Using Twitter Post Data to Ascertain the Sentiment of Alcohol-related Blackouts in the United States

2023 | HICSS | Citations: 0

Authors: Merrill, Jennifer; Riordan, Benjamin; Ward, Rose Marie; Raubenheimer, Jacques

Abstract: Research shows variability in how alcohol-related blackouts (periods of memory l ... Expand

Abstract: Research shows variability in how alcohol-related blackouts (periods of memory loss during/after drinking) are subjectively evaluated. We accessed 3.5 million original Tweets written in the U.S. between July 2009 and February 2020 that referenced blackouts, and coded the sentiment (positive or negative) of those Tweets, using the machine learning function of a Twitter-sponsored commercial platform. The sentiment of Tweets was examined by day of week and compared to the sentiment of blackout Tweets on certain holidays to non-celebration matched days. Tweets were more likely to have a positive (73%) than negative sentiment, and positive Tweets were more common during weekends. Relative to typical non-celebratory weekends, a greater proportion of blackout Tweets were positive around Thanksgiving and New Year’s Eve, though differences were not observed relative to several other celebratory periods (e.g., Superbowl). Results have implications for online interventions, which can use social networking sites to target alcohol during high-risk periods. Collapse

Topics: Twitter online learning system application programming interface social media social network

Methods: statistical hypothesis test paired t-test computational algorithm machine learning machine learning process

Reuse, Reduce, Support: Design Principles for Green Data Mining

2023 | Business & Information Systems Engineering | Citations: 1

Authors: Schneider, Johannes; Seidel, Stefan; Basalla, Marcus; vom Brocke, Jan

Abstract: This paper reports on a design science research (DSR) study that develops design ... Expand

Abstract: This paper reports on a design science research (DSR) study that develops design principles for “green” – more environmentally sustainable – data mining processes. Grounded in the Cross Industry Standard Process for Data Mining (CRISP-DM) and on a review of relevant literature on data mining methods, Green IT, and Green IS, the study identifies eight design principles that fall into the three categories of reuse, reduce, and support. The paper develops an evaluation strategy and provides empirical evidence for the principles’ utility. It suggests that the results can inform the development of a more general approach towards Green Data Science and provide a suitable lens to study sustainable computing. Collapse

Topics: green IS environmental sustainability data quality computer hardware database system

Methods: machine learning design principle Cross-Industry Standard Process for Data Mining computational algorithm machine learning process

Data-driven decision-making in credit risk management: The information value of analyst reports

2022 | Decision Support Systems | Citations: 0

Authors: Roeder, Jan; Palmer, Matthias; Muntermann, Jan

Abstract: Other than banks, non-financial companies also continuously monitor and analyze ... Expand

Abstract: Other than banks, non-financial companies also continuously monitor and analyze their credit risk exposure to avoid possible counterparty defaults. Credit default swaps are commonly used financial instruments that provide information on a counterparty’s creditworthiness. Although this metric can provide crucial insights, the un­ derlying price dynamics often remain unknown and require further explanation. Data-driven decision-making is a key concept for identifying these reasons and supporting and justifying decisions. In this paper, we provide such justifications by applying sentiment and topic analysis to company-related financial analyst reports. While the contents of financial news have been analyzed in the past, analyst reports can offer additional insights, as seasoned analysts use them to disseminate in-depth research to experienced investors. This analysis examines 3386 analyst reports covering constituents of the Dow Jones Industrial Average Index in the period from 2009 to 2020. The results suggest that even when established credit risk indicators and financial news are considered, the sentiment and a subset of topics are correlated with changes in the credit default swap spread, indicating a fundamental relationship between quantitative risk metric and analyst reports. We find that analyst reports contain information related to the change in credit default swap spreads, an insight that helps to improve our understanding of existing risk assessments. The outcome indicates that banks or corporate risk managers can benefit from complementing established financial metrics and even financial news data with new insights derived from analyst reports. Collapse

Topics: accounting risk management decision making mergers and acquisition Python

Methods: regression analysis method topic model machine learning longitudinal research regression diagnostic

Explainability and fairness of RegTech for regulatory enforcement: Automated monitoring of consumer complaints

2022 | Decision Support Systems | Citations: 0

Authors: Siering, Michael

Abstract: The application of regulatory technology (RegTech) for monitoring comprehensive ... Expand

