Article
Computer Science, Artificial Intelligence
Farnaz Mahan, Maryam Mohammadzad, Seyyed Meysam Rozekhani, Witold Pedrycz
Summary: Several classification algorithms using fuzzy decision trees have been proposed in recent years, with a key issue being the selection of the division feature. While traditional algorithms commonly use the information gain criterion, it has been pointed out that this can lead to an unfair selection of features with a large number of values. The paper presents a new algorithm that extends MFlexDT and utilizes chi-square based fuzzy partitioning to improve accuracy in stream data classification.
APPLIED SOFT COMPUTING
(2021)
Article
Mathematics
Yeling Yang, Feng Yi, Chuancheng Deng, Guang Sun
Summary: This paper analyzes the accuracy of the CHAID algorithm, introduces the reasons, applicable conditions, and application scope of the CHAID algorithm, and highlights the differences in branching principles between the CHAID algorithm and other common decision tree algorithms. Through a case study and comparison of vehicle customer satisfaction data, the results show that CHAID can analyze data well and reliably detect correlated factors. This paper provides the necessary information to understand the CHAID algorithm, enabling better choices for accurate results when using decision tree algorithms.
Article
Environmental Sciences
Jionghua Wang, Haowen Luo, Wenyu Li, Bo Huang
Summary: This study develops a method of function label classification using integrated features derived from remote sensing and crowdsensing data, and verified on a dataset from Shenzhen, China. It was found that basic building attributes and POIs contributed most to the classification process, while crowdsensing data becomes increasingly important in more complicated classification tasks.
Article
Computer Science, Information Systems
Jiahao Ye, Jingjing Yang, Jiang Yu, Siqiao Tan, Feng Luo, Zheming Yuan, Yuan Chen
Summary: This paper introduces an adaptive multi-branch decision tree (CMDT) algorithm that improves the generalization ability of decision trees by introducing unequal interval optimization and local chi-square test. In imbalanced datasets, CMDT may be more reliable than other methods.
Article
Computer Science, Artificial Intelligence
Xinlei Zhou, Sudong Chen, Nianjiao Peng, Xinpeng Zhou, Xizhao Wang
Summary: A model tree is a hybrid learning algorithm that combines decision trees and embedding models, offering a simple structure and high interpretability. However, existing research on model trees fails to estimate the reliability of the output. This study introduces the output uncertainty of the embedding model into the model tree building process, providing guidance for model selection and optimization.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Mathematics, Interdisciplinary Applications
Harris Farooq, David Ryckelynck, Samuel Forest, Georges Cailletaud, Aldo Marano
Summary: The paper explores data pruning via hyperreduction modeling, introducing a lossy data compression/decompression approach for polycrystalline data based on a hyperreduction scheme that preserves data driven modeling capabilities. The focus is on reconstructing data through an oblique projection of selected original data by solving reduced mechanical equations. The relevance of pruned data is tested for statistics on predicted strain after the pruning process.
COMPUTATIONAL MECHANICS
(2021)
Article
Computer Science, Artificial Intelligence
Md Anisur Rahman, Mirko Duradoni, Andrea Guazzini
Summary: Research on Phubbing has traditionally focused on linear statistics, but more recent studies have utilized data mining and machine learning to identify patterns related to Phubbing behavior. These studies have shown that Phubbing is not solely determined by addiction measures, but also by non-linear relationships with ICT measures and social anxiety. Machine learning approaches have been found to be more effective at predicting Phubbing compared to traditional linear statistics.
NEURAL COMPUTING & APPLICATIONS
(2022)
Article
Computer Science, Artificial Intelligence
Teddy Lazebnik, Svetlana Bunimovich-Mendrazitsky
Summary: Decision tree is a popular technique in data mining, especially in the clinical domain. To improve the generalization capability and reduce the model size, we propose a new SAT-based pruning algorithm called SAT-PP. Experimental results on medical-related classification data sets show that the SAT-PP algorithm significantly reduces the model size and computation time (6.8%) while maintaining the same accuracy and F1 score as the unpruned decision tree.
DATA & KNOWLEDGE ENGINEERING
(2023)
Article
Computer Science, Information Systems
Guoqiang Li, Bowen Liu, Anbang Chen
Summary: In this paper, a neural network pruning technique based on data variance is proposed, which is robust and not affected by the number of data batches and training sessions. Additionally, a pruning compensation technique is introduced to retain pruned information effectively. Experimental results on standard datasets demonstrate the superiority of the proposed method.
JOURNAL OF VISUAL COMMUNICATION AND IMAGE REPRESENTATION
(2023)
Article
Plant Sciences
Yuxing Fu, Yuyang Xia, Huiming Zhang, Meng Fu, Yong Wang, Wei Fu, Congju Shen
Summary: This study proposes a method of pruning point identification based on skeleton information to realize automatic pruning of jujube trees. The RGB-D camera is used to collect multi-view information and build a complete point cloud information model. The space colonization algorithm generates the skeleton of the jujube trees, and the pruning points are identified on the primary branches based on pruning rules. The visual model of the pruned jujube tree is established through the skeleton information.
