Article
Mathematics
Matevz Kunaver, Arpad Burmen, Iztok Fajfar
Summary: This article introduces an improvement to recommender systems using grammatical evolution to automatically initialize and optimize the algorithm, allowing it to produce optimal results without requiring prior or in-depth knowledge from the user. Additionally, the approach is able to detect over-saturation in large datasets.
Article
Computer Science, Information Systems
Abdullah Khan, Junaid Bukhari, Javed Iqbal Bangash, Asfandyar Khan, Muhammad Imran, Muhammad Asim, Muhammad Ishaq, Arshad Khan
Summary: Classification is a common problem in various fields of life and a key challenging task in data mining. This paper proposes a model based on a high-order functional link neural network and an accelerated particle swarm optimization algorithm for medical data classification. Experimental results show that the proposed model performs well in terms of mean square error and accuracy.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2022)
Article
Chemistry, Multidisciplinary
Djamila Romaissa Beddiar, Mourad Oussalah, Tapio Seppanen, Rachid Jennane
Summary: In medical image captioning, a method that combines medical concepts with visual features to generate captions is proposed. An end-to-end trainable network is used to combine semantic features, visual features, and LSTM model, and beam search is employed to select the best next word. Experimental results demonstrate the effectiveness and efficiency of the proposed approach.
APPLIED SCIENCES-BASEL
(2022)
Article
Computer Science, Artificial Intelligence
Qurrat Ul Ain, Harith Al-Sahaf, Bing Xue, Mengjie Zhang
Summary: This study analyzes GP-based approaches to skin image classification, which improve the performance of machine learning classification algorithms by constructing features, thereby enhancing diagnostic efficiency and assisting dermatologists in diagnosis.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Computer Science, Software Engineering
Mohammad D. Alahmadi
Summary: The paper presents vid2XML, a method that leverages visual and textual information of video frames to locate XML regions, locate the currently opened file, and extract XML data for each file presented in video frames. Through a comprehensive empirical evaluation, the results show that vid2XML is able to accurately locate XML regions, locate the bounding box of the selected file, and extract, fix, and merge XML data for each file opened/created in a video.
IEEE TRANSACTIONS ON SOFTWARE ENGINEERING
(2023)
Article
Computer Science, Artificial Intelligence
S. Sreelakshmi, V. S. Anoop
Summary: Machine learning provides powerful techniques for automated disease diagnosis through medical image classification. Deep learning approaches have shown significant improvements in performance and accuracy compared to shallow learning techniques. This paper proposes a framework that combines deep convolutional neural networks with an enhanced feature extraction technique for classifying medical data. Experimental results demonstrate that the proposed approach outperforms chosen baselines by achieving an accuracy of 98.91% in classifying COVID-19 images from computed tomography scans.
Article
Computer Science, Software Engineering
Nikolaos Anastasopoulos, Ioannis G. Tsoulos, Alexandros Tzallas
Summary: A genetic programming tool based on Grammatical Evolution technique is proposed for data classification, utilizing multicore computing systems with the OpenMP library. It generates simple rules in a C-like programming language, and allows for user-defined BNF grammar, showing very promising results in a wide range of classification problems when compared to traditional techniques.
Article
Computer Science, Artificial Intelligence
Dimmy Magalhaes, Ricardo H. R. Lima, Aurora Pozo
Summary: Text classification is a NLP task aiming to label textual elements. Deep Learning-based approaches have been widely used in this context, particularly with DNNs. This paper presents a study on the application of a grammar-based evolutionary approach to design DNNs for text classification, using models based on CNNs, LSTM, and GNNs. Different grammars were proposed to capture the features of each type of network, and their impact on generated designs and performance of the models were evaluated.
APPLIED SOFT COMPUTING
(2023)
Article
Computer Science, Information Systems
Sarab AlMuhaideb, Reem Binghannam, Nourah Alhelal, Shatha Alduheshi, Fatimah Alkhamees, Raghad Alsuhaibani
Summary: This study focuses on applying classification methods to medical datasets in order to predict diagnoses or prognoses. A novel MH algorithm, IWD-Miner, which combines Intelligent Water Drops and AntMiner+, was developed and tested on 21 publicly available medical datasets. Results suggest that IWD-Miner is more efficient than AntMiner+ but has lower accuracy compared to J48, with the latter producing larger models.
