An ensemble model for link prediction based on graph embedding

2022 | Decision Support Systems | Citations: 0

Authors: Chen, Yen-Liang; Hsiao, Chen-Hsin; Wu, Chia-Chi

Abstract: A network is a form of data representation and is widely used in many fields. Fo ... Expand

Abstract: A network is a form of data representation and is widely used in many fields. For example, in social networks, we regard nodes as individuals or groups, and the edges between nodes are called links, that is, the interaction between people. By analyzing the interaction of nodes, we can learn more about network relationships. The core idea of link prediction is to predict whether there is a new relationship between a pair of nodes or to discover hidden links in the network. Link prediction has been applied to many fields such as social networking, ecommerce, bioinformatics, and so on. In addition, many studies have used graph embedding for link prediction, which effectively preserves the network structure and converts node information into a low-dimensional vector space. In this research, we used three graph embedding approaches: matrix decomposition based methods, random walk based methods, and deep learning based methods. Since each method has its own advantages and disadvantages, we propose an ensemble model to combine these graph embeddings into a new representation of each node. Then, we designed a two-stage link prediction model based on a multi-classifier ensemble and took the new node representation as its input. Performance evaluation was conducted on multiple data sets. Exper­ imental results show that the integration of multiple embedding methods and multiple classifiers can signifi­ cantly improve the performance of link prediction. Collapse

Topics: electronic mail wiki social network Wikipedia

Methods: logistic regression machine learning deep learning support vector machine ensemble learning

Facing Big Data System Architecture Deployments: Towards an Automated Approach Using Container Technologies for Rapid Prototyping

2022 | HICSS | Citations: 0

Authors: Volk, Matthias; Staegemann, Daniel; Islam, Ashraful; Turowski, Klaus

Abstract: Within the last decade, big data became a promising trend for many application a ... Expand

Abstract: Within the last decade, big data became a promising trend for many application areas, offering immense potential and a competitive edge for various organizations. As the technical foundation for most of today´s data-intensive projects, not only corresponding infrastructures and facilities but also the appropriate knowledge is required. Currently, several projects and services exist that not only allow enterprises to utilize but also to deploy related technologies and systems. However, at the same time, the use of these is accompanied by various challenges that may result in huge monetary expenditures, a lack of modifiability, or the risk of vendor lock-ins. To overcome these shortcomings, in the contribution at hand, modern container and task automation technologies are used to wrap complex big data technologies into re-usable and portable resources. Those are subsequently incorporated in a framework to automate the deployment of big data architectures in private and limited resources. Collapse

Topics: Cassandra Docker big data hard disk drive usability

Methods: data modeling literature study design artifact testing design science survey

Data Mining Meta-analysis Coding to Develop Smart Learning Systems that Dynamically Customize Scaffolding

2017 | Americas Conference on Information Systems | Citations: 0

Authors: Kim, Nam Ju; Belland, Brian; Kim, Yong Seog

Abstract: Scaffolding, defined as support to help students perform and gain skill at compl ... Expand

Abstract: Scaffolding, defined as support to help students perform and gain skill at complex tasks, has been integrated into instruction in a wide range of educational levels, disciplines, and problem-centered instructional approaches. In the underlying project, we aim to use machine learning to dynamically customize computerbased scaffolding. In this paper, we estimate the predictive power of machine learning classifiers with different types of input variables such as scaffolding characteristics, study characteristics, student characteristics, and study quality characteristics. Three classifiers (i.e., Naïve Bayes, support vector machines, and decision tree) can accurately predict with an averaged accuracy of over 50% the effectiveness of scaffolding using only scaffolding characteristics information. In addition, Decision Tree classifier leads to 91.34% accuracy when scaffolding characteristics and study characteristics are classified. Our findings from this preliminary experiment provide a foundation to set up initial parameters that a smart learning system can utilize to provide individualized and customized scaffolding. Collapse

Topics: online learning system data accuracy database system IS effectiveness

Methods: machine learning support vector machine sensitivity analysis experiment decision tree classification

QUO VADIS REAL-TIME BUSINESS INTELLIGENCE? A DESCRIPTIVE LITERATURE REVIEW AND FUTURE DIRECTIONS

2016 | European Conference On Information Systems | Citations: 0

Authors: Nadj, Mario

Abstract: This paper presents a systematic review of real-time Business Intelligence (RTBI ... Expand

