How Tourist Group Books Hotels Meeting the Majority Affective Expectations: A Group Selection Frame with Kansei Text Mining and Consensus Coordinating

2023 | Group Decision and Negotiation | Citations: 0

Authors: Chang, Jia-Li; Li, Hui; Wu, Jian

Abstract: The decision on which hotel to book is a high-cost process for individuals and e ... Expand

Abstract: The decision on which hotel to book is a high-cost process for individuals and even more for groups due to the information explosion on online travel agent platforms. To assist group selecting the optimal hotel that meets the majority affective expectations, this study proposes a group hotel selection frame for booking with Kansei text mining and consensus coordinating. Firstly, to quantify the tourists’ affective needs for hotels, a Kansei-related dictionary towards hotel domain is constructed by Kansei text mining. Secondly, the bilattice-based Kansei score of hotels is defined and measured to represent hotel affective information scientifically. Then, a group consensus model with minimum adjustment cost is introduced to coordinate consistency of individual preferences in a tourist group. Moreover, a group hotel ranking based on group affective preference is interactively generated to meet the group personalized demands. Finally, a case study on Trip.com is conducted to demonstrate the proposed method. Collapse

Topics: knowledge representation online review personality website

Methods: natural language processing word2vec computational algorithm sentiment analysis case study

Global reconstruction of language models with linguistic rules – Explainable AI for online consumer reviews

2022 | Electronic Markets | Citations: 0

Authors: Binder, Markus; Heinrich, Bernd; Hopf, Marcus; Schiller, Alexander

Abstract: Analyzing textual data by means of AI models has been recognized as highly relev ... Expand

Abstract: Analyzing textual data by means of AI models has been recognized as highly relevant in information systems research and practice, since a vast amount of data on eCommerce platforms, review portals or social media is given in textual form. Here, language models such as BERT, which are deep learning AI models, constitute a breakthrough and achieve leading-edge results in many applications of text analytics such as sentiment analysis in online consumer reviews. However, these language models are “black boxes”: It is unclear how they arrive at their predictions. Yet, applications of language models, for instance, in eCommerce require checks and justifications by means of global reconstruction of their predictions, since the decisions based thereon can have large impacts or are even mandatory due to regulations such as the GDPR. To this end, we propose a novel XAI approach for global reconstructions of language model predictions for token-level classifications (e.g., aspect term detection) by means of linguistic rules based on NLP building blocks (e.g., part-of-speech). The approach is analyzed on different datasets of online consumer reviews and NLP tasks. Since our approach allows for different setups, we further are the first to analyze the trade-off between comprehensibility and fidelity of global reconstructions of language model predictions. With respect to this trade-off, we find that our approach indeed allows for balanced setups for global reconstructions of BERT’s predictions. Thus, our approach paves the way for a thorough understanding of language model predictions in text analytics. In practice, our approach can assist businesses in their decision-making and supports compliance with regulatory requirements. Collapse

Topics: explainable artificial intelligence laptop computer experience good electronic market artificial intelligence

Methods: BERT language model natural language processing machine learning product review

Analyzing e-government design science artifacts: A systematic literature review

2022 | International Journal of Information Management | Citations: 0

Authors: Carter, Lemuria; Yoon, Victoria; Liu, Dapeng

Abstract: Design science as a research paradigm is gaining popularity in the information s ... Expand

Abstract: Design science as a research paradigm is gaining popularity in the information systems (IS) discipline. E-gov­ ernment research explores IS artifacts designed to improve the quality and efficiency of public administration and service. This paper utilizes the E-government Design Research Model (EgovDR Model) to review e-gov­ ernment designs. Through a comprehensive literature review, this paper identifies prototypical e-government tasks for design science implementation and the corresponding solutions. We demonstrate whether and how the development and evolution of e-government designs continue to gain relevance in design science research. Additionally, this paper identifies and analyzes the theories employed in the literature to illustrate how the EgovDR Model has facilitated e-government designs in increasing rigor over time. Our findings indicate the majority of the workpractice tasks in existing e-government designs are decision-support tasks. Collapse

Topics: government system criminality decision support system social identity knowledge representation

Methods: design science literature study design artifact literature sample synthesis

Theories: information theory activity theory social disorganization theory portfolio theory structuration theory

An Ontology of Emotion Process to Support Sentiment Analysis

2022 | Journal of the Association for Information Systems | Citations: 1

Authors: Storey, Veda; Park, Eun Hee

Abstract: Sentiment analysis is used to mine text data from many sources, including blogs ... Expand

