Decision support tool to define the optimal pool testing strategy for SARS-CoV-2

2023 | Decision Support Systems | Citations: 0

Authors: Barracosa, Bruno; Felício, João; Carvalho, Ana; Moreira, Leonilde M.; Mendes, Filipa; Verde, Sandra Cabo; Pinto-Varela, Tânia

Abstract: This work proposes a holistic decision support tool to be used by diverse stakeh ... Expand

Abstract: This work proposes a holistic decision support tool to be used by diverse stakeholders and decision-makers, integrating experimental data. Considering the characteristics of each region, the tool aids in defining the most suitable testing strategies per country or region: number of tests, time, cost, type of pool strategy based on prevalence rate interval, and the use of centralization vs. decentralization of testing strategies. The tool con­ tributes to the field of decision support systems theoretically by introducing a novel model for testing decisions in extreme situations within a specific region, and practically, by demonstrating how laboratory data can be applied to real-life decision-making in that same region. This tool has been validated with two real case studies (Portugal and France) and has applied, as input data, real experimental results. The results suggested tree- and matrixbased strategies for low and high prevalence rates. A decentralized testing strategy improves the financial and time gains introduced by pool testing; however, centralizing the decisions shows a more straightforward implementation and resource management of the pool testing system with a low-performance loss. Therefore, there might be a solution to consider overcoming operational restrictions. Collapse

Topics: decision support system pandemic

Methods: case study optimization model experiment simulation machine learning

Robust and High-Accessibility Ranking Method for Crowdsourcing-Based Decision Making

2023 | Group Decision and Negotiation | Citations: 0

Authors: Nguyen, Phan-Anh-Huy; Hsu, Ping-Yu

Abstract: With the advancement of online technologies in recent years, crowdsourcing data ... Expand

Abstract: With the advancement of online technologies in recent years, crowdsourcing data has been used for numerous applications in many fields. The preference sequences obtained through crowdsourcing are valuable resources for ranking. However, the aggregation of incomplete and inconsistent preferences is complicated. To address these challenges, this study proposed a novel method termed robust crowd ranking (RCR) based on a consistent fuzzy c-means approach to increase the robustness and accessibility of aggregated preference sequences obtained through crowdsourcing. To verify the robustness, accessibility, and accuracy of RCR, comprehensive experiments were conducted using synthetic and real data. The simulation results validated that the RCR outperforms Borda Count, Dodgson, IRV and Tideman methods. Collapse

Topics: crowdsourcing Microsoft Windows programming language operating system internet technology

Methods: computational algorithm experiment simulation descriptive statistic cluster analysis

Examining the Low- Resolution Electromagnetic Tomography Technique for EEG Brain Mapping: Towards Methodological Advancement for IS Research

2023 | ACM SIGMIS Database | Citations: 0

Authors: Tejay, Gurvirender P.; Mohammed, Zareef A.

Abstract: NeuroIS presents a new opportunity for information systems research. Often used ... Expand

Abstract: NeuroIS presents a new opportunity for information systems research. Often used neuroscience techniques include brain mapping with the functional magnetic resonance imaging (fMRI) device or eventrelated potential time-domain studies with the electroencephalogram (EEG). The critics of EEG consider the poor spatial resolution as justification for EEG's inadequacy to brain mapping studies. However, the low-resolution electromagnetic tomography (LORETA) technique provides strong estimation parameters allowing EEG to perform brain mapping. This paper presents EEG (with lower number of channels) and LORETA techniques as an effective approach for exploratory investigation specially when researchers are constrained with lack of resources (specially at significantly lower costs). We demonstrate the effectiveness of EEG using sLORETA with respect to fMRI as proof-of-concept approach to study IS phenomenon. The results of such studies can serve as a preliminary step for further analysis with the use of more sophisticated neuroscience devices. This can enhance IS research by taking advantage of both high temporal and spatial resolution leading to reduced estimation errors of neural activity and stronger basis for correlating neural activity and specific tasks. We also present a set of guidelines for using the LORETA family of techniques in IS research. Collapse

Topics: digital distribution mobile application market corporate social responsibility usability

Methods: electroencephalography functional magnetic resonance imaging design artifact NeuroIS Student's t-test

Which voice are you satisfied with? Understanding the physician–patient voice interactions on online health platforms

2022 | Decision Support Systems | Citations: 0

Authors: Liu, Shan; Si, Guangsen; Gao, Baojun

Abstract: Patients are usually sensitive to physicians’ attitude and behavior when physici ... Expand

