Asian Journal of Research in Computer Science
Not a member yet
792 research outputs found
Sort by
Autonomous Database Systems – A Systematic Review of Self-Healing and Self-Tuning Database Systems
No full textProblem Statement: Autonomous database systems represent a significant change in the management of databases, utilizing Machine Learning (ML) and Artificial Intelligence (AI) in order to carry out self-healing and self-tuning with minimal human intervention.
Objectives: This systematic review investigates the defining characteristics, AI/ML techniques, challenges and the future trends of self-healing and self-tuning autonomous databases.
Methodology: The research questions were answered integrating findings from 35 current literatures between 2020 and 2025. These literatures were obtained from several reputable databases.
Results: From the study, self-healing databases employ techniques such as autoencoders, hidden Markov models, clustering algorithms, reinforcement learning, Bayesian optimization, neural networks and surrogate models to detect and recover from faults, enhancing operational resilience. On the other hand, self-tuning databases employ reinforcement learning, neural networks, multi-armed bandit techniques, decision trees, regression models, Bayesian optimization, and anomaly detection to optimize query execution, indexing, and resource allocation. Challenges in applying AI/ML in autonomous databases study include data quality dependencies, and adaptation to dynamic workload still exists and integration into existing infrastructures.
Conclusion: The deeper integration of deep learning techniques and predictive modelling serves as future trends to improve this autonomy
Construction of a Random Forest-based Machine Learning Model for Depression Prediction: Application to the Analysis of Disordered Behaviors
No full textThe topics covered in this article include the creation of a bootstrap learning model for depression predictions based on the Random Forest technique. Depression is a severe mental illness that affects millions of people worldwide. This condition causes great pain and affects quality of life, and in the most severe cases, the person takes their own life. Depression has a high incidence, but diagnosis is always complex and often delayed. It is made on the basis of clinical assessment, which is subjective, and questionnaires, which are often inaccurate and cannot identify people at risk early enough because a person\u27s subjective perception can often be distorted.
In this context, and to illustrate our point, we aim to show how AI, and more specifically machine learning, can provide innovative applications that can be used to improve early detection cf. prevention of depression risk. Instead of stupidly defining score intervals for a child, we can train a model on a dataset to identify patterns and correlations that escape simple regression analyses. Then, we can anticipate the first signs of log-in with depression, or we can identify which combinations of self and family history are most concerning. To complement our study, we chose the decision tree ensemble algorithm.
The article highlights the need for more objective and effective prediction tools for depression, and proposes a machine learning-based solution to achieve this, potentially leading to earlier intervention and better patient care
Dual-Stage Deep Learning Framework for Hybrid Coconut Maturity Classification and Harvest Timeline Prediction
No full textCoconut maturity estimation is critical in agriculture, as harvesting directly affects product quality, oil yield, and economic returns. Traditional methods, such as tapping or visual inspection, are subjective and inconsistent. This research develops a dual-stage deep learning framework that enables classification of coconut maturity stages and prediction of harvest timelines. The framework integrates a hybrid convolutional neural network (EfficientNetB0 + DenseNet121) for classification with a MobileNetV2-based regression model for predicting harvest time in immature coconuts. Images were collected, preprocessed, and augmented to balance classes. The models were trained and validated using accuracy, F1-score, mean absolute error, and root mean square error. A Gradio-based web application was developed to enable real-time image upload, classification, and timeline estimation. The hybrid classifier achieved over 99% accuracy, outperforming single-model baselines, while the regression model recorded an MAE of 36 days and an RMSE of 27 days, confirming reliable predictions. The web interface demonstrates practical usability and accessibility for farmers. While the dataset was limited in size and scope, which may affect generalizability, this study introduces the first dual-stage coconut framework that combines classification and predictive modeling into a practical, scalable system deployable on mobile and edge devices. Beyond its practical contributions, the study also advances agricultural AI research by extending coconut maturity studies from static classification into predictive modeling, a direction that remains underexplored. Future research will focus on expanding the dataset to diverse environments, integrating multimodal variables such as weather and soil data, and enhancing robustness under real-world conditions
Hybrid Sleep Scheduling for Energy-Efficient IoT Sensor Networks in Smart Poultry Monitoring
No full textThe integration of Internet of Things (IoT) technologies into precision poultry farming has revolutionized environmental monitoring; yet, high energy consumption in sensor networks remains a significant barrier to scalability and sustainability. This study presents a hybrid sleep scheduling algorithm for energy-efficient IoT-based poultry environmental monitoring. The algorithm enables dynamic transitions between Active, Modem Sleep, and Light Sleep modes according to environmental stability and data variability. Analytical models of system cycle time and power consumption were developed to optimise node behaviour under varying farm conditions. A prototype built with Wemos D1 Mini microcontrollers, DHT22, and MQ135 sensors was experimentally validated in a live poultry environment. Results show an average energy reduction of 68.4% compared to always-active systems, while maintaining latency below 2 seconds and measurement errors within ±0.4°C, ±1.3% RH, and ±7 ppm. The proposed framework offers a scalable, low-power architecture suitable for remote, battery-powered farms, advancing the sustainability of IoT-enabled livestock management and supporting the United Nations’ SDGs 2 and 12
Big Data Driven Cyber Threat Intelligence Framework for U.S. Critical Infrastructure Protection
No full textThe key infrastructure systems found throughout the U.S.—energy, transportation, healthcare, and water systems—are becoming ever more dependent on connected virtual online networks, thus increasing their vulnerability to both more ubiquitous and sophisticated cyber threats. Traditional security measures are unable to adapt to the volume, velocity, and variety of data generated by today’s cyber-attacks. This paper offers a Big Data-Driven Cyber Threat Intelligence Framework (BD-CTIF) that simultaneously takes advantage of real-time networking and IoT data-sharing, distributed analytics, and AI-based anomaly detection at the speed of business to provide proactive threat intelligence for U.S. critical infrastructures. Tests showed low latency, and high accuracy of detection demonstrating the framework\u27s utility for protecting U.S. national critical infrastructure. The proposed BDA methods borrowed from artificial intelligence (AI), machine learning (ML), natural language processing (NLP), and then employ deep knowledge to cite massive data sets for anomalies and respond to potential threats with high accuracy. This paper reviews the interrelationship of machine learning, artificial intelligence, and biological warfare with fresh insights into how those converge in relationship to cyber security for critical infrastructures. A review of the advantages, challenges, and options for operational use are considered in the discussion. Ultimately, this work demonstrates unrealized potential for any of the areas of artificial intelligence (AI)
Autoencoders for Clinical Data Analysis: Application of Neural Network-Based Dimensionality Reduction on Fine-Needle Aspiration Breast Data
No full textObjective: Machine learning provides powerful tools for analyzing large datasets; however, it faces challenges such as high computational costs and overfitting. To overcome these issues techniques that reduce the dimensionality of data are frequently used. Dimensionality reduction aims to eliminate redundant or unnecessary information in the dataset thereby reducing computational load and improving the model\u27s ability to generate more accurate results. The primary objective of this study is to evaluate the performance of the Autoencoder algorithm, one of the dimensionality reduction methods. This study will thoroughly examine the effectiveness of the Autoencoder algorithm in terms of data loss processing time and the model’s performance on new data.
Materials and Methods: Breast masses can be effectively analyzed using quantitative features of cell nuclei obtained from fine-needle aspiration (FNA) samples. This study aimed to evaluate the performