The Cost of Fairness in AI: Evidence from E-Commerce

2022 | Business & Information Systems Engineering | Citations: 0

Authors: von Zahn, Moritz; Feuerriegel, Stefan; Kuehl, Niklas

Abstract: Contemporary information systems make widespread use of artificial intelligence ... Expand

Abstract: Contemporary information systems make widespread use of artificial intelligence (AI). While AI offers various benefits, it can also be subject to systematic errors, whereby people from certain groups (defined by gender, age, or other sensitive attributes) experience disparate outcomes. In many AI applications, disparate outcomes confront businesses and organizations with legal and reputational risks. To address these, technologies for so-called “AI fairness” have been developed, by which AI is adapted such that mathematical constraints for fairness are fulfilled. However, the financial costs of AI fairness are unclear. Therefore, the authors develop AI fairness for a real-world use case from e-commerce, where coupons are allocated according to clickstream sessions. In their setting, the authors find that AI fairness successfully manages to adhere to fairness requirements, while reducing the overall prediction performance only slightly. However, they find that AI fairness also results in an increase in financial cost. Thus, in this way the paper’s findings contribute to designing information systems on the basis of AI fairness. Collapse

Topics: artificial intelligence electronic commerce web analytics marketing management website

Methods: computational algorithm artificial neural network machine learning deep learning experiment

Theories: socio technical theory

A probabilistic Bayesian inference model to investigate injury severity in automobile crashes

2021 | Decision Support Systems | Citations: 2

Authors: Topuz, Kazim; Delen, Dursun

Abstract: Big data analytics examines millions, if not billions of records, to unmask hidd ... Expand

Abstract: Big data analytics examines millions, if not billions of records, to unmask hidden patterns, provide actionable insights and interpretable results for various domains. One area that has great potential to leverage the value of big data and analytics is the critical analysis of traffic accidents. Investigation results help in providing an indepth understanding of the risks and provide measures to potentially prevent these risk factors hence enhancing the well-being of individuals who may experience such accidents. This study explains existing models and proposes a data science methodology in a field where probabilistic modeling makes much sense for faster, better decision-making. The main objective of this data analytics study is to identify the high-risk factors with their apparent significance to influence the probability of injury severity on automobile crashes using a geographically representative car crash dataset. To obtain reliable, accurate, and intuitive results, a multi-step probabilistic inference model based on Bayesian Belief Network— highly-acclaimed machine learning method­ ology—is proposed. The underlying inference model provides researchers with a causally accurate way to explore the domain (with the subject matter expert inputs) while disengaging issues related to statistical cor­ relations and causal effects. In this study, we also used the data to create a web-based probabilistic inference simulator, a Bayesian inference decision support tool, which will be a publicly available/accessible tool, to help decision-makers better understand and to conduct what-if analysis on variable interdependencies. Collapse

Topics: missing data Python decision support system decision making database system

Methods: Bayesian analysis machine learning probabilistic graphical model simulation Bayesian network

Theories: information theory

Applying Multi-Criteria Decision-Making for the Selection of Big Data Technologies

2021 | Americas Conference on Information Systems | Citations: 0

Authors: Volk, Matthias; Staegemann, Daniel; Bischoff, Dennis; Turowski, Klaus

Abstract: For many years now big data presents itself as a technical starting point for a ... Expand

Abstract: For many years now big data presents itself as a technical starting point for a multitude of different trends, including data science, artificial intelligence, the internet of things, industry 4.0, and others. However, many decision makers are still challenged by the realization of related projects. One of the major problems arises from the multitude of existing technologies as well their selection and application. Due to that, in this paper, applying the design science research methodology, a comprehensive investigation about the application of multi-criteria decision-making (MCDM) methods is conducted. Apart from a structured literature review that seeks to identify relevant requirements for the technology selection, a conceptual model, as well as a prototypical implementation in the form of a decision support system, are presented. In the future, further MCDM approaches shall be tested and the system implemented in larger processes, including general sensemaking, system architecture modeling and rapid deployment. Collapse

Topics: big data decision support decision support system Apache Hadoop electronic authentication

Methods: design artifact structured literature research experiment design science multiple criteria decision making

Designing Decision-Support Systems for Secondary School Counselors

2020 | Americas Conference on Information Systems | Citations: 0

Authors: Crandall, Kodey; Noteboom, Cherie; Bishop, Dave

Abstract: In the current era of accountability, school counselors are expected to use dat ... Expand