Abstract: The application of regulatory technology (RegTech) for monitoring comprehensive data sources has gained increased importance. Nevertheless, previous research neglects that the output of RegTech applications has to be explainable and non-discriminatory. Within this study, we propose design principles and features for a RegTech approach which provides automated assessments of financial consumer complaints. We follow three main design principles. First, we build upon information diagnosticity theory to address the need for explainable classifica­ tions empowering regulators to justify their actions. Second, we consider a bag-of-words representation and ensemble learning to ensure high classification accuracy. Third, we take into account author characteristics to avoid discriminating classifications. We evaluate our approach in the financial services industry and show its value for identifying consumer complaints resulting in monetary compensations. The proposed design principles and features are highly relevant for regulators, corporations as well as consumers. Collapse

Topics: online review database system user-generated content experience good information overload

Methods: design principle machine learning design requirement McNemar's test kernel theory

Fair CRISP-DM: Embedding Fairness in Machine Learning (ML) Development Life Cycle

2022 | HICSS | Citations: 0

Authors: Singh, Vivek; Singh, Anshuman; Joshi, Kailash

Abstract: With the rapid adoption of machine learning (ML) technologies, organizations are ... Expand

Abstract: With the rapid adoption of machine learning (ML) technologies, organizations are constantly exploring efficient processes to develop such technologies. The cross-industry standard process for data mining (CRISP-DM) provides industry and technologyindependent model for organizing ML project development. However, the model lacks fairness concerns related to ML technologies. To address this significant theoretical and practical gap in the literature, we propose a new model – Fair CRISP-DM, which groups and presents fairness concerns relevant to each phase of an ML project development. We contribute to the literature on ML development and fairness. Specifically, ML researchers and practitioners can use our model to check and mitigate fairness concerns in each phase of an ML project development. Collapse

Topics: artificial intelligence decision making societal challenge

Methods: Cross-Industry Standard Process for Data Mining machine learning literature study machine learning process

Organizational Learning for Intelligence Amplification Adoption: Lessons from a Clinical Decision Support System Adoption Project

2022 | Information Systems Frontiers | Citations: 1

Authors: Wijnhoven, Fons

Abstract: Intelligence amplification exploits the opportunities of artificial intelligence ... Expand

Abstract: Intelligence amplification exploits the opportunities of artificial intelligence, which includes data analytic techniques and codified knowledge for increasing the intelligence of human decision makers. Intelligence amplification does not replace human decision makers but may help especially professionals in making complex decisions by well-designed human-AI system learning interactions (i.e., triple loop learning). To understand the adoption challenges of intelligence amplification systems, we analyse the adoption of clinical decision support systems (CDSS) as an organizational learning process by the case of a CDSS implementation for deciding on administering antibiotics to prematurely born babies. We identify user-oriented single and double loop learning processes, triple loop learning, and institutional deutero learning processes as organizational learning processes that must be realized for effective intelligence amplification adoption. We summarize these insights in a system dynamic model—containing knowledge stocks and their transformation processes—by which we analytically structure insights from the diverse studies of CDSS and intelligence amplification adoption and by which intelligence amplification projects are given an analytic theory for their design and management. From our case study, we find multiple challenges of deutero learning that influence the effectiveness of IA implementation learning as transforming tacit knowledge into explicit knowledge and explicit knowledge back to tacit knowledge. In a discussion of implications, we generate further research directions and discuss the generalization of our case findings to different organizations. Collapse

Topics: organization learning implicit knowledge artificial intelligence usage intention explicit knowledge

Methods: case study qualitative interview survey personal interview literature study

Theories: unified theory of acceptance and use of technology

Competitor identification: A review of use cases, data sources, and algorithms

2022 | International Journal of Information Management | Citations: 0

Authors: Werle, Marcel; Laumer, Sven

Abstract: Businesses have to deal with and survive in competitive, constantly changing lan ... Expand

