Related references
Note: Only part of the references are listed.
Article
Public, Environmental & Occupational Health
Maria Oskarsdottir et al.
Summary: The study proposes a strategy to detect fraud by extracting information from the social network of claims, showing that models using network features perform well in fraud detection and contribute to a more effective fraud investigation process.
Article
Computer Science, Artificial Intelligence
Guangyuan Gao et al.
Summary: This paper introduces two data-driven neural network approaches for utilizing telematics car driving data to enhance classical actuarial regression models for claim frequency prediction. The study concludes that both classical actuarial risk factors and telematics car driving data are necessary to achieve the best predictive models, emphasizing the interaction and complementarity of these two sources of information.
Article
Computer Science, Artificial Intelligence
Arne De Brabandere et al.
Summary: This paper introduces an automated feature construction system called TSFuse, which supports fusion of sensor data and efficiently explores the search space. Through empirical evaluation on real-world time series classification datasets, the results show that TSFuse can find better feature representations compared to existing feature construction systems for univariate time series data.
Article
Computer Science, Artificial Intelligence
Bart Baesens et al.
Summary: Financial institutions are increasingly relying on data-driven methods to develop fraud detection systems, and have proposed several data processing techniques to improve the performance of analytical models while retaining interpretability.
DECISION SUPPORT SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Ozge Surer et al.
Summary: CTR is an algorithm that helps discover group structures to fit regression models, automating the creation of new features for improved predictive modeling in domains with unknown predictor relationships. It offers outstanding computational and predictive performance, as well as substantial interpretive advantages.
Article
Computer Science, Artificial Intelligence
Ekaterina Loginova et al.
Summary: This study contributes to the literature on directional cryptocurrency price returns prediction by expanding the set of meaningful features extracted from textual data with sentiment analysis and comparing their usefulness across multiple data sources. The use of fine-grained topic-sentiment features, such as aspect-based sentiment analysis models, leads to interpretable topics and an improvement in predictive performance. The dataset collected from sources like Reddit, Bitcointalk, and CryptoCompare demonstrates the effectiveness of the proposed features.
Article
Computer Science, Artificial Intelligence
Wenya Shi et al.
Summary: This study proposed two algorithms for solving high-dimension and large-sample dense and sparse trace-ratio problems, with theoretical results established to show the rationality and efficiency of the methods. Numerical experiments on real-world data sets demonstrated the superiority of the proposed algorithms over many state-of-the-art algorithms for high-dimension and large-sample dimensionality reduction problems.
Article
Computer Science, Artificial Intelligence
Vitor Cerqueira et al.
Summary: The study investigates extending auto-regressive processes using statistics and introduces a novel framework called VEST for automatic feature engineering of time series. The research finds that combining features generated by VEST with auto-regression significantly improves forecasting performance in a database of time series with high sampling frequency.
Article
Computer Science, Artificial Intelligence
Jakob Raymaekers et al.
Summary: This article proposes modifications to the Box-Cox and Yeo-Johnson transformations, using a robust estimator for the transformation parameter to make the transformed data approximately normal in the center while allowing for some outliers. It outperforms existing techniques in simulation studies and real data analysis.
Article
Computer Science, Artificial Intelligence
Emrah Hancer
Summary: Fuzzy mutual information is a popular method in information theory for quantifying the information between random variables, capable of handling different types of variables effectively. Recently, it has been integrated into evolutionary filter feature selection approaches to significantly improve the computational efficiency and performance of classification algorithms on real-world datasets.
Article
Computer Science, Artificial Intelligence
Yanou Ramon et al.
Summary: Machine learning models based on behavioral and textual data can be highly accurate but difficult to interpret. Rule-extraction techniques aim to combine predictive accuracy with global explainability, with higher-level metafeatures showing better fidelity in mimicking black-box prediction models.
Article
Computer Science, Theory & Methods
Jie Tao et al.
JOURNAL OF BIG DATA
(2020)
Article
Computer Science, Artificial Intelligence
Maria Oskarsdottir et al.
APPLIED SOFT COMPUTING
(2019)
Article
Computer Science, Artificial Intelligence
Maria Oskarsdottir et al.
EXPERT SYSTEMS WITH APPLICATIONS
(2017)
Article
Management
Veronique Van Vlasselaer et al.
MANAGEMENT SCIENCE
(2017)
Article
Information Science & Library Science
Joeran Beel et al.
INTERNATIONAL JOURNAL ON DIGITAL LIBRARIES
(2016)
Article
Computer Science, Artificial Intelligence
Veronique Van Vlasselaer et al.
DECISION SUPPORT SYSTEMS
(2015)
Article
Computer Science, Artificial Intelligence
Wouter Verbeke et al.
APPLIED SOFT COMPUTING
(2014)
Article
Management
B Baesens et al.
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH
(2002)
