International Journal of Electrical and Computer Engineering Systems

Optimizing Gastric Cancer Classification with QCNN and Fine-Tuning

Aroop D S — 2025-04-24

Cancer ranks as one of the primary contributors to morbidity and mortality worldwide, standing as the second leading causeof death on a global scale. According to data from the National Cancer Registry Program of the Indian Council of Medical Research, over1300 individuals in India lose their lives daily as a result of cancer-related causes. Gastric cancer is among the top five most prevalent cancersglobally, after cancer in the lung, breast, colorectum, and prostate, highlighting the importance of accurate classification for effectivetreatment strategies. In this study, a novel approach utilizing a Quadratic Convolutional Neural Network combined with Extreme Learningand Fine-Tuning technique, a deep learning architecture specifically designed to capture intricate patterns and features within medicalimaging data. Fine tuning technique is used to enhance the model's generalization capability and adaptability to diverse datasets. Throughextensive experimentation and validation on a comprehensive dataset comprising gastric cancer images, the proposed approach achievesan impressive accuracy of 94%. The findings indicate the efficacy of the proposed approach for classifying gastric cancer. With its highaccuracy and robust performance, the developed QCNN model holds promise for assisting clinicians in accurate diagnosis and prognosis ofgastric cancer patients, ultimately contributing to improved patient outcomes and personalized treatment strategies.

Efficient Privacy-Utility Optimization for Differentially Private Deep Learning: Application to Medical Diagnosis

Rafika Benladgham — 2025-04-25

The optimization of differentially private deep learning models in medical data analysis using efficient hyper-parameter tuning is still a challenging task. In this context, we address the fundamental issue of balancing privacy guarantees with model utility by simultaneously optimizing model parameters and privacy parameters across two primary medical datasets, with additional validation on PathMNIST. Our framework encompasses both tabular data (Wisconsin Breast Cancer dataset) and medical imaging (BreastMNIST and PathMNIST), implementing four distinct optimization approaches: Grid Search, Random Search, Bayesian Optimization, and Bat Algorithm. Through extensive experimentation, we demonstrate a promising performance: achieving 93.62% accuracy with strong privacy guarantees (ε = 0.5) for tabular data, and 74.91% accuracy for medical imaging, with the Bat Algorithm discovering an unprecedented privacy level (ε = 0.293). Further validation on PathMNIST histopathology images demonstrated the framework's scalability, achieving 44.71% accuracy with privacy guarantees (ε = 2.603). Our comparative analysis reveals that different medical data types require distinct optimization strategies, with Bayesian Optimization excelling in tabular data applications and Random Search providing efficient solutions for image processing. The experiments with PathMNIST histopathology images provided valuable insights into the framework's behavior with complex medical data, revealing configuration-dependent performance variations and computational trade-offs. Our framework incorporates Pareto analysis and visualization techniques to enable systematic exploration of privacy-utility trade-offs, while early stopping mechanisms optimize privacy budget utilization. This comprehensive approach, validated across diverse medical imaging complexities and data modalities, establishes practical guidelines for implementing privacy-preserving machine learning in healthcare settings while highlighting the importance of balanced optimization strategies and computational efficiency in secure and efficient medical data analysis.

A Semantic Analysis Approach to Extract Personality Traits from Tweets (X)

Marouane Echhaimi — 2025-05-05

The utilization of social networks has experienced a substantial surge in the past decade, with individuals routinely exchanging and consuming personal data. This data, subject to analysis and utilization across diverse contexts, has spurred scholarly interest in discerning the personality traits of social network users. Personality, as an intrinsic characteristic, distinguishes individuals in terms of cognition, emotion, and behavior, thereby influencing social relationships and interactions. Among the extensively studied frameworks elucidating personality variance is the Five Factor Model, commonly referred to as the "Big Five," encompassing Openness, Conscientiousness, Extroversion, Agreeableness, and Neuroticism (OCEAN). Personality assessment holds practical utility across domains such as education, security, marketing, e-learning, healthcare, and personnel management. Prior investigations have demonstrated the feasibility of automatic text analysis in personality discernment. This paper introduces a multi-agent methodology grounded in semantic similarity metrics for personality trait recognition via automatic text analysis of Tweets. Our approach leverages WordNet and information content-based semantic similarity measures to analyze tweet content and classify users' personality traits. Experimental results demonstrate the effectiveness of our method, achieving a remarkable 96.28% accuracy in identifying personality traits from Tweets. This high success rate underscores the potential of our semantic analysis approach in accurately profiling social media users' personalities, offering valuable insights for various applications in behavioral analysis and personalized services.