Abstract: This paper presents a systematic review of real-time Business Intelligence (RTBI) literature through a grounded theory lens in five RTBI research areas: Presentation (of information), Analytics, Data Integration, Processes, and Soft Factors. The main contribution of the paper is in the summary of metainformation, such as research methods and industry sectors, and in the synopsis of RTBI related research in these five research streams. Collapse

Topics: business intelligence data warehouse logistics management customer relationship management risk management

Methods: literature study grounded theory literature sample structured literature research mathematical model

Ontology-Based Mapping for Automated Document Management: A Concept-Based Technique for Word Mismatch and Ambiguity Problems in Document Clustering

2015 | ACM Transactions on Management Information Systems | Citations: 0

Authors: Lee, Yen-Hsien; Hu, Paul Jen-Hwa; Tu, Ching-Yi

Abstract: Document clustering is crucial to automated document management, especially for ... Expand

Abstract: Document clustering is crucial to automated document management, especially for the fast-growing volume of textual documents available digitally. Traditional lexicon-based approaches depend on document content analysis and measure overlap of the feature vectors representing different documents, which cannot effectively address word mismatch or ambiguity problems. Alternative query expansion and local context discovery approaches are developed but suffer from limited efficiency and effectiveness, because the large number of expanded terms create noise and increase the dimensionality and complexity of the overall feature space. Several techniques extend lexicon-based analysis by incorporating latent semantic indexing but produce less comprehensible clustering results and questionable performance. We instead propose a concept-based document representation and clustering (CDRC) technique and empirically examine its effectiveness using 433 articles concerning information systems and technology, randomly selected from a popular digital library. Our evaluation includes two widely used benchmark techniques and shows that CDRC outperforms them. Overall, our results reveal that clustering documents at an ontology-based, concept-based level is more effective than techniques using lexicon-based document features and can generate more comprehensible clustering results. Collapse

Topics: knowledge representation operating system Microsoft Windows Java information retrieval

Methods: ontological modelling cluster analysis experiment literature sample existing ontologies

Tracking&Tracing-systems in the Wood Supply Chain: Opportunities and Challenges

2013 | Americas Conference on Information Systems | Citations: 0

Authors: Appelhanz, Shanna

Abstract: Tracking&Tracing-systems have been used for years in the logistics and food indu ... Expand

Abstract: Tracking&Tracing-systems have been used for years in the logistics and food industry for the monitoring, control and management of processes. In general, the usage of systems includes applications such as finding the source of recalls, proof of product quality and origin, improving logistics and production processes. The benefits of these systems in the specific environment of the wood supply chain have not been explored to this point. Whether the data from Tracking&Tracingsystems could contribute to overcoming wood-specific challenges such as fulfillment of legislative requirements, certification of wood origin, minimizing illegal felling, support of sustainable forest management and thus the environmental sustainability is the main question of the investigation carried out in this paper. The investigation includes a comprehensive overview of possibilities and challenges of these systems in the wood supply chain. Collapse

Topics: supply chain management environmental sustainability database system

Methods: literature study literature sample

Simple decision forests for multi-relational classification

2013 | Decision Support Systems | Citations: 0

Authors: Bina, Bahareh; Schulte, Oliver; Crawford, Branden; Qian, Zhensong; Xiong, Yi

Abstract: An important task in multi-relational data mining is link-based classification w ... Expand

Abstract: An important task in multi-relational data mining is link-based classification which takes advantage of attributes of links and linked entities, to predict the class label. The relational Naive Bayes classifier exploits independence assumptions to achieve scalability. We introduce a weaker independence assumption to the effect that information from different data tables is independent given the class label. The independence assumption entails a closed-form formula for combining probabilistic predictions based on decision trees learned on different database tables. Logistic regression learns different weights for information from different tables and prunes irrelevant tables. In experiments, learning was very fast with competitive accuracy. Collapse

Topics: database system internet technology anonymity

Methods: decision tree classification logistic regression machine learning naïve bayes experiment

The Generative Mechanisms of Digital Infrastructure Evolution

2013 | Management Information Systems Quarterly | Citations: 609

Authors: Henfridsson, Ola; Bygstad, Bendik

Abstract: The current literature on digital infrastructure offers powerful lenses for conc ... Expand