Abstract: Patients are usually sensitive to physicians’ attitude and behavior when physicians make decisions in patients’ consultations. However, the consequences of physicians’ interactive voices on patients’ perception have rarely been explored. This study develops a multistage online interaction model to explore the effects of the key vocal emotional cues of physician voice on patient-perceived satisfaction. We collect data of over 44,000 voice in­ teractions regarding lung disease from 889 physicians on a leading online health platform. By using the theo­ retical lens of emotional contagion, we find that the voice pitch and intensity of physicians, respectively, positively and negatively affect patient-perceived satisfaction, in turn positively affecting physician economic returns in the postinteraction stage. However, the negative effect of intensity is low for physicians with high recommendation popularity (a factor in the preinteraction stage), suggesting that physician-inherent information has an anchoring effect. These findings not only provide new insights into the literature on online phys­ ician–patient interaction by identifying the novel nonverbal cues of physician voice, their consequent effects, and boundary conditions, but also provide decision support for physicians and platform managers. Collapse

Topics: electronic health decision support Python

Methods: cluster analysis support vector machine descriptive statistic machine learning natural language processing

Theories: social contagion theory

Combination Generalized Grey Target Decision Method for Mixed Attributes Based on Zero-Sum Game Theory

2022 | Group Decision and Negotiation | Citations: 0

Authors: Ma, Jinshan; Yuan, Zhiqi; Zheng, Guanghua; Lang, Fushan

Abstract: Generalized grey target decision method (GGTDM) for mixed attributes has differe ... Expand

Abstract: Generalized grey target decision method (GGTDM) for mixed attributes has different decision-making bases (DMBs) owing to different mechanisms. A combination GGTDM based on zero-sum game theory is proposed to obtain the desirable decision-making effect by fully using the characteristics and information of different DMB based GGTDMs. First, two players are set to compete in the zero-sum game. One player takes different DMB based GGTDMs as the set of his strategies, while the other player adopts the alternatives as the other player’s set of strategies. Then, the ranking values of all alternatives decided by different GGTDMs are converted into scores by equivalent way, and the scores are regarded as the payoffs of the two players. Next, the zero-sum game based linear programming model is built to obtain the solutions when reaching the Nash equilibrium. Finally, the combination decision-making values of all alternatives can be calculated based on the strategy solutions for all GGTDMs. And the decision making is based on the value with which the smaller the better. The application analysis verifies that the proposed method has its feasible and can make full use of the advantages of different DMB based GGTDMs and avoid their shortcomings. Collapse

Topics: decision making systems analysis innovation management

Methods: linear optimization descriptive statistic trade-off analysis method multiple attribute decision making cross sectional research

Theories: game theory

Using a FMEA–TIFIAD Approach to Identify the Risk of Railway Dangerous Goods Transportation System

2021 | Group Decision and Negotiation | Citations: 0

Authors: Huang, Wencheng; Li, Yueyang; Kou, Xingyi; Wang, Wenzhe; Xu, Yifei

Abstract: In this paper a systemic approach that combines Failure Mode and Effect Analysis ... Expand

Abstract: In this paper a systemic approach that combines Failure Mode and Effect Analysis (FMEA) and trapezoidal intuitionistic fuzzy information axiomatic design (TIFIAD) is proposed to identify the risk of railway dangerous goods transportation system (RDNGTS), which has seven steps: (1) identify potential risk factors and risk sub-indicators of RDNGTS; (2) define the failure criterion of severity of failures (S), portability of occurrence (O) and possibility of detection (D) of FMEA based on trapezoidal intuitionistic fuzzy numbers (TrIFNs); (3) invite experts to score the TrIFNs of S, O and D for each risk sub-indicator; (4) aggregate the TrIFNs of S, O and D for each risk sub-indicator; (5) calculate the TIFIAD of each risk sub-indicator; (6) calculate the total TIFIAD of each risk sub-indicator; (7) calculate the total TIFIAD of each risk factor. Compared with FMEA, FMEA–TIFIAD has the following improvements: (1) The TrIFNs are applied to score the S, O and D of each risk sub-indicator of RDNGTS; (2) The TIFIAD of the S, O and D are calculated for each risk sub-indicator, the product of the S, O and D could replace the Risk Priority Number (RPN) of FMEA; (3) The Entropy Weight Method is used to calculate the weight of each risk sub-indicator. The calculation results show that the potential human risk should be paid more attentions; the FMEA–TIFIAD is more reliable and accurate than FMEA. Finally the improvement suggestions for managers and operators of RDNGTS are given. Collapse

Topics: transportation management logistics management

Methods: case study literature study multiple case study

Theories: fuzzy set theory

User Community Development in Social Networks to Support AI-Enabled Knowledge Provision

2021 | International Conference on Information Systems | Citations: 0

Authors: Skulimowski, Andrzej

Abstract: Efficient building and maintenance of user networks is a fundamental problem fo ... Expand

