Human-machine collaboration for feature selection and integration to improve congestive Heart failure risk prediction

2023 | Decision Support Systems | Citations: 0

Authors: Ben-Assuli, Ofir; Heart, Tsipi; Klempfner, Robert; Padman, Rema

Abstract: The issue of harnessing machine learning (ML) algorithms for the prediction of a ... Expand

Abstract: The issue of harnessing machine learning (ML) algorithms for the prediction of adverse medical events is important considering the prevalence of vast repositories of patient-level medical data, and the quest to reduce healthcare costs while elevating the quality of care. Of particular interest is identifying a set of features that can better predict the risk of unplanned readmissions of chronically ill patients within 30 days post discharge, among them patients with Congestive Heart Failure (CHF). While numerous studies have employed ML algorithms to identify sets of significant features, only a handful have compared features extracted by various methods, and even fewer have compared the predictive power of feature sets selected by human experts to those selected by ML. Hence, this research aimed to elicit a comprehensive feature set selected by four ML algorithms and compare its predictive performance to that of a feature set selected by experts. We then evaluated a merged list of the two feature sets, constructing a human-machine collaborative set to predict the likelihood of 30-day unplanned readmission of CHF patients using data on 10,763 CHF patients hospitalized during the years 2010–2017. Our models achieved Area under the ROC Curve (AUC) above 0.8 and an Accuracy of ~0.89, comparable to the best models in the extant literature. Moreover, the Human-Machine Collaborative model significantly outperformed (best AUC 0.836) other human-only selected models as well as the machine-selected set. This study contributes to a better understanding of the power of using ML for patient risk stratification with special attention to the benefits of human-machine collaboration, even when these two entities work separately or in parallel, for improved clinical decision making. Collapse

Topics: healthcare data decision making big data missing data data warehouse

Methods: machine learning computational algorithm XGBoost descriptive statistic random forest classification

Roadside Unit-based Unknown Object Detection in Adverse Weather Conditions for Smart Internet of Vehicles

2023 | ACM Transactions on Management Information Systems | Citations: 0

Authors: Chen, Yu-Chia; Jhong, Sin-Ye; Hsia, Chih-Hsien

Abstract: For Internet of Vehicles applications, reliable autonomous driving systems usual ... Expand

Abstract: For Internet of Vehicles applications, reliable autonomous driving systems usually perform the majority of their computations on the cloud due to the limited computing power of edge devices. The communication delay between cloud platforms and edge devices, however, can cause dangerous consequences, particularly for latency-sensitive object detection tasks. Object detection tasks are also vulnerable to significantly degraded model performance caused by unknown objects, which creates unsafe driving conditions. To address these problems, this study develops an orchestrated system that allows real-time object detection and incrementally learns unknown objects in a complex and dynamic environment. A you-only-look-once–based object detection model in edge computing mode uses thermal images to detect objects accurately in poor lighting conditions. In addition, an attention mechanism improves the system’s performance without significantly increasing model complexity. An unknown object detector automatically classifies and labels unknown objects without direct supervision on edge devices, while a roadside unit (RSU)-based mechanism is developed to update classes and ensure a secure driving experience for autonomous vehicles. Moreover, the interactions between edge devices, RSU servers, and the cloud are designed to allow efficient collaboration. The experimental results indicate that the proposed system learns uncategorized objects dynamically and detects instances accurately. Collapse

Topics: cloud computing autonomous driving system database system internet technology internet of things

Methods: computational algorithm cluster analysis design artifact experiment artificial neural network

A Reference Architecture for a Workflow Management System Front End Designed for Augmented Reality Headsets

2023 | European Conference On Information Systems | Citations: 0

Authors: Damarowsky, Johannes; Kuehnel, Stephan; Böhmer, Martin; Sackmann, Stefan

Abstract: A well-known approach to managing and controlling workflows in organizations is ... Expand

Abstract: A well-known approach to managing and controlling workflows in organizations is the workflow management system (WFMS). Recently, approaches utilizing augmented reality headsets as WFMS front ends have been discussed, enabling higher efficiency, effectiveness, and usability for certain application scenarios. However, existing design-oriented approaches lack tangible guidance for implementation. A well-known approach to address such knowledge gaps is a reference architecture, which inter alia reduces development times and risks and facilitates collaboration between developers. Based on an existing tentative design theory for an augmented reality-based WFMS front end, we contribute a reference architecture containing an extended design theory, user interface design, and UML models for use cases, components, classes, and sequence flows. The reference architecture was successfully operationalized in a prototype and positively evaluated via a survey of respective users. Collapse

Topics: usability user interface augmented reality smart device enterprise architecture

Methods: design science design artifact unified modeling language business process model and notation modeling language

Theories: unified theory of acceptance and use of technology

Causal Relationship and Ranking Technique (CRRT): A Novel Group Decision-Making Model and Application in Students’ Performance Assessment in Indian High School Context

2023 | Group Decision and Negotiation | Citations: 0

Authors: Dhara, Suvojit; Goswami, Adrijit

Abstract: In the field of multi-attribute group decision-making (MAGDM), “decision-making ... Expand