Abstract: The current literature on digital infrastructure offers powerful lenses for conceptualizing the increasingly interconnected information system collectives found in contemporary organizations. However, little attention has been paid to the generative mechanisms of digital infrastructure, that is, the causal powers that explain how and why such infrastructure evolves over time. This is unfortunate, since more knowledge about what drives digital infrastructures would be highly valuable for managers and IT professionals confronted by the complexity of managing them. To this end, this paper adopts a critical realist view for developing a configurational perspective of infrastructure evolution. Our theorizing draws on a multimethod research design comprising an in-depth case study and a case survey. The in-depth case study, conducted at a Scandinavian airline, distinguishes three key mechanisms of digital infrastructure evolution: adoption, innovation, and scaling. The case survey research of 41 cases of digital infrastructure then identifies and analyzes causal paths through which configurations of these mechanisms lead to successful evolution outcomes. The study reported in this paper contributes to the infrastructure literature in two ways. First, we identify three generative mechanisms of digital infrastructure and how they contingently lead to evolution outcomes. Second, we use these mechanisms as a basis for developing a configurational perspective that advances current knowledge about why some digital infrastructures evolve successfully while others do not. In addition, the paper demonstrates and discusses the efficacy of critical realism as a philosophical tradition for developing substantive contributions in the field of information systems. Collapse

Topics: information technology infrastructure innovation management artificial intelligence database system digital ecosystem

Methods: survey case study literature study descriptive statistic theory development

Theories: theory of affordance

Style Composition in Action Research Publication

2012 | Management Information Systems Quarterly | Citations: 40

Authors: Mathiassen, Lars; Chiasson, Mike; Germonprez, Matt

Abstract: Examining action research publications in leading Information Systems journals a ... Expand

Abstract: Examining action research publications in leading Information Systems journals as a particular genre of research communication, we develop the notion of style composition to understand how authors structure their arguments for a research contribution. We define style composition as the activity through which authors select, emphasize, and present elements of their research to establish premises, develop inferences, and present contributions in publications. Drawing on this general notion, we identify a set of styles that is characteristic of how IS action researchers compose their argument. Premise styles relate to the dual goals of action research through practical or theoretical positioning of the argument; inference styles combine insights from the problem-solving and the research cycles through inductive or deductive reasoning; and contribution styles focus on different types of contributions-experience report, field study, theoretical development, problem solving method, and research method. Based on the considered sample, we analyze the styles adopted in selected publications and show that authors have favored certain styles while leaving others under explored;further, we reveal important strengths and weaknesses in the composition of styles within the IS discipline. Based on these insights, we discuss how action research practices and writing can be improved, as well as how to further develop style compositions to support the publication of engaged scholarship research Collapse

Topics: problem solving IS discipline systems development IT risk management brainstorming

Methods: action research grounded theory theory development design artifact literature study

Theories: structuration theory actor network theory media theory literary theory deterrence theory

Managing Changes in Collaborative Networks: A Conceptual Approach

2011 | International Conference on Information Systems | Citations: 4

Authors: Zarvic, Novica; Fellmann, Michael; Thomas, Oliver

Abstract: Collaborative Networks represent organizational forms that became omnipresent in ... Expand

Abstract: Collaborative Networks represent organizational forms that became omnipresent in today’s way of making business. Such organizational forms are often established in order to satisfy a complex customer need, which one company could not satisfy on its own. This means that the participating companies are to a certain degree dependent on each other. Managing inter-firm relationships by means of inter-organizational interdependencies represents an important Business-IT Alignment issue. In this paper, we present the Dependency-based Alignment Framework, which represents a conceptual approach for managing changes in Collaborative Networks from a holistic perspective. A detailed and methodologically well-founded approach in the definition and design of our framework is accompanied by a detailed investigation of relevant properties of this design artifact. To demonstrate the applicability of our framework in practice, we introduce a case study, which uses Semantic Media Wiki and the SPARQL query language. Finally, we evaluate our results in an argumentative and deductively descriptive way. Collapse

Topics: IT alignment wiki database languages system support problem solving

Methods: literature study design science design artifact case study conceptual modelling