coupled with the aggregate energy supply data recorded by master meters provide a new opportunity to pinpoint the source of electricity theft. Existing works on electricity theft pinpointing either assume linear attack modes which often limit their capability in identifying nonlinear electricity theft behaviours, or incur extra cost for model training or sensor installation. Our insight hinges upon the fact that the value of electricity theft loss (ETL) should be more correlated to the meter readings of energy thieves than to those of honest consumers. Guided by this insight, we formulate the problem of electricity theft pinpointing as a time-series correlation analysis problem which does not require linearity assumption of attack modes or any cost of training. Two coefficients are defined to evaluate the suspicion level of a consumer's reported energy consumption pattern. A comprehensive set of experiments has been conducted on a real-world energy usage dataset with several types of attacks, and the results show that our proposed technique significantly improves the pinpointing accuracy when compared with other state-of-the-art methods.

Predicting the antecedents of trust in social commerce – A hybrid structural equation modeling with neural network approach

Trust is an essential concern in s-commerce. Though existing research has studied the association between trust and purchasing intention; the determinants of the formation of trust in s-commerce remain largely unexplored. This study examines the determinants of trust in s-commerce based on social presence and social support. Unlike most business research, we applied a hybrid SEM-ANN approach that can detect non-linear and non-compensatory relationships. Linear and compensatory models assume that a shortfall in one factor may be compensated by other factors. However, consumer decision-making processes are complicated and non-compensatory and linear models tend to oversimplify these processes. Criterion sampling was used to gather 462 datasets of social commerce users using a mall intercept technique. Information support has the strongest effect followed by the social presence of interaction with the sellers, income and social presence of others. The integrated model predicts 76.9% trust in s-commerce. Theoretical and managerial contributions are discussed.

A novel algorithm for estimation of Twitter users location using public available information

Abstract Social media networks are an attractive and hot research area in the big data community because of their numerous active users. One of the most widely studied topics in social networks is the prediction from the public available data. Recently, researchers have successfully predicted many statistical and human properties from social media networks using different machine learning algorithms. In this paper, a new efficient and accurate algorithm is proposed to predict the country location of a Twitter user using his or her public information only. The proposed algorithm employs the public information of the Twitter user and that of his or her followers and friends to predict his or her location without using GPS data. A convenient data set of Twitter users is gathered and used to test our proposed algorithm using KNIME software. The proposed algorithm is compared with other state-of-the-art algorithms, and results showed that our proposed algorithm significantly outperforms other location detection algorithms by using Twitter users from different countries.

Fuzzy ARTMAP driven features selection: towards a system to design information fusion for the geographic discovery

The availability of ground-level and overhead images is increasing fast. However, the potentialities of joint usage of these images on information fusion systems for in-deep geographic discovery, by answering the typical binomial quest: what and where are still to be completely evaluated. It is, therefore, necessary to establish a variety of procedures to study these potentialities including on how to assess domain-invariance between this imagery related to ground-level and overhead capture domains for example. We addressed in this paper a novel framework for one prospect for an optimal set of handcrafted domain-invariant features for ground the optimized usage of multiple sources of information on a feature level based fusion system. All of this was done framed by the usage of small datasets and, also, of interpretable classifiers (human-centered computing) and it paves the way towards the proposition of application software to be used potentially as a didactic tool with emphasis on the botanic domain. The results also endorse a domain-invariance that was already implicit on the results of previously published papers. Broadly evaluating, these results connecting image datasets on contrasting scales can launch new investigations in the scope of scale-free phenomena.

Potential-Driven Model for Influence Maximization in Social Networks

The adoption of a new technology, the spreading of ideologies, the transmission of contagions, are examples of spreading phenomena that can be modelled at two distinct levels of resolution, a macroscopic and a microscopic level.In the present work, we consider the network at the microscopic level and represent its nodes as a system of interacting particles and propose a novel, physics-inspired approach. This approach assumes that nodes interact via `forces' that derive from the `potential' that each node creates at the location of the other nodes, leading to a potential gradient that indicates the `natural' direction of diffusion through the network. A set of influencers in the network, is determined from strategically selected nodes based on the value of their net potential. We use synthetic networks of various sizes and compare the influence spread resulting from a seed set determined from the potential-based model and alternate approaches, including the greedy algorithm of Kempe et al. Our findings indicate that this approach achieves comparable results to those of the greedy algorithm without the prohibitive computational cost, and consistently outperforms the other approaches by a large margin. We then apply our methodology to information spreading related to the Twitter Higgs Boson user dataset. The results are analyzed in the context of influence maximization, and we provide insights into the general application of this technique for information dissemination.

Benchmark and Parameter Sensitivity Analysis of Single-Cell RNA Sequencing Clustering Methods.

Single-cell RNA-seq (scRNAseq) is a powerful tool to study heterogeneity of cells. Recently, several clustering based methods have been proposed to identify distinct cell populations. These methods are based on different statistical models and usually require to perform several additional steps, such as preprocessing or dimension reduction, before applying the clustering algorithm. Individual steps are often controlled by method-specific parameters, permitting the method to be used in different modes on the same datasets, depending on the user choices. The large number of possibilities that these methods provide can intimidate non-expert users, since the available choices are not always clearly documented. In addition, to date, no large studies have invistigated the role and the impact that these choices can have in different experimental contexts. This work aims to provide new insights into the advantages and drawbacks of scRNAseq clustering methods and describe the ranges of possibilities that are offered to users. In particular, we provide an extensive evaluation of several methods with respect to different modes of usage and parameter settings by applying them to real and simulated datasets that vary in terms of dimensionality, number of cell populations or levels of noise. Remarkably, the results presented here show that great variability in the performance of the models is strongly attributed to the choice of the user-specific parameter settings. We describe several tendencies in the performance attributed to their modes of usage and different types of datasets, and identify which methods are strongly affected by data dimensionality in terms of computational time. Finally, we highlight some open challenges in scRNAseq data clustering, such as those related to the identification of the number of clusters.

