Unveiling and Unraveling Aggregation and Dispersion Fallacies in Group MCDM

2023 | Group Decision and Negotiation | Citations: 0

Authors: Mohammadi, Majid; Tamburri, Damian A.; Rezaei, Jafar

Abstract: Priorities in multi-criteria decision-making (MCDM) convey the relevance prefere ... Expand

Abstract: Priorities in multi-criteria decision-making (MCDM) convey the relevance preference of one criterion over another, which is usually reflected by imposing the non-negativity and unit-sum constraints. The processing of such priorities is different than other unconstrained data, but this point is often neglected by researchers, which results in fallacious statistical analysis. This article studies three prevalent fallacies in group MCDM along with solutions based on compositional data analysis to avoid misusing statistical operations. First, we use a compositional approach to aggregate the priorities of a group of DMs and show that the outcome of the compositional analysis is identical to the normalized geometric mean, meaning that the arithmetic mean should be avoided. Furthermore, a new aggregation method is developed, which is a robust surrogate for the geometric mean. We also discuss the errors in computing measures of dispersion, including standard deviation and distance functions. Discussing the fallacies in computing the standard deviation, we provide a probabilistic criteria ranking by developing proper Bayesian tests, where we calculate the extent to which a criterion is more important than another. Finally, we explain the errors in computing the distance between priorities, and a clustering algorithm is specially tailored based on proper distance metrics. Collapse

Topics: decision making

Methods: descriptive statistic computational algorithm cluster analysis Wilcoxon signed-rank test nonparametric test

Theories: SERVQUAL

Data engineering for fraud detection

2021 | Decision Support Systems | Citations: 1

Authors: Baesens, Bart; Höppner, Sebastiaan; Verdonck, Tim

Abstract: Financial institutions increasingly rely upon data-driven methods for developing ... Expand

Abstract: Financial institutions increasingly rely upon data-driven methods for developing fraud detection systems, which are able to automatically detect and block fraudulent transactions. From a machine learning perspective, the task of detecting suspicious transactions is a binary classification problem and therefore many techniques can be applied. Interpretability is however of utmost importance for the management to have confidence in the model and for designing fraud prevention strategies. Moreover, models that enable the fraud experts to understand the underlying reasons why a case is flagged as suspicious will greatly facilitate their job of investigating the sus­ picious transactions. Therefore, we propose several data engineering techniques to improve the performance of an analytical model while retaining the interpretability property. Our data engineering process is decomposed into several feature and instance engineering steps. We illustrate the improvement in performance of these data engineering steps for popular analytical models on a real payment transactions data set. Collapse

Topics: electronic authentication fraud detection electronic commerce online banking mobile application

Methods: machine learning descriptive statistic experiment feature engineering mathematical model

Opinion Formation Threshold Estimates from Different Combinations of Social Media Data-Types

2019 | HICSS | Citations: 0

Authors: Asher, Derrik E; Caylor, Justine; Doyle, Casey; Neigel, Alexis R; Korniss, Gyorgy; Szymanski, Bolek K

Abstract: Passive consumption of a quantifiable amount of social media information related ... Expand

Abstract: Passive consumption of a quantifiable amount of social media information related to a topic can cause individuals to form opinions. If a substantial amount of these individuals are motivated to take action from their recently established opinions, a movement or public opinion shift can be induced independent of the information’s veracity. Given that social media is ubiquitous in modern society, it is imperative that we understand the threshold at which social media data results in opinion formation. The present study estimates population opinion formation thresholds by querying 2222 participants about the number of various social media data-types (i.e., images, videos, and/or messages) that they would need to passively consume to form opinions. Opinion formation is assessed across three dimensions, 1) data-type(s), 2) context, 3) and source. This work provides a theoretical basis for estimating the amount of data needed to influence a population through social media information. Collapse

Topics: social media crowdsourcing social influence

Methods: experiment parametric test ANOVA descriptive statistic goodness of fit statistical test

See No Evil, Hear No Evil? Dissecting the Impact of Online Hacker Forums

2019 | Management Information Systems Quarterly | Citations: 26

Authors: Yue, Wei T.; Qiu-Hong Wang; Kai-Lung Hui

Abstract: Online hacker forums offer a prominent avenue for sharing hacking knowledge. Usi ... Expand

Abstract: Online hacker forums offer a prominent avenue for sharing hacking knowledge. Using a field dataset culled from multiple sources, we find that online discussion of distributed denial of service (DDOS) attacks in hackforums.net decreases the number of DDOS-attack victims. A 1% increase in discussion decreases DDOS attacks by 0.032% to 0.122%. This means that two DDOS-attack posts per day could reduce the number of victims by 700 to 2,600 per day. We find that discussion topics with similar keywords can variously increase or decrease DDOS attacks, meaning we cannot ascertain the impact of the discussion just by the post nature. Mentioning botnets, especially new botnets, increases the attacks, but the follow-up discussion decreases the attacks. Our results suggest that online-hacker-forum discussion may exhibit the dual-use characteristic. That is, it can be used for both good and bad purposes. We draw related managerial implications. Collapse

Topics: denial of service attack internet technology computer virus Python open source

Methods: term frequency–inverse document frequency latent dirichlet allocation forums posting qualitative content analysis topic model

Prognostic Model Development with Missing Labels: A Condition-Based Maintenance Approach Using Machine Learning

2019 | Business & Information Systems Engineering | Citations: 0

Authors: Zschech, Patrick; Heinrich, Kai; Bink, Raphael; Neufeld, Janis S.