Abstract: In the field of multi-attribute group decision-making (MAGDM), “decision-making trial and evaluation laboratory” (DEMATEL) is a popular method that has been deployed to discover hidden relationships among many inter-twined factors of numerous real-world problems. This paper discusses several limitations of the DEMATEL method in mining the causal relationships and ranking the factors. Also, in this paper, we develop a novel decision-making framework that will be able to overcome the issues found in the DEMATEL method and tackle both formal and informal decisions. This model, which we call “Causal Relationship and Ranking Technique (CRRT)”, uses a data-deterministic expert weighing setup rather than assigning equal weights or pre-defined weights to each expert. The model proposes a better technique to mine factor inter-relations and proposes two new types of digraphs portraying the causal inter-relationships among the factors, and also proposes a new two-step factor ranking method based on the derived causal relationships. The intuitionistic fuzzy setting is considered in this work in order to capture the uncertainty of expert opinions. Furthermore, our proposed framework is applied to a real-life case study on discovering various factors of a high school administration, that have a significant influence on students’ performance. Comparison with standard methods such as TOPSIS and VIKOR for student performance data and also some other datasets such as supplier selection data, and sustainability data proves that the model is able to rank the factors better than the existing models. Collapse

Topics: decision making

Methods: case study TOPSIS trade-off analysis method cross sectional research descriptive statistic

FINFLUENCERS: OPINION MAKERS OR OPINION FOLLOWERS?

2023 | European Conference On Information Systems | Citations: 0

Authors: Haase, Frederic; Rath, Oliver; Kurka, Marco; Schoder, Detlef

Abstract: This paper explores the concept of Finfluencers: financial social network actor ... Expand

Abstract: This paper explores the concept of Finfluencers: financial social network actors with high potential social influence. Our research aims to clarify whether Finfluencers drive or are influenced by the broader social network sentiment, thereby establishing their role as either opinion makers or opinion followers. Using a dataset of 71 million tweets focusing on stocks and cryptocurrencies, we grouped actors by their social networking potential (SNP). Next, we derived sentiment time series using state-ofthe-art sentiment models and applied the technique of Granger causality. Our findings suggest that the sentiment of Finfluencer actors on Twitter has short-term predictive power for the sentiment of the larger group of actors. We found stronger support for cryptocurrencies in comparison to stocks. From the perspective of financial market regulation, this study emphasizes the relevance of understanding sentiment on social networks and high social influence actors to anticipate scams and fraud. Collapse

Topics: Twitter social network social influence cryptocurrency social media

Methods: longitudinal research time series analysis Granger causality descriptive statistic sentiment analysis

Co-evolution of neural architectures and features for stock market forecasting: A multi-objective decision perspective

2023 | Decision Support Systems | Citations: 0

Authors: Hafiz, Faizal; Broekaert, Jan; La Torre, Davide; Swain, Akshya

Abstract: In a multi-objective setting, a portfolio manager’s highly consequential decisio ... Expand

Abstract: In a multi-objective setting, a portfolio manager’s highly consequential decisions can benefit from assessing alternative forecasting models of stock index movement. The present investigation proposes a new approach to identify a set of non-dominated neural network models for further selection by the decision-maker. A new coevolution approach is proposed to simultaneously select the features and topology of neural networks (collectively referred to as neural architecture), where the features are viewed from a topological perspective as input neurons. Further, the co-evolution is posed as a multi-criteria problem to evolve sparse and efficacious neural architectures. The well-known dominance and decomposition based multi-objective evolutionary algorithms are augmented with a non-geometric crossover operator to diversify and balance the search for neural architectures across conflicting criteria. Moreover, the co-evolution is augmented to accommodate the data-based implications of distinct market behaviors prior to and during the ongoing COVID-19 pandemic. A detailed comparative evaluation is carried out with the conventional sequential approach of feature selection followed by neural topology design, as well as a scalarized co-evolution approach. The results on three market indices (NASDAQ, NYSE, and S&P500) in pre- and peri-COVID time windows convincingly demonstrate that the proposed co-evolution approach can evolve a set of non-dominated neural forecasting models with better generalization capabilities. Collapse

Topics: covid 19 pandemic decision support system decision making

Methods: computational algorithm artificial neural network data transformation machine learning principal component analysis

Theories: co-evolution theory

FROM EMERGENCY REMOTE ASSESSMENT TO A NEW STATUS QUO? – LESSONS LEARNED FROM ONLINE ASSESSMENTS DURING THE COVID-19 PANDEMIC

2023 | European Conference On Information Systems | Citations: 0

Authors: Hartmann, Philipp; Hobert, Sebastian

Abstract: Before the COVID-19 pandemic, studying at German universities was traditionally ... Expand