Abstract: In the current era of accountability, school counselors are expected to use data to drive program decisionmaking, implement evidence-based interventions to create systemic change, and utilize emerging technology. Research shows it is difficult for school counselors to meet any of these expectations. A decision-support system is a technology that takes minimal effort to learn and can assist in decision-making processes. In an attempt to solve the problems facing school counselors, this design science research focuses on building such a system using the design principles of Information Quality, Interface Quality, and Customization and measuring their influence on user satisfaction. Collapse

Topics: decision support system user satisfaction usability computer system knowledge base

Methods: survey design principle qualitative coding design artifact field study

Theories: SERVQUAL

An Ontology and Multi-Agent Based Decision Support Framework for Prefabricated Component Supply Chain

2020 | Information Systems Frontiers | Citations: 0

Authors: Du, Juan; Jing, Hengqing; Choo, Kim-Kwang Raymond; Sugumaran, Vijayan; Castro-Lacouture, Daniel

Abstract: Due to industrialization and informatization of the construction industry, prefa ... Expand

Abstract: Due to industrialization and informatization of the construction industry, prefabricated construction has attracted wide attention from both research and practitioner communities. In prefabricated construction, there are exacting requirements for information sharing. Also, data in a prefabricated component supply chain tend to be dispersed in design, production, transportation and other stages. In other words, such data are significantly multi-source heterogeneous. Without an effective way of participating in supply chain dynamic collaboration, decision-making at various stages and resource allocation can be extremely challenging. This paper proposes a decision support framework for prefabricated component supply chain based on ontology and multi-agent. The framework comprises the ontology layer (i.e. provides data support for the model), the agent interaction layer (i.e. serves as the communication hub to coordinate the data transmission between modules), and the agent simulation layer (i.e. simulates interaction behavior of participants, and supports decision making). Using the Shanghai Chenxiang Road Station complex project as a case study, the paper demonstrates the validity of the proposed ontology and multi-agent based decision support framework. Collapse

Topics: knowledge representation supply chain management decision support resource allocation decision making

Methods: ontological modelling simulation qualitative interview design artifact parametric test

Conceptualizing smart service systems

2019 | Electronic Markets | Citations: 178

Authors: Beverungen, Daniel; Müller, Oliver; Matzner, Martin; Mendling, Jan; vom Brocke, Jan

Abstract: Recent years have seen the emergence of physical products that are digitally net ... Expand

Abstract: Recent years have seen the emergence of physical products that are digitally networked with other products and with information systems to enable complex business scenarios in manufacturing, mobility, or healthcare. These “smart products”, which enable the co-creation of “smart service” that is based on monitoring, optimization, remote control, and autonomous adaptation of products, profoundly transform service systems into what we call “smart service systems”. In a multi-method study that includes conceptual research and qualitative data from in-depth interviews, we conceptualize “smart service” and “smart service systems” based on using smart products as boundary objects that integrate service consumers’ and service providers’ resources and activities. Smart products allow both actors to retrieve and to analyze aggregated field evidence and to adapt service systems based on contextual data. We discuss the implications that the introduction of smart service systems have for foundational concepts of service science and conclude that smart service systems are characterized by technology-mediated, continuous, and routinized interactions. Collapse

Topics: smart device mobile system participatory design value creation boundary spanning

Methods: qualitative interview design methodology literature study personal interview qualitative coding

Theories: boundary objects theory

Health Improvement Path: Ontological Approach to Self-management Support in Personal Health Management Systems

2019 | HICSS | Citations: 0

Authors: Pak, Jinie; Song, Yeong-tae

Abstract: Ontologies have been used for knowledge modeling and reasoning in healthcare dom ... Expand

Abstract: Ontologies have been used for knowledge modeling and reasoning in healthcare domain (e.g., homecare, hospital clinical procedure, mHealth, etc.), but few in a context of self-management in healthcare with no sufficient reasoning rules to specify a systematic health management plan for an individual. In response to such needs, we aim to provide a generic ontology model for organizing the broad range of multidisciplinary knowledge required in personal health management by applying the ontology design patterns as well as for being extensible to more specific activity ontologies (e.g., physical exercises, diet, medication intake, etc.). The scope of a proposed ontology is to classify core concepts and relations in health self-management process and to build axioms for health improvement plans to meet an individual’s needs and health capability/maturity level. The proposed ontology is developed based on our previous work, health capability maturity model (HCMM) and can be integrated with existing health-related ontologies for further specification in health management processes. Collapse

Topics: knowledge representation business process management internet of things mobile system inference engine

Methods: ontological modelling design pattern capability maturity model design science

Overcoming individual process model matcher weaknesses using ensemble matching

2017 | Decision Support Systems | Citations: 2

Authors: Meilicke, Christian; Leopold, Henrik; Kuss, Elena; Stuckenschmidt, Heiner; Reijers, Hajo A.