Abstract: Efficient building and maintenance of user networks is a fundamental problem for the management of open-access information systems whose business model relies directly on social impact. This article presents a methodology of user community building and management for cloud-based AI-enabled knowledge repositories (AILPs) which are considered a relevant part of enterprise learning ecosystems. Their assessment is based on the number of users, user engagement and satisfaction. Therefore, the indicators of user community growth are a basis for evaluating the quality of IS management and services. To optimize IS business performance, we applied a social impact model, derived from multi-level simulation and a self-contained multicriteria analysis framework. The model combines controlled discrete-event systems, learning cellular automata, and anticipatory decision support. These all serve to optimize user network development strategies. The primary application of this model is the exploitation of an innovative digital knowledge platform, developed within a recent Horizon 2020 research project. Collapse

Topics: decision support system knowledge repository artificial intelligence online learning system cloud computing

Methods: simulation computational algorithm optimization model multi objective optimization logistic regression

Serial Integration, Real Innovation: Roles of Diverse Knowledge and Communicative Participation in Crowdsourcing

2021 | HICSS | Citations: 0

Authors: Sun, Yao; Majchrzak, Ann; Malhotra, Arvind

Abstract: Despite a burgeoning public and scholarly interest on open innovation and crowds ... Expand

Abstract: Despite a burgeoning public and scholarly interest on open innovation and crowdsourcing, how to enable members of online temporary crowd to maintain knowledge integration and innovation remains underexplored. This study seeks to understand the ways in which online crowd members collectively generate more innovative and serial integrative solutions to crowdsourced open innovation challenges. Analyzing 3,200 unique posts generated by 486 participants of 21 organization-sponsored online crowdsourcing innovation challenges, this research demonstrates that crowd members contribute more innovative solutions when being exposed to explicitly shared diverse knowledge, and that crowd members’ communicative participation acts as a catalyst for the production of both innovation and serial knowledge integration. Findings suggest that managers who seek to generate knowledge integration and innovation should endeavor to implement systems that afford highlevel communicative participation, as well as encourage crowd members to make their diverse knowledge explicit while minimizing their cognitive load in knowledge sharing. Collapse

Topics: crowdsourcing innovation management open innovation knowledge management knowledge sharing

Methods: statistical hypothesis test descriptive statistic regression analysis method k-means clustering

Design of electronic-commerce recommendation systems based on outlier mining

2021 | Electronic Markets | Citations: 0

Authors: Xia, Huosong; Wei, Xiang; An, Wuyue; Zhang, Zuopeng Justin; Sun, Zelin

Abstract: Prior studies mostly consider outliers as noise data and eliminate them, resulti ... Expand

Abstract: Prior studies mostly consider outliers as noise data and eliminate them, resulting in the loss of outlier knowledge. Based on the existing technology of recommendation systems and outlier detection, this research develops a new e-commerce recommended model from the perspective of outlier knowledge management. Specifically, we apply outlier data mining and integrate local outlier coefficients into the recommendation algorithm. The experimental results show that the proposed outlier extent recommendation model performs better than the traditional recommendation systems based on the collaborative filtering algorithm, which can effectively improve the quality of recommendation, enhance customer satisfaction and loyalty, and create potential benefits for the business. Our study contributes to the design of e-commerce recommending systems with some novel ideas and provides useful guidelines for developing the outlier extent. Collapse

Topics: recommender system electronic commerce collaborative filtering customer satisfaction knowledge management

Methods: computational algorithm experiment local outlier factor machine learning anomaly detection

Easy and Efficient Hyperparameter Optimization to Address Some Artificial Intelligence “ilities”

2020 | HICSS | Citations: 0

Authors: Bihl, Trevor J; Schoenbeck, Joseph; Steeneck, Daniel; Jordan, Jeremy

Abstract: Artificial Intelligence (AI), has many benefits, including the ability to find ... Expand

Abstract: Artificial Intelligence (AI), has many benefits, including the ability to find complex patterns, automation, and meaning making. Through these benefits, AI has revolutionized image processing among numerous other disciplines. AI further has the potential to revolutionize other domains; however, this will not happen until we can address the "ilities": repeatability, explain-ability, reliability, use-ability, trust-ability, etc. Notably, many problems with the "ilities" are due to the artistic nature of AI algorithm development, especially hyperparameter determination. AI algorithms are often crafted products with the hyperparameters learned experientially. As such, when applying the same algorithm to new problems, the algorithm may not perform due to inappropriate settings. This research aims to provide a straightforward and reliable approach to automatically determining suitable hyperparameter settings when given an AI algorithm. Results, show reasonable performance is possible and end-to-end examples are given for three deep learning algorithms and three different data problems. Collapse

Topics: artificial intelligence usability systems development Python

Methods: computational algorithm deep learning Cross-Industry Standard Process for Data Mining parametric test Student's t-test