Abstract: Before the COVID-19 pandemic, studying at German universities was traditionally dominated by inclass activities. During the pandemic, technical and organizational foundations were laid not only for teaching but also for exams. With the resumption of in-class events, the question is to what extent online exams will also be used in the future. Here, examiners and educational institutions can learn from the experience during the pandemic to increase the students' intention to participate in online exams in the future. To identify the problems encountered and derive lessons from them, a mixed-method approach was followed. For this purpose, 478 examinees were surveyed over 3 semesters using a quantitative questionnaire, followed by in-depth semi-structured follow-up interviews with 11 students. As a result, 5 recommendations for action could be derived, covering the topics of exam routine and familiarity, problem-solving, cheating, and teaching-examination arrangements. Collapse

Topics: computer based testing pandemic knowledge type knowledge sharing

Methods: survey descriptive statistic personal interview data transformation design principle

Resource Allocation for Heterogeneous Computing Tasks in Wirelessly Powered MEC-enabled IIOT Systems

2023 | ACM Transactions on Management Information Systems | Citations: 0

Authors: Hu, Yixiang; Deng, Xiaoheng; Zhu, Congxu; Chen, Xuechen; Chi, Laixin

Abstract: Integrating wireless power transfer with mobile edge computing (MEC) has become ... Expand

Abstract: Integrating wireless power transfer with mobile edge computing (MEC) has become a powerful solution for increasingly complicated and dynamic industrial Internet of Things (IIOT) systems. However, the traditional approaches overlooked the heterogeneity of the tasks and the dynamic arrival of energy in wirelessly powered MEC-enabled IIOT systems. In this article, we formulate the problem of maximizing the product of the computing rate and the task execution success rate for heterogeneous tasks. To manage energy harvesting adaptively and select appropriate computing modes, we devise an online resource allocation and computation offloading approach based on deep reinforcement learning. We decompose this approach into two stages: an offloading decision stage and a stopping decision stage. The purpose of the offloading decision stage is to select the computing mode and dynamically allocate the computation round length for each task after learning from the channel state information and the task experience. This stage allows the system to support heterogeneous computing tasks. Subsequently, in the second stage, we adaptively adjust the number of fading slots devoted to energy harvesting in each round in accordance with the status of each fading slot. Simulation results show that our proposed algorithm can better allocate resources for heterogeneous tasks and reduce the ratio of failed tasks and energy consumption when compared with several existing algorithms. Collapse

Topics: resource allocation edge computing decision quality

Methods: computational algorithm reinforcement learning simulation experiment PyTorch

Examining Experiences Concerning Goals Pursued and Smart Dimensions to Develop in Cities on their Path to Become Smart. The Case of Costa Rican Cities

2023 | HICSS | Citations: 0

Authors: Pereira Piedra, I.; Rodríguez Bolívar, M.; Alcaide Muñoz, L.

Abstract: The rapid growth of the population in cities has given rise to major urban chal ... Expand

Abstract: The rapid growth of the population in cities has given rise to major urban challenges. This has led city governments to use technical solutions based on Information and Communication Technologies (ICTs), arising the concept of smart city (SC). However, the strategy followed to become smart is still very diffuse. This paper seeks to analyze the opinion of the ICT managers of local governments in their initial stage of cities to become smart. To achieve this aim, we have conducted empirical research in Costa Rican cities, which are initiating the SC process, but emphasizing the technological implementation strategy they are using. Concretely, this paper analyzes variables such as the number of ICT projects, dimensions, goals and expected impacts as well as their association. Findings provide technological strategies that can be used and thus facilitate the implementation process of cities to become smart as well as the achievement of results. Collapse

Topics: IT manager IT project systems implementation decision making mobile application

Methods: survey descriptive statistic case study statistical power analysis

Leveraging Interactions for Stationary and Dynamic Financial Distress Prediction: A Spatio-Temporal Financial Graph Attention Network

2023 | International Conference on Information Systems | Citations: 0

Authors: Qi, Qi; Xing, Frank

Abstract: Financial distress prediction is a prominent research topic in information syst ... Expand

Abstract: Financial distress prediction is a prominent research topic in information systems, with two primary modelling categories: stationary and dynamic modelling. Recent stationary modelling works have leveraged company interactions to improve prediction performance, considering the heterogeneity of interactions while ignoring the dynamicity. However, few dynamic modelling works utilized interactions. To address the inconsistency and limitation of stationary and dynamic modelling works in leveraging interactions, we propose the Spatio-Temporal Financial Graph Attention Network with Metalearning (STFGAN-Meta). STFGAN-Meta leverages interactions' spatial heterogeneity via the Spatial Aggregation module and temporal dynamicity via the Temporal Aggregation module. STFGAN-Meta introduces the Meta-learning Optimization module to unify stationary and dynamic modelling. Our experimental evaluation demonstrates that leveraging dynamicity and heterogeneity of interactions outperforms leveraging dynamicity or heterogeneity alone. Meta-learning succeeds in providing a generalized approach between stationary and dynamic modelling. STFGAN-Meta can be a promising risk assessment and decision-making tool in the financial industry. Collapse

Topics: decision making database system accounting Python

Methods: meta learning experiment machine learning metamodel design science