Blackmarket-Driven Collusion Among Retweeters–Analysis, Detection, and Characterization

The growth of online social media has led to a huge increase in the number of users who want to share and publicize various kinds of information. Twitter, the most popular micro-blogging platform, has become a hotbed for users who are involved in different activities such as news publishing, job hunting, recruiting, advertising and publicity. Retweeting a tweet is a major action to broadcast a user’s message out to millions of users. Retweet action has two major advantages: (i) gaining quick exposure to the content, and (ii) increasing likelihood of gaining new Twitter followers in return. The organic way of gaining a larger number of retweets is a time consuming process, which leads to the creation of unfair methods to gain retweets. Thus, Twitter users often approach various blackmarket services to gain retweets inorganically in a short duration. Blackmarkets spread their collusive ecosystem in such a way that Twitter is unable to detect them even after devoting significant effort to purge the platform off bots, trolls, and fake accounts. One major reason behind the evasion is that the collusive users involved in blackmarket services exhibit a mix of organic and inorganic behavior – they organically reweet some genuine tweets; at the same time, they inorganically retweet tweets submitted to blackmarket services. This paper is the first attempt to provide a thorough study of the collusive users involved in two types of blackmarket services – Premium and Freemium. We collect a novel dataset of collusive users comprising of users from both types of blackmarket services. We provide network-centric, profile-centric, timeline-centric and retweet-centric characteristics of these users and show how users involved in premium blackmarket services exhibit diverse behavior as compared to those involved in freemium services. We further employ human annotators to label collusive users into three types: bots, promotional customers, and normal customers. We then curate 63 novel features to run state-of-the-art classifiers in two settings – binary classification (collusive vs. genuine) and multi-class classification (bot, promotional customers, normal customers, and genuine users). Bagging achieves the best accuracy (macro F1-score of 0.892) in the former setting, whereas Random Forest outperforms others (macro F1-score of 0.791) in the latter setting. We also develop a chrome extension, SCoRe++ which can detect collusive retweeters in real time.

An Analysis of Information Systems Literature: Contributions to Fraud Research

This study analyzes the knowledge and methods used in information systems (IS) journals in the area of financial statement fraud. The purpose of this analysis is to provide tools and ideas to support interdisciplinary research in accounting and information systems for financial statement fraud topics. The study presents an analysis of five top ranking IS journals (MIS Quarterly, Information Systems Research, Communications of the ACM, Management Science, and Journal of MIS) and five top ranking IS conferences [International Conference on Information Systems (ICIS), Hawaii International Conference on System Sciences (HICSS), International Federation for Information Processing (IFIP), International Conference on Decision Support Systems (DSS), and Decision Sciences Institute National Conference (DSI)]. The literature found from these sources are categorized and presented by year, journal, contribution, type of study, methodology, data set usage, and research design. Although the literature varies, a common thread in many studies is the use of data mining and/or machine learning models to detect fraud.

Cyber-Security: The Use of Big Data Analytic Model for Network Intrusion Detection Classification

Cybersecurity is seen as a major player in the protection of Internet-connected systems, including hardware, software and data, from cyberattacks and other malicious crimes in today’s densely connected world-Internet of Things (IoTs). The divers challenge facing Internet users as private and business entities is being advocated as not enough hinderance to seamless interfacing of Mobile Computing and Internet Applications presently making waves. Technology such as Intrusion Detection Systems (IDS) application into cyber-security is an evolving computing mechanism designed as a counter-measure to incessant network threats and intruders. It is one of most reliable pro-defensive tools and has gained significance over time. Meanwhile network traffic data being generated within the context of enormous Internet users requires the application of big data analytical tools for its analysis. This paper, therefore, employs the use of big data analytical tools with its machine learning algorithm on an open-source data set-KDD’99. The full data set was used in the analysis. Predictive model was built in less than 5 minutes time with 99.91% prediction accuracy. Computational challenge and only 10% data set usage, which could only be accounted for in previous research were overcome. Therefore, IDS could be better designed with integration of this classification model result. Keywords: Cyber-Security, Internet of Things, Intrusion, Mobile, Big data, network DOI : 10.7176/CEIS/10-7-02 Publication date: November 30th 2019

Credit Card Transaction Based on Face Recognition Technology

This paper proposes a method for credit card transaction system which will make use of face recognition and face detection technology, using Haar Cascade and GLCM algorithm. The main problem faced by credit card users is attack to lot of privacy issues such as credit card. This generally happens when users give their credit card number to unknown people or when the card is lost. So, we are proposing a system that will reduce the risk of credit card frauds. The system we are proposing will match the image of user’s face with dataset of respective user. A database will be maintained for authentication purpose. If the image matches, that means user is genuine and he will be allowed to proceed otherwise, the user will be denied to do the transaction.