Abstract: In recent years, a considerable number of process model matching techniques have ... Expand

Abstract: In recent years, a considerable number of process model matching techniques have been proposed. The goal of these techniques is to identify correspondences between the activities of two process models. However, the results from the Process Model Matching Contest 2015 reveal that there is still no universally applicable matching technique and that each technique has particular strengths and weaknesses. It is hard or even impossible to choose the best technique for a given matching problem. We propose to cope with this problem by running an ensemble of matching techniques and automatically selecting a subset of the generated correspondences. To this end, we propose a Markov Logic based optimization approach that automatically selects the best correspondences. The approach builds on an adaption of a voting technique from the domain of schema matching and combines it with process model specific constraints. Our experiments show that our approach is capable of generating results that are significantly better than alternative approaches. Collapse

Topics: knowledge representation web page business process management inference engine

Methods: experiment business process modeling machine learning computational algorithm optimization model

Theories: set theory

A hybrid decision support system for managing humanitarian relief chains

2017 | Decision Support Systems | Citations: 1

Authors: Sahebjamnia, Navid; Torabi, S. Ali; Mansouri, S. Afshin

Abstract: Decisions regarding location, allocation and distribution of relief items are am ... Expand

Abstract: Decisions regarding location, allocation and distribution of relief items are among the main concerns of the humanitarian relief chain (HRC) managers in response to no-notice large-scale disasters such as earthquakes. In this paper, a hybrid decision support system (HDSS) consisting of a simulator, a rule-based inference engine, and a knowledge-based system (KBS) is developed to configure a three level HRC. Three main performance measures including the coverage, total cost, and response time are considered to make an explicit trade-off analysis between the cost efficiency and responsiveness of the designed HRC. In the first step, the simulator calculates the performance measures of the different configurations of the HRC under a number of generated disaster scenarios. Then, the rule-based inference engine attempts to build the best configuration of the HRC including facilities' locations, relief items' allocation and distribution plan of the scenario under investigation based on the calculated performance measures. Finally, the best configuration for each scenario is stored in the KBS as the extracted knowledge from the above analyses. In this way, the HRC managers can retrieve the most appropriate HRC configuration in accordance with the realized post-disaster scenario in an effective and timely manner. The results of a real case study in Tehran demonstrate that the developed HDSS is an effective tool for fast configuration of HRCs using stochastic data. Collapse

Topics: earthquake decision support system knowledge base Microsoft Windows website

Methods: case study simulation literature study optimization model cross sectional research

Consumer decision support systems for novice buyers – a design science approach

2017 | Information Systems Frontiers | Citations: 0

Authors: Tang, Heng; Lee, Chang Boon Patrick; Choong, Kwee Keong

Abstract: For business-to-consumer e-commerce to have a significant impact, consumer decis ... Expand

Abstract: For business-to-consumer e-commerce to have a significant impact, consumer decision support systems (CDSS) must accommodate a wide variety of online buyers. The popular parameter-based CDSS (PCDSS) are limited by their inability to help novice consumers in their purchase of highly differentiated products. Drawing on Cognitive Fit Theory, we propose that a needs-based CDSS (NCDSS) that derives product selections based on user specified usage context is suitable for buyers without sufficient expertise in the product category. We adopt a design science paradigm to develop several approaches that serve this purpose. The effectiveness of our design is empirically evaluated through a user study in a simulated scenario. The results show that in general, the proposed needs-based decision supporting mechanisms significantly outperform the conventional PCDSS when applied to aid e-buyers to make purchase decisions. Moreover, they can also be utilized as the complement of the popular PCDSS in order to enable purchase decision support for online shoppers at various expertise levels. Collapse

Topics: decision support online word of mouth online review word of mouth inference engine

Methods: design artifact experiment computational algorithm sentiment analysis Bayesian network

Theories: cognitive fit theory