CMC-COMPUTERS MATERIALS & CONTINUA
(2021)
Article
Computer Science, Artificial Intelligence
Joanna Jedrzejowicz, Piotr Jedrzejowicz
Summary: The paper introduces an incremental Gene Expression Programming classifier for mining imbalanced datasets, which adapts to the requirements of the imbalanced data environment by reusing minority class instances and applying the incremental learning paradigm. Through extensive computational experiments, it demonstrates competitive performance against other state-of-the-art learners in many cases.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Mathematics
Mohammad H. Nadimi-Shahraki, Ali Fatahi, Hoda Zamani, Seyedali Mirjalili
Summary: This paper proposes an efficient binary version of the quantum-based avian navigation optimizer algorithm (BQANA) for feature selection from high-dimensional medical datasets. It compares the performance of different binary versions of BQANA and other well-known metaheuristic algorithms to demonstrate the superiority of BQANA in finding optimal feature subsets.
Article
Computer Science, Information Systems
Masood Nekoei, Seyed Amirhossein Moghaddas, Emadaldin Mohammadi Golafshani, Amir H. Gandomi
Summary: In the field of artificial intelligence automatic programming, artificial bee colony expression programming (ABCEP) presents new solutions by using expression sharing to improve performance. Experimental results indicate that predictions generated by ABCEP outperform other automatic programming algorithms based on successful runs, mean fitness values, and convergence rate.
INFORMATION SCIENCES
(2021)
Article
Computer Science, Artificial Intelligence
Mohammed Eshtay, Hossam Faris, Ali Asghar Heidari, Ala' M. Al-Zoubi, Ibrahim Aljarah
Summary: Random Weight Networks have been widely used in various applications but traditional training techniques may suffer from local optima stagnation. This paper proposes a training method based on competitive swarm optimization, which can automatically adjust network parameters to obtain reasonable prediction results.
NEURAL COMPUTING & APPLICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Sundar Rengasamy, Punniyamoorthy Murugesan
Summary: A new family memory dimension was introduced in the PSO algorithm, resulting in the development of a modified MPSO algorithm. This algorithm was tested and found to significantly outperform other clustering algorithms in terms of multi-criteria inventory classification.
SWARM AND EVOLUTIONARY COMPUTATION
(2021)
Article
Computer Science, Interdisciplinary Applications
Mostefa Mesbah, Mohamed S. Khlif, Siamak Layeghy, Christine E. East, Shiying Dong, Amy Brodtmann, Paul B. Colditz, Boualem Boashash
Summary: This study proposed a novel automatic fetal movement recognition algorithm utilizing wearable tri-axial accelerometers placed on the maternal abdomen. By extracting multiple features and using various classifiers for identification and artefact removal, the Bagging classifier algorithm was found to perform the best in distinguishing fetal movements.
COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
(2021)
Article
Engineering, Biomedical
Dipak Kumar Patra, Tapas Si, Sukumar Mondal, Prakash Mukherjee
Summary: Breast DCE-MRI segmentation is crucial for accurate diagnosis of breast cancer, with optimization techniques playing an important role in differentiating lesions. The proposed SPBO algorithm achieves high accuracy levels and outperforms existing segmentation methods.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2021)
Article
Computer Science, Artificial Intelligence
Tapas Si, Jayri Bagchi, Pericles B. C. Miranda
Summary: This study investigates the performance of different metaheuristics as learning algorithms for training Artificial Neural Networks (ANN) in medical data classification tasks. The results establish that the Equilibrium Optimizer algorithm outperforms other algorithms in terms of performance.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Computer Science, Artificial Intelligence
Tapas Si, Dipak Kumar Patra, Sukumar Mondal, Prakash Mukherjee
Summary: The prevalence of breast cancer in women has been increasing dramatically in recent times. This article presents an entropy-based multilevel thresholding method for segmenting breast lesions in DCE-MRI using the Chimp Optimization Algorithm (ChOA). The proposed method achieves high sensitivity, accuracy, and Dice Efficient Coefficient (DSC) levels in the segmentation of breast lesions. The performance of the proposed method is compared with other existing algorithms, and the results demonstrate its superiority in both quantitative and qualitative outcomes.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Genetics & Heredity
Sana Munquad, Tapas Si, Saurav Mallik, Asim Bikas Das, Zhongming Zhao
Summary: In this study, a deep learning framework based on a convolutional neural network was developed for the accurate identification of GBM subtypes. The deep learning model outperformed traditional machine learning algorithms in subtype identification. The study also identified genotype-phenotype relationships of GBM subtypes and subtype-specific predictive biomarkers, providing potential diagnostic and treatment options.
FRONTIERS IN GENETICS
(2022)
Article
Computer Science, Interdisciplinary Applications
Dipak Kumar Patra, Tapas Si, Sukumar Mondal, Prakash Mukherjee
Summary: This paper proposes two breast lesion segmentation methods in DCE-MRI using MFO and WOA algorithms, with experimental results showing that these methods outperform other algorithms in terms of accuracy and sensitivity.