Abstract: Condition-based maintenance (CBM) has emerged as a proactive strategy for deter ... Expand

Abstract: Condition-based maintenance (CBM) has emerged as a proactive strategy for determining the best time for maintenance activities. In this paper, a case of a milling process with imperfect maintenance at a German automotive manufacturer is considered. Its major challenge is that only data with missing labels are available, which does not provide a sufficient basis for classical prognostic maintenance models. To overcome this shortcoming, a data science study is carried out that combines several analytical methods, especially from the field of machine learning (ML). These include time-domain and time-frequency domain techniques for feature extraction, agglomerative hierarchical clustering and time series clustering for unsupervised pattern detection, as well as a recurrent neural network for prognostic model training. With the approach developed, it is possible to replace decisions that were made based on subjective criteria with data-driven decisions to increase the tool life of the milling machines. The solution can be employed beyond the presented case to similar maintenance scenarios as the basis for decision support and prognostic model development. Moreover, it helps to further close the gap between ML research and the practical implementation of CBM. Collapse

Topics: decision support decision making

Methods: cluster analysis hierarchical clustering computational algorithm descriptive statistic longitudinal research

Crisis Communication on Twitter during a Global Crisis of Volkswagen – The Case of “Dieselgate”

2018 | HICSS | Citations: 0

Authors: Stieglitz, Stefan; Mirbabaie, Milad; Potthoff, Tobias

Abstract: In this study, we investigate the communication behaviour in Twitter during the ... Expand

Abstract: In this study, we investigate the communication behaviour in Twitter during the rise of a corporate crisis. In September 2015, the emission scandal of Volkswagen (also known as “Dieselgate”) became public. We collected Twitter data and analysed approximately 400,000 tweets regarding the Volkswagen crisis. We take different perspectives on the data, by 1) separating the overall communication in peak and quiet phases, 2) analysing the sentiment in each phase, 3) looking at specific tweet contents, and 4) using statistical analyses to determine the significance of differences. Furthermore, we mapped the publishing behaviour of official Volkswagen accounts to the situational crisis communication theory (SCCT). The findings suggest that Volkswagen followed a strategy that is not covered by SCCT, i.e. keeping silent. Volkswagen’s tweets were not able to reduce the emotionality and sentiment of the ongoing Twitter discussion. Instead, even during quiet phases, the communication remained rather negative. Collapse

Topics: Twitter social media application programming interface user behavior customer service

Methods: sentiment analysis nonparametric test case study Kruskal Wallis test descriptive statistic

Theories: communication theory

Market sentiment dispersion and its effects on stock return and volatility

2017 | Electronic Markets | Citations: 4

Authors: See-To, Eric. W. K.; Yang, Yang

Abstract: Behavioral economics has revealed that investor sentiment can profoundly affect ... Expand

Abstract: Behavioral economics has revealed that investor sentiment can profoundly affect individual behavior and decision-making. Recently, the question is no longer whether investor sentiment affects stock market valuation, but how to directly measure investor sentiment and quantify its effects. Before the era of big data, research uses proxies as a mediator to indirectly measure investor sentiment, which has proved elusive due to insufficient data points. In addition, most of extant sentiment analysis studies focus on institutional investors instead of individual investors. This is despite the fact that United States individual investors have been holding around 50% of the stock market in direct stock investments. In order to overcome difficulties in measuring sentiment and endorse the importance of individual investors, we examine the role of individual sentiment dispersion in stock market. In particular, we investigate whether sentiment dispersion contains information about future stock returns and realized volatility. Leveraging on development of big data and recent advances in data and text mining techniques, we capture 1,170,414 data points from Twitter and used a text mining method to extract sentiment and applied both linear regression and Support Vector Regression; found that individual sentiment dispersion contains information about stock realized volatility, and can be used to increase the prediction accuracy. We expect our results contribute to extant theories of electronic market financial behavior by directly measuring the individual sentiment dispersion; raising a new perspective to assess the impact of investor opinion on stock market; and recommending a supplementary investing approach using user-generated content. Collapse

Topics: Twitter social media big data accounting user-generated content

Methods: naïve bayes sentiment analysis descriptive statistic F-test parametric test