FusionNet- A Hybrid Deep Learning Approach for Accurate Drug-Target Binding Prediction

Tintu Vijayan — 2025-05-05

Identifying drug-target binding affinities (DTBA) is crucial in drug discovery, to understand how effectively drugs interact with their targets. However, traditional methods often struggle to accurately capture the complex relationships in biological data, leading to limitations in their predictive power. This paper introduces FusionNet, an advanced deep-learning model designed to improve DTBA prediction. FusionNet combines the strengths of Convolutional Neural Networks (CNNs), Long Short-Term Memory networks (LSTMs), and Transformers, to better understand both short-range and long-range interactions in biological sequences and employs the Layer-wise Adaptive Moments (LAMB) optimizer, which ensures the model is more efficient and stable, especially when working with large datasets. FusionNet achieved an MSE of 0.20 and an rm2 of 0.681 on the Davis dataset and an MSE of 0.18 and an rm2 of 0.71 on the KIBA dataset, significantly outperforming existing models like SimBoost, GANsDTA, DeepCDA, and DeepDTA, making it a powerful tool for drug discovery and bioinformatics. This work not only enhances the accuracy of DTBA prediction but also sets new performance standards by integrating advanced neural network architectures and optimizing their training process. FusionNet effectively addresses the limitations of previous approaches, offering a more reliable and efficient method for predicting drug-target interactions.

An Improved MPPT Scheme for Photovoltaic Systems Using a Novel MRAC-FUZZY Controller

Mahbouba Brahmi — 2025-05-13

This paper presents a novel and highly effective fuzzy model reference adaptive control for MPPT based on a boost converter. The design of Model-Referenced Adaptive Control (MRAC) and the adaptive gain selection are discussed. The adjustment of the adaptation gains by a fuzzy logic subsystem and a simplified fuzzy MRAC procedure are presented. The suggested algorithm is assessed through a comprehensive simulation in MATLAB/Simulink. Various scenarios and environmental conditions are considered to assess its robustness and adaptability. The results indicate that the suggested MRAC-Fuzzy MPPT control is extremely robust, with tracking efficiency that can reach 99.97%. Furthermore, it consistently operates the photovoltaic system at or around the MPP, effectively reducing oscillations, improving energy efficiency, and enhancing power production. Under real operating conditions, this new controller can be used for photovoltaic pumping applications.

3D-based Convolutional Neural Networks for Medical Image Segmentation: A Review

Siti Raihanah Abdani — 2025-04-17

Medical image segmentation is essential for disease screening and diagnosis, particularly through techniques like anatomical and lesion segmentation that can be used to isolate critical regions of interest. However, manual segmentation is labor-intensive, costly, and susceptible to subjective bias, underscoring the need for automation. Deep learning, particularly convolutional neural networks (CNNs), has significantly advanced segmentation accuracy and efficiency. With the introduction of 3D imaging, research has evolved from 2D CNNs to 3D CNNs, which leverage inter-slice information to improve segmentation precision. This paper aims to provide a literature review of studies published between 2018 and 2024 on platforms such as Google Scholar and ScienceDirect, where the identified relevant research are "3D segmentation" and "3D medical imaging". This study outlines the key stages of 3D CNN segmentation that include preprocessing, region-of-interest extraction, and post-processing. Furthermore, this study emphasizes the application of 3D CNN architectures to complex lung imaging scenarios, such as lung cancer and COVID-19. Although 3D CNNs outperform 2D CNNs in preserving spatial continuity across slices, they present notable limitations. Key challenges include heavy computational and high memory demands, as well as a dependency on large annotated datasets, which are often scarce in medical imaging. Additionally, effective multiscale feature learning remains a challenging issue, with current architectures struggling to generalize the features of interest across several usage variations. To further improve the segmentation performance, future research should prioritize developing adaptive algorithms and fostering interdisciplinary collaboration between computer scientists and medical professionals to design efficient and scalable models, designed specifically for clinical applications. This future research direction will enhance diagnostic accuracy and segmentation quality in 3D medical imaging.