PATTERN RECOGNITION AND IMAGE ANALYSIS
(2022)
Article
Biotechnology & Applied Microbiology
Sana Munquad, Tapas Si, Saurav Mallik, Aimin Li, Asim Bikas Das
Summary: Classifying lower-grade gliomas (LGGs) accurately is crucial for effective therapeutic intervention. In this study, a machine learning based classification framework using transcriptome data was designed to diagnose LGG subtypes and grades. The framework achieved superior accuracy compared to other machine learning methods, and it was found that the accuracy of subtype classification was always higher in a specific grade rather than in mixed grade cancer. Furthermore, predictive biomarkers were identified through co-expression, gene set enrichment, and survival analysis.
BRIEFINGS IN FUNCTIONAL GENOMICS
(2022)
Article
Computer Science, Artificial Intelligence
Tapas Si, Dipak Kumar Patra, Sukumar Mondal, Prakash Mukherjee
Summary: This paper proposes a lesion segmentation method for breast DCE-MRI using the opposition-based Sine Cosine Algorithm (SCA), which improves segmentation efficiency through multi-level thresholding optimization. The MR images are denoised and intensity inhomogeneities are corrected using the Anisotropic Diffusion Filter. The lesions are then located from the segmented images. Experimental results show that the proposed method outperforms other algorithms in both quantitative and qualitative measures.
PATTERN ANALYSIS AND APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Tapas Si, Pericles B. C. Miranda, Debolina Bhattacharya
Summary: This article investigates the application of Opposition-based Learning (OBL) to the search process of Salp Swarm Algorithm (SSA) and develops five enhanced hybrid SSA-OBL algorithms. Experimental results show that the opposition-based SSAs statistically outperform traditional SSA and other competitive algorithms in terms of coverage, accuracy, exploration and exploitation, and convergence.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Computer Science, Artificial Intelligence
Shubhankar Bhakta, Utpal Nandi, Tapas Si, Sudipta Kr Ghosal, Chiranjit Changdar, Rajat Kumar Pal
Summary: This paper presents a novel optimizer called diffMoment, which adjusts the step size for each parameter based on changing information between the 1st and 2nd moment estimates. Experimental results show that diffMoment outperforms AdaGrad, Adam, AdaDelta, RAdam, and RMSProp optimizers, especially when training Convolutional Neural Networks (CNN) with different activation functions.
APPLIED INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Cleber A. C. F. da Silva, Daniel Carneiro Rosa, Pericles B. C. Miranda, Filipe R. Cordeiro, Tapas Si, Andre C. A. Nascimento, Rafael F. L. Mello, Paulo S. G. de Mattos Neto
Summary: This study proposes a method that uses a multi-objective grammatical evolution framework to automatically generate and optimize CNNs for image classification problems. The results show that the proposed method is able to generate simpler networks that statistically outperform state-of-the-art CNNs in terms of accuracy and F1-score.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Computer Science, Information Systems
Partha Chowdhuri, Pabitra Pal, Tapas Si
Summary: This study presents an efficient authentication scheme for digital image steganography on medical images using Support Vector Machine (SVM) and Integer Wavelet Transform (IWT). It employs SVM to separate the Region of Interest (ROI) from Non-Region of Interest (NROI) in the medical image, and IWT to embed secret information within the NROI part of the image.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Steffano Pereira, Pericles Miranda, Thiago Franca, Carmelo J. A. Bastos-Filho, Tapas Si
Summary: This paper proposes using a many-objective algorithm to generate synthetic datasets from four different characteristics of real-world problems, in order to assess the performance of classification algorithms. The results showed that the approach could optimize conflicting objectives, generating synthetic datasets with different complexities.
2022 IEEE LATIN AMERICAN CONFERENCE ON COMPUTATIONAL INTELLIGENCE (LA-CCI)
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Pericles B. C. Miranda, Rafael Ferreira Mello, Andre C. A. Nascimento, Tapas Si
Summary: This article proposes a multi-objective optimization method based on combinatorial optimization to determine the most promising sequence of sampling algorithms in order to improve the performance of classifiers in terms of accuracy and F-1 score. The results demonstrate that the proposed method is capable of finding optimized sequences that significantly improve the performance of classifiers compared to competing methods in most unbalanced problems.
2022 IEEE CONGRESS ON EVOLUTIONARY COMPUTATION (CEC)
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Cleber A. C. F. da Silva, Daniel Carneiro Rosa, Pericles B. C. Miranda, Filipe R. Cordeiro, Tapas Si, Andre C. A. Nascimento, Rafael F. L. Mello, Paulo S. G. de Mattos Neto
Summary: This paper proposes a Multi-Objective Grammatical Evolution framework to automatically generate suitable CNN architectures for a given classification problem, achieving good results on the CIFAR-10 dataset.
2021 IEEE CONGRESS ON EVOLUTIONARY COMPUTATION (CEC 2021)
(2021)