Results

Found 7 articles

Clear filters

Incorporating Multiple Knowledge Sources for Targeted Aspect-based Financial Sentiment Analysis
PDF
2023 | ACM Transactions on Management Information Systems | Citations: 0
Authors: Du, Kelvin; Xing, Frank; Cambria, Erik

Abstract: Combining symbolic and subsymbolic methods has become a promising strategy as re ... Expand

Abstract: Combining symbolic and subsymbolic methods has become a promising strategy as research tasks in AI grow increasingly complicated and require higher levels of understanding. Targeted Aspect-based Financial Sentiment Analysis (TABFSA) is an example of such complicated tasks, as it involves processes like information extraction, information specification, and domain adaptation. However, little is known about the design principles of such hybrid models leveraging external lexical knowledge. To fill this gap, we define anterior, parallel, and posterior knowledge integration and propose incorporating multiple lexical knowledge sources strategically into the fine-tuning process of pre-trained transformer models for TABFSA. Experiments on the Financial Opinion mining and Question Answering challenge (FiQA) Task 1 and SemEval 2017 Task 5 datasets show that the knowledge-enabled models systematically improve upon their plain deep learning counterparts, and some outperform state-of-the-art results reported in terms of aspect sentiment analysis error. We discover that parallel knowledge integration is the most effective and domain-specific lexical knowledge is more important according to our ablation analysis. Collapse

Semantic filters: XLNet

Topics: microblog domain-specific language Twitter social media knowledge graph

Methods: sentiment analysis RoBERTa transformer model BERT XLNet
A deep multi-embedding model for mobile application recommendation
PDF
2023 | Decision Support Systems | Citations: 0
Authors: Liu, Yi-Hung; Chen, Yen-Liang; Chang, Po-Ya

Abstract: With the explosion in the number of mobile applications, it becomes difficult an ... Expand

Abstract: With the explosion in the number of mobile applications, it becomes difficult and time-consuming for users to find the most suitable and interesting application from the millions of applications that exist today. Therefore, mobile app recommendation becomes a top priority to help people solve this problem. In previous studies, there are two common methods to solve this problem, one is to apply matrix factorization to the rating data to predict the target user's ratings for other applications, and the other is to predict ratings based on the user's review text. In addition to combining the above two traditional methods, this study also combines deep learning methods to generate better models. For rating data, we use a deep matrix factorization method to obtain latent factor vectors of users and items from the rating matrix to improve the embedding of ratings. For review text, we improve the text representation by using a multi-embedding approach. Most importantly, our model also incorporates deep neural networks with nonlinear transformation capabilities to extract more representative abstract features. We conduct experiments on datasets collected from the real world. Extensive experimental results show that our proposed method has better performance than other existing methods. Collapse

Semantic filters: XLNet

Topics: mobile application mobile application market recommender system digital distribution mobile system

Methods: BERT deep learning experiment sensitivity analysis artificial neural network
Improving Information Systems Sustainability by Applying Machine Learning to Detect and Reduce Data Waste
PDF
2023 | Communications of the Association for Information Systems | Citations: 0
Authors: Savarimuthu, Bastin Tony Roy; Corbett, Jacqueline; Yasir, Muhammad; Lakshmi, Vijaya

Abstract: Big data are key building blocks for creating information value. However, infor ... Expand

Abstract: Big data are key building blocks for creating information value. However, information systems are increasingly plagued with useless, waste data that can impede their effective use and threaten sustainability objectives. Using a constructive design science approach, this work first, defines digital data waste. Then, it develops an ensemble artifact comprising two components. The first component comprises 13 machine learning models for detecting data waste. Applying these to 35,576 online reviews in two domains reveals data waste of 1.9% for restaurant reviews compared to 35.8% for app reviews. Machine learning can accurately identify 83% to 99.8% of data waste; deep learning models are particularly promising, with accuracy ranging from 96.4% to 99.8%. The second component comprises a sustainability cost calculator to quantify the social, economic, and environmental benefits of reducing data waste. Eliminating 5948 useless reviews in the sample would result in saving 6.9 person hours, $2.93 in server, middleware and client costs, and 9.52 kg of carbon emissions. Extrapolating these results to reviews on the internet shows substantially greater savings. This work contributes to design knowledge relating to sustainable information systems by highlighting the new class of problem of data waste and by designing approaches for addressing this problem. Collapse

Semantic filters: XLNet

Topics: mobile application online review Python data management Google+

Methods: deep learning product review computational algorithm design artifact RoBERTa
What Do Customers Say About My Products? Benchmarking Machine Learning Models for Need Identification

2023 | HICSS | Citations: 0
Authors: Stahlmann, S.; Ettrich, O.; Kurka, M.; Schoder, D.