FRONTIERS IN PLANT SCIENCE
(2023)
Article
Mathematical & Computational Biology
Chao-Peng Yang, Jian-Qing Zhu, Tan Yan, Qiu-Ling Su, Li-Xin Zheng
Summary: This paper studies a lightweight pneumonia classification network based on deep learning methods to meet the requirements of clinical pneumonia auxiliary diagnosis. Through channel pruning and compression, the proposed method effectively reduces the number of parameters, improves speed, and achieves good experimental results in pneumonia auxiliary diagnosis.
COMPUTATIONAL AND MATHEMATICAL METHODS IN MEDICINE
(2022)
Article
Ecology
Sonam Sah, Dipanwita Haldar, Subhash Chandra, Ajeet Singh Nain
Summary: This study introduces a novel approach using Synthetic Aperture Radar (SAR) data to discriminate between different rice cultural types and subtypes and identify their optimum stages for representing crop growth profiles. The findings show that C-band co and cross-polarized SAR data can accurately detect differences between different rice types and subtypes, enabling effective crop growth monitoring.
ECOLOGICAL INFORMATICS
(2023)
Article
Computer Science, Artificial Intelligence
Chenxia Jin, Fachao Li, Shijie Ma, Ying Wang
Summary: Obtaining comprehensible classification rules is crucial in real applications, and decision-tree methods are commonly used. However, their performance is unsatisfactory and lacks theoretical support in big data scenarios. This study introduces a sampling-based classification rule mining (SCRM) method to improve the adaptability and generalization ability of classification rules in big data environments. The SCRM was evaluated using seven UCI datasets and showed good classification ability.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Jung-Sik Hong, Jeongeon Lee, Min K. Sim
Summary: This study proposes a novel Concise Algorithm to effectively remove irrelevant conditions from classification rules, aiming to enhance the interpretability of machine learning models.
EXPERT SYSTEMS WITH APPLICATIONS
(2024)
Article
Mathematics
Dejan Djordjevic, Dragan Cockalo, Srdjan Bogetic, Mihalj Bakator
Summary: This study analyzed entrepreneurial intentions among Serbian youth and found that demographic characteristics, close social environment, attitudes, awareness of incentive means, and environment assessment can partially predict entrepreneurial intentions. The results differed from similar studies, possibly due to the robust dataset, expanding the current literature and inviting future research.
Article
Computer Science, Artificial Intelligence
Stefania-Iuliana Soiman, Ionela Rusu, Stefan-Gheorghe Pentiuc
ADVANCES IN ELECTRICAL AND COMPUTER ENGINEERING
(2015)
Article
Computer Science, Information Systems
Ionut-Alexandru Zaiti, Stefan-Gheorghe Pentiuc, Radu-Daniel Vatavu
PERSONAL AND UBIQUITOUS COMPUTING
(2015)
Article
Computer Science, Software Engineering
Vasile Purdila, Stefan-Gheorghe Pentiuc
SOFTWARE-PRACTICE & EXPERIENCE
(2016)
Article
Multidisciplinary Sciences
Laura Bianca Bilius, Stefan Gheorghe Pentiuc
Correction
Multidisciplinary Sciences
Bilius Laura Bianca, Pentiuc Stefan Gheorghe
Article
Chemistry, Analytical
Laura Bianca Bilius, Stefan Gheorghe Pentiuc
Proceedings Paper
Computer Science, Theory & Methods
Laura-Bianca Bilius, Stefan Gheorghe Pentiuc
2020 15TH INTERNATIONAL CONFERENCE ON DEVELOPMENT AND APPLICATION SYSTEMS (DAS)
(2020)
Article
Chemistry, Physical
Stelian Alaci, Ilie Musca, Stefan-Gheorghe Pentiuc
Proceedings Paper
Computer Science, Theory & Methods
Stefan-Gheorghe Pentiuc, Oana-Mihaela Vultur
2018 14TH INTERNATIONAL CONFERENCE ON DEVELOPMENT AND APPLICATION SYSTEMS (DAS)
(2018)
Proceedings Paper
Computer Science, Theory & Methods
Oana-Mihaela Vultur, Stefan-Gheorghe Pentiuc, Valeriu Lupu
2016 13TH INTERNATIONAL CONFERENCE ON DEVELOPMENT AND APPLICATION SYSTEMS (DAS 2016)
(2016)
Proceedings Paper
Automation & Control Systems
Stefania-Iuliana Soiman, Ionela Rusu, Stefan-Gheorghe Pentiuc
2015 20TH INTERNATIONAL CONFERENCE ON CONTROL SYSTEMS AND COMPUTER SCIENCE
(2015)
Proceedings Paper
Computer Science, Hardware & Architecture
Felicia Giza-Belciug, Stefan-Gheorghe Pentiuc
2015 14TH ROEDUNET INTERNATIONAL CONFERENCE - NETWORKING IN EDUCATION AND RESEARCH (ROEDUNET NER)
(2015)
Article
Neurosciences
Mirella Amelia Mioc, Stefan-Gheorghe Pentiuc
BRAIN-BROAD RESEARCH IN ARTIFICIAL INTELLIGENCE AND NEUROSCIENCE
(2015)
Proceedings Paper
Education & Educational Research
Gabriel Cramariuc, Stefan Gheorghe Pentiuc
RETHINKING EDUCATION BY LEVERAGING THE ELEARNING PILLAR OF THE DIGITAL AGENDA FOR EUROPE!, VOL. I
(2015)
