Article
Biochemical Research Methods
Marija Stanojevic, Jovan Andjelkovic, Adrienne Kasprowicz, Louise A. Huuki, Jennifer Chao, S. Blair Hedges, Sudhir Kumar, Zoran Obradovic
Summary: The TimeTree project has manually curated a TimeTree of Life from research articles containing timetrees, but the process has become costly and time-consuming. The researchers developed a text-mining approach and the TimeTreeFinder tool to automatically find articles with timetrees, potentially doubling the knowledge accessible to the scientific community.
Review
Oncology
Ming Kuang, Hang-Tong Hu, Wei Li, Shu-Ling Chen, Xiao-Zhou Lu
Summary: AI transforms medical images into mineable data, with machine learning algorithms being applied for lesion detection, segmentation, diagnosis, and prognosis. The number of articles utilizing AI, including in ultrasound, has significantly increased in recent years. Due to the unique properties of ultrasound, additional attention should be paid when assessing studies that rely on ultrasound AI.
FRONTIERS IN ONCOLOGY
(2021)
Article
Engineering, Civil
Mashrekur Rahman, Jonathan M. Frame, Jimmy Lin, Grey S. Nearing
Summary: The study reveals that individual water science and hydrology research articles are becoming more diverse in terms of topics, indicating an increase in interdisciplinary research. The topics that experienced the largest increases in popularity were Climate Change Impacts, Water Policy & Planning, and Pollutant Removal, while Stochastic Models and Numerical Models saw the largest decreases in popularity. In terms of journals, Water Resources Research, Journal of Hydrology, and Hydrological Processes are the most topically diverse among the studied corpus.
JOURNAL OF HYDROLOGY
(2022)
Article
Plant Sciences
Shriprabha R. Upadhyaya, Philipp E. Bayer, Cassandria G. Tay Fernandez, Jakob Petereit, Jacqueline Batley, Mohammed Bennamoun, Farid Boussaid, David Edwards
Summary: Gene model prediction is a complex process with potential false positive results. This study developed a machine learning approach using gene and protein-based characteristics to classify potential low confidence gene models. The optimized models showed high prediction accuracy and F-1 scores, which can be useful for supporting future gene annotation processes.
Article
Information Science & Library Science
Raji Raman, Ruba Aljafari, Viswanath Venkatesh, Vernon Richardson
Summary: This study examines the impact of textual content from business journals on cumulative abnormal returns, using sentiment analysis and machine learning techniques. The research finds that negative sentiments have a greater effect on cumulative abnormal returns compared to positive sentiments, and the effect of positive sentiments weakens when past quantitative measures are high. This study makes important contributions to the practice of sentiment analysis in financial markets as information sources continue to emerge on the web.
INTERNATIONAL JOURNAL OF INFORMATION MANAGEMENT
(2022)
Article
Multidisciplinary Sciences
Juhyeon Kim, Insung Ahn
Summary: The emergence of newly infectious diseases poses critical threats to human health and economy, emphasizing the importance of predicting and preparing for them. Due to the unpredictable and rapid spread of infectious diseases, detecting emerging patterns and building prediction models using relevant data is essential.
SCIENTIFIC REPORTS
(2021)
Article
Engineering, Electrical & Electronic
Dongqi Han, Zhiliang Wang, Ying Zhong, Wenqi Chen, Jiahai Yang, Shuqiang Lu, Xingang Shi, Xia Yin
Summary: This study systematically examines the use of ML/DL in NIDS, proposing a method for evaluating NIDS robustness through adversarial attacks in gray/black-box traffic space and offering a defense scheme. Experimental results demonstrate the effectiveness of the proposed attack and the ability of the defense method to mitigate such attacks.
IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS
(2021)
Article
Engineering, Civil
Mohsen Zaker Esteghamati, Thomas Gernay, Srishti Banerji
Summary: This study develops explainable data-driven models to predict the fire resistance of timber columns, and compares their predictive capabilities to available prescriptive equations. The results show that the random forest-based model provides the best performance, with accurate and balanced predictions. Column capacity and cross-section dimension are the main factors influencing fire resistance.
ENGINEERING STRUCTURES
(2023)
Article
Automation & Control Systems
Ara Carballo-Meilan, Lewis McDonald, Wanawan Pragot, Lukasz Michal Starnawski, Ali Nauman Saleemi, Waheed Afzal
Summary: This study employs information and data science algorithms to predict the outcome of a chemical engineering experiment. By using a systematic review and data mining method, a specific knowledge community is identified, and optimal synthesis conditions for producing the desired compound are determined.
CHEMOMETRICS AND INTELLIGENT LABORATORY SYSTEMS
(2022)
Article
Computer Science, Interdisciplinary Applications
Giovanna Nicora, Miguel Rios, Ameen Abu-Hanna, Riccardo Bellazzi
Summary: Interest in applying machine learning in clinical and biological problems is growing, but how to determine the reliability of predictions remains a challenge. This paper reviews methods to identify unreliable predictions and proposes an integrative framework for evaluating prediction reliability in specific scenarios.
JOURNAL OF BIOMEDICAL INFORMATICS
(2022)
Article
Multidisciplinary Sciences
Guiyu Wei, Ruliang Zhou
Summary: This study evaluated suitable areas for cultivating premium teas and found that certain regions in Yunnan are highly suitable for growing these high-quality teas. This is of significant importance for the development of the tea industry and the economic prosperity of minority regions.
Article
Biochemical Research Methods
Emilio Fenoy, Alejando A. Edera, Georgina Stegmayer
Summary: This article discusses the application of representation methods in bioinformatics, focusing on protein representation learning methods. It points out the lack of fair benchmark studies evaluating the predictive performance of existing proposals on large sets of proteins, which hinders the acceleration of protein functional characterization. Therefore, this study conducts comparative experiments to evaluate the performance of different protein sequence representation learning methods on several bioinformatics tasks.
BRIEFINGS IN BIOINFORMATICS
(2022)
Article
Business
Xiaoyu Ma, Yizhi Hao, Xiao Li, Jun Liu, Jiasen Qi
Summary: This study investigates national intelligence innovation through machine learning methods and proposes a global intelligence innovation index (GIII) to evaluate the global landscape of intelligence innovation. The study develops a conceptual framework of national intelligence innovation based on the innovation ecosystem theory and measures GIII using machine learning methods. The results show interesting relationships between intelligence innovation and factors such as unemployment, aging, and economic sectors. GIII provides a reference for intelligence innovation development and helps decision-makers formulate effective policies.
TECHNOLOGICAL FORECASTING AND SOCIAL CHANGE
(2023)
Article
Computer Science, Information Systems
Ghada Elkhawaga, Omar Elzeki, Mervat Abuelkheir, Manfred Reichert
Summary: Although machine learning predictions are highly accurate, understanding the underlying mechanisms and explaining the outcomes remain challenging. This paper proposes a technique to extract important features from a data perspective and introduces metrics to measure the ability of an explainability method to convey the underlying concepts. It also evaluates the capability of an eXplainable Artificial Intelligence (XAI) method to reason about the reliance of a Machine Learning (ML) model on the extracted features, providing a means to differentiate explainability methods.
Article
Clinical Neurology
Mathias Baumert, Simon Hartmann, Huy Phan
Summary: Using the deep neural network XSleepNet2, we trained and tested four separate sleep stage classifiers on polysomnograms from children, adults, and older adults. The underrepresentation of certain age groups, especially children, significantly affects the performance of automated sleep staging systems and limits their clinical use.