Abstract: Businesses have to deal with and survive in competitive, constantly changing landscapes due to new entrants and exogenous shocks like the COVID-19 pandemic. Furthermore, decision-makers have to deal with staggering amounts of data produced in their corporate environment. Hence, research and practice have developed hybrid socio-technical systems of data mining and expert knowledge to identify competitors early on. To better un­ derstand the ergonomic design of such systems and how to integrate them in strategic management driven by artificial intelligence, we engaged in a systematic literature review to summarize the field and guide future research. We identified use cases, data source types, and algorithms that have been developed in 40 publications between 2002 and 2019. We found that existing approaches neglect to identify indirect and potential competitors at the periphery of a company’s vision. Such a blind spot exposes companies to increased risks of disruption, given that disruptive change often starts in the form of weak signals. Furthermore, we recommend using artificial intelligence to advance search strategies, allow document collection updates, and harmonize multiple data source types. Collapse

Topics: social media user-generated content Twitter decision making artificial intelligence

Methods: machine learning natural language processing literature study machine learning process grounded theory

A method for objective performance benchmarking of teams with process mining and DEA

2021 | European Conference On Information Systems | Citations: 0

Authors: Aysolmaz, Banu; Nemeth, Marcel; Iren, Deniz

Abstract: Process performance is influenced by the teams involved in those processes. How ... Expand

Abstract: Process performance is influenced by the teams involved in those processes. However, objectively evaluating team performance is challenging. Process mining literature has mostly focused on the characteristics of social networks rather than their performance. Following design science research principles, we designed a method for objectively benchmarking team performance using process mining, data envelopment analysis, and social network analysis. Our stepwise method provides guidelines on choosing appropriate data, extracting teams, identifying and calculating relevant metrics, benchmarking team performance, and determining improvement areas by considering the process context. We demonstrate the application on a real-life loan management process. We conducted interviews with business professionals to validate the usefulness and applicability. The results show that the method can provide an objective and transparent mechanism for comparing the performance of teams. The method can be used by organizations to form efficient teams and facilitate an increase in productivity. Collapse

Topics: team productivity process mining business process management productivity organizational value

Methods: event log qualitative interview design science machine learning process social network analysis

Achieving Success in Community Crowdsourcing: Lessons from the Field

2021 | HICSS | Citations: 0

Authors: De Vreede, Triparna; De Vreede, Gert-Jan; Alawi, Naif

Abstract: Community crowdsourcing is a relatively new phenomenon where local institutions, ... Expand

Abstract: Community crowdsourcing is a relatively new phenomenon where local institutions, such as cities and neighborhoods, invite citizens to engage in a public discussion and solve problems that directly affect them. While community crowdsourcing has been around for over a decade, relatively little is known about what drives the success of these initiatives. In this exploratory study, we analyze field data from over 1,000 community crowdsourcing projects that were hosted on a professional community crowdsourcing platform. Our exploration reveals interesting insights into characteristics of community crowdsourcing projects that are associated with higher levels of user engagement. These insights allow us to speculate on guidelines to organize and execute community crowdsourcing initiatives. Collapse

Topics: crowdsourcing evaluation criteria geographic information system analytical information system data integration

Methods: statistical hypothesis test multivariate analysis of variance descriptive statistic data transformation machine learning

Designing a collaboration process for big data analytics projects

2021 | European Conference On Information Systems | Citations: 0

Authors: Hagen, Janine

Abstract: Although there is broad agreement that big data analytics (BDA) creates busines ... Expand

Abstract: Although there is broad agreement that big data analytics (BDA) creates business value, knowledge about how organizations should proceed to extract value remains limited. Scholars have found that utilizing BDA insights in business greatly contributes to BDA value, but existing BDA processes tend to focus more on data science activities and do not adequately outline collaboration activities with business users. However, this relationship seems troubled and may even impede BDA success. To alleviate this discrepancy, we conduct an action research project with a tool manufacturer. The goal is to design a collaboration process for BDA. We apply social capital theory to evaluate the process, as social capital can improve collaboration. This paper gives the first glimpse of the collaboration process which intends to contribute to literature and practice by outlining critical collaboration activities that improve the relationship between business and data science and increase BDA use in business.conceptual iteration are directly applied to the content automation project to improve collaboration. The insights regarding the collaboration challenges and mitigation approaches from the diagnosis phase went directly into the first process concept. Principle of learning through reflection: We are aware of our dual responsibility as action researchers. Details on our expected practical and academic contribution can be found in section 5.Mar Collapse

Topics: social capital business process design organizational value IT skill marketing management

Methods: action research qualitative interview survey business process modeling machine learning process

Theories: social capital theory