Abstract: Needmining is the process of extracting customer needs from user-generated cont ... Expand

Abstract: Needmining is the process of extracting customer needs from user-generated content by classifying it as either informative or uninformative regarding need content. Contemporary studies achieve this by utilizing machine learning. However, models found in the literature cannot be compared to each other because they use private data for training and testing. This study benchmarks all previously suggested needmining models including CNN, SVM, RNN, and RoBERTa. To ensure an unbiased comparison, this study samples and annotates a dataset of customer reviews for products from 4 different categories from amazon. Henceforth, the dataset is publicly available and serves as a gold-set for future needmining benchmarks. RoBERTa outperformed other classifiers and seems to be best suited for needmining. The relevance of this study is reinforced by the fact that this benchmark creates a different hierarchy between models than otherwise suggested by comparing the results of previous studies. Collapse

Semantic filters: XLNet

Topics: user-generated content Wikipedia Python online review Twitter

Methods: support vector machine RoBERTa naïve bayes machine learning XLNet
Fake Review Detection - The Value of Domain-Specificity

2023 | Americas Conference on Information Systems | Citations: 0
Authors: Theuerkauf, René; Peters, Ralf

Abstract: The issue of fake reviews, a perennial challenge faced by e-commerce companies, ... Expand

Abstract: The issue of fake reviews, a perennial challenge faced by e-commerce companies, has been worsened in recent years. In this paper, we investigate the trade-off between the application of general-purpose and domain-specific model by selecting training data using transformer-based models to identify fake reviews. Therefore, we compare two scenarios using different set ups of data selection. First, a general-purpose model was identified and applied on specific domains. Afterwards, domain-specific models were trained and tested. Then, the results from these scenarios were compared, yielding the conclusion that models trained on data from a similar product category outperform general-purpose models in classification performance up to 21% (on average by 4%). Our findings send an important message to e-commerce companies to rethink their strategy on training general-purpose models to identify fake reviews to create several, more domainspecific models on that task according to present data domains. Collapse

Semantic filters: XLNet

Topics: fake review online review Facebook electronic commerce recommender system

Methods: RoBERTa DeBERTa BERT transformer model Cross-Industry Standard Process for Data Mining
Development of an Automated Physician Review Classification System: A hybrid Machine Learning Approach

2022 | International Conference on Information Systems | Citations: 0
Authors: THIMMANAYAKANAPALYA, Sagarika Suresh; Mulgund, Pavankumar; Sharman, Raj

Abstract: ... Expand

Abstract: Collapse

Semantic filters: XLNet

Topics: online learning system artificial intelligence website online review user requirement

Methods: BERT machine learning design science computational algorithm artificial neural network
SymptomID: A Framework for Rapid Symptom Identification in Pandemics Using News Reports
PDF
2021 | ACM Transactions on Management Information Systems | Citations: 1
Authors: Gu, Kang; Vosoughi, Soroush; Prioleau, Temiloluwa

Abstract: The ability to quickly learn fundamentals about a new infectious disease, such a ... Expand

Abstract: The ability to quickly learn fundamentals about a new infectious disease, such as how it is transmitted, the incubation period, and related symptoms, is crucial in any novel pandemic. For instance, rapid identification of symptoms can enable interventions for dampening the spread of the disease. Traditionally, symptoms are learned from research publications associated with clinical studies. However, clinical studies are often slow and time intensive, and hence delays can have dire consequences in a rapidly spreading pandemic like we have seen with COVID-19. In this article, we introduce SymptomID, a modular artificial intelligence–based framework for rapid identification of symptoms associated with novel pandemics using publicly available news reports. SymptomID is built using the state-of-the-art natural language processing model (Bidirectional Encoder Representations for Transformers) to extract symptoms from publicly available news reports and cluster-related symptoms together to remove redundancy. Our proposed framework requires minimal training data, because it builds on a pre-trained language model. In this study, we present a case study of SymptomID using news articles about the current COVID-19 pandemic. Our COVID-19 symptom extraction module, trained on 225 articles, achieves an F1 score of over 0.8. SymptomID can correctly identify well-established symptoms (e.g., “fever” and “cough”) and less-prevalent symptoms (e.g., “rashes,” “hair loss,” “brain fog”) associated with the novel coronavirus. We believe this framework can be extended and easily adapted in future pandemics to quickly learn relevant insights that are fundamental for understanding and combating a new infectious disease. Collapse

Semantic filters: XLNet

Topics: pandemic artificial intelligence application programming interface

Methods: BERT cluster analysis named entity recognition literature sample case study

Page 1 of 1

AIS11

Journals

Conferences

Semantic Filter

Found 7 articles

Incorporating Multiple Knowledge Sources for Targeted Aspect-based Financial Sentiment Analysis

A deep multi-embedding model for mobile application recommendation

Improving Information Systems Sustainability by Applying Machine Learning to Detect and Reduce Data Waste

What Do Customers Say About My Products? Benchmarking Machine Learning Models for Need Identification

Fake Review Detection - The Value of Domain-Specificity

Development of an Automated Physician Review Classification System: A hybrid Machine Learning Approach

SymptomID: A Framework for Rapid Symptom Identification in Pandemics Using News Reports