Abstract: Sentiment analysis is used to mine text data from many sources, including blogs, support forums, and social media, in order to extract customers' opinions and attitudes. The results can be used to make important assessments about a customer's attitude toward a company and if and how a company should respond. However, much research on sentiment analysis uses simple classification, where the polarity of a text that is mined is classified as positive, negative, or neutral. This research creates an ontology of emotion process to support sentiment analysis, with an emphasis on obtaining a more fine-grained assessment of sentiment than polarity. The ontology is grounded in a theory of emotion process and consists of concepts that capture the generation of emotion all the way from the occurrence of an event to the resulting behaviors of the person expressing the sentiment. It includes two lexicons: one for affect and one for appraisal. The ontology is applied to posts obtained from customer support forums of large companies to show its applicability in a multilevel evaluation. Doing so provides an example of a complete ontology assessment effort. Collapse

Topics: knowledge representation online review Bing evaluation criteria website

Methods: ontological modelling design artifact sentiment analysis experiment aspect-based sentiment analysis

Universal Spam Detection using Transfer Learning of BERT Model

2022 | HICSS | Citations: 0

Authors: Tida, Vijay Srinivas; Hsu, Sonya Hy

Abstract: Deep learning transformer models become important by training on text data based ... Expand

Abstract: Deep learning transformer models become important by training on text data based on self-attention mechanisms. This manuscript demonstrated a novel universal spam detection model using pre-trained Google's Bidirectional Encoder Representations from Transformers (BERT) base uncased models with four datasets by efficiently classifying ham or spam emails in real-time scenarios. Collapse

Topics: email spam Google+ electronic mail fake news Wikipedia

Methods: machine learning BERT computational algorithm transformer model deep learning

Motivating participation in crowdsourcing contests: The role of instruction-writing strategy

2022 | Information & Management | Citations: 0

Authors: Yin, Xicheng; Zhu, Kevin; Wang, Hongwei; Zhang, Jiaping; Wang, Wei; Zhang, Heng

Abstract: Task instructions are seeker-generated content aiming to disclose information an ... Expand

Abstract: Task instructions are seeker-generated content aiming to disclose information and persuade solvers to participate in crowdsourcing contests. There are various writing strategies for task instructions, with different levels of informativeness and persuasiveness for solvers. This work examines how requirement-oriented and rewardoriented strategies affect solver participation. The empirical results reveal a U-shaped relationship between the usage of restrictive words and the number of participants. In contrast, an inverted U-shaped relationship is found between terminology usage and participation. Emphasizing intrinsic and extrinsic rewards through issuerelevant information statements and emotional appeals also proves to be effective in motivating solvers’ involvement. Collapse

Topics: crowdsourcing Python knowledge representation task complexity database system

Methods: natural language processing machine learning word2vec statistical hypothesis test descriptive statistic

Theories: expectancy theory of motivation dual-system theory

A novel approach for fraudulent reviewer detection based on weighted topic modelling and nearest neighbors with asymmetric Kullback–Leibler divergence

2022 | Decision Support Systems | Citations: 0

Authors: Zhang, Wen; Xie, Rui; Wang, Qiang; Yang, Ye; Li, Jian

Abstract: The task of detecting fraudulent reviewers is of great importance to E-commerce ... Expand

Abstract: The task of detecting fraudulent reviewers is of great importance to E-commerce platforms. Existing research has invested much effort into developing comprehensive features and advanced techniques to detect fraudulent reviewers. However, most of these studies have ignored the data imbalance problem inherent in fraudulent reviewer detection: non-fraudulent reviewers are the majority, while fraudulent reviewers are the minority in real practice. To fill this gap, we propose a novel approach called ImDetector to detect fraudulent reviewers while handling data imbalance based on weighted latent Dirichlet allocation (LDA) and Kullback–Leibler (KL) divergence. Specifically, we develop a weighted LDA model to extract the latent topics of reviewers distributed on the review features. Asymmetric KL divergence is adopted to make the similarity measure between reviewers biased toward the fraudulent minority when using the K-nearest-neighbor for classification. By mapping the reviewers to the latent topics of features derived from the weighted LDA model and measuring the similarities between reviewers using asymmetric KL divergence, the data imbalance problem in fraudulent reviewer detection is alleviated. Extensive experiments on the Yelp.com dataset demonstrate that the proposed ImDetector approach is superior to the state-of-the-art techniques used for fraudulent reviewer detection. We also explain the experimental results and present the managerial implications of this paper. Collapse

Topics: online review fraud detection knowledge representation network effect website

Methods: machine learning experiment nonparametric test Mann Whitney U test supervised learning

Ontology-Based Approach for the Measurement of Privacy Disclosure

2022 | Information Systems Frontiers | Citations: 0

Authors: Zhu, Nafei; Chen, Baocun; Wang, Siyu; Teng, Da; He, Jingsha

Abstract: Privacy protection has received a lot of attention in recent years since in the ... Expand

Abstract: Privacy protection has received a lot of attention in recent years since in the era of big data, abundant information about individuals can be easily acquired. Meanwhile, as a prerequisite for effective privacy protection, the measurement of privacy disclosure is essential. Although some work has been done on the evaluation of privacy disclosure via quantification for the protection of privacy, not much attention has been placed on exploring the relationships between privacy information, resulting in underestimation, if not ill-formed reasoning, of privacy disclosure. In this paper, we propose an ontology-based approach to measure privacy disclosure by exploring the relationships between privacy information based on the WordNet. We first propose an algorithm for deriving or measuring privacy disclosure based on a set of words or concepts from text data related to individuals to ensure that the disclosure of certain user privacy can still be deduced and measured even if the set of words or concepts don’t seem to be much related to it. We then perform a set of experiment by applying the proposed algorithm to some public information of ten public figures from different walks of life to evaluate the effectiveness of the algorithm and to shed some light on the characteristics of privacy disclosure in the real world in the era of big data. The work can thus serve as the foundation for the development of mechanisms for limiting or reducing privacy disclosure to achieve better protection of individual privacy. Collapse

Topics: privacy knowledge representation search engine internet technology Java

Methods: computational algorithm experiment WordNet ontological modelling literature sample

An Enhanced Approach to Map Domain-Specific Words in Cross-Domain Sentiment Analysis

2021 | Information Systems Frontiers | Citations: 0

Authors: Geethapriya, A.; Valli, S.

Abstract: Domain adaptation in sentiment analysis is one of the areas where a classifier t ... Expand

Abstract: Domain adaptation in sentiment analysis is one of the areas where a classifier trained in one domain often classifies sentiments poorly when applied to another domain due to domain-specific words. Extracting features and their relevant opinion words from different domain sources and mapping them to the target domains are herculean tasks as far as domain adaptation is concerned. In this paper, the feature extraction technique is refined by which the mapping task is enhanced. The feature extraction technique uses both the syntactic and semantic properties of the features for extracting similar words. The features are further refined by merging synonyms and by replacing negative polarity terms with the appropriate antonyms. This refinement in the feature selection improves the mapping functionality of the domain adaptation and also exploits the relationship between domain-specific words and domain-independent words from different domains. Collapse

Topics: DVD enterprise architecture knowledge representation online review website

Methods: computational algorithm sentiment analysis experiment literature study product review

Social Network Positions, Peer Effects, and Evaluation Updating: An Experimental Test in the Entrepreneurial Context

2021 | Organization Science | Citations: 0

Authors: Greenberg, Jason

Abstract: In many facets of life, individuals make evaluations that they may update after ... Expand

Abstract: In many facets of life, individuals make evaluations that they may update after consulting with others in their networks. But not all individuals have the same positional opportunities for social interaction in a given network or the ability and desire to make use of those opportunities that are available to them. The configuration of a person’s network can also alter how information is spread or interpreted. To complicate matters further, scant research has considered how positions in social networks and the valence of network content interact because of the difficulty of (a) separating the “player” from the position in networks and (b) measuring all germane content in a particular network. This research develops a novel experimental platform that addresses these issues. Participants viewed and evaluated an entrepreneurial video pitch and were then randomly assigned to different networks, and positions within networks, and thus various opportunities for peer influence that were orthogonal to their network history, inclinations, attributes, or capabilities. Furthermore, all the content of social interaction, including its valence, was recorded to test underlying assumptions. Results reveal that those assigned to a position with brokerage opportunities in a network updated their evaluations of the entrepreneurial video considerably more negatively. Collapse

Topics: social influence relational database knowledge representation social network usability

Methods: experiment regression analysis method chi squared test instrumental variables estimation descriptive statistic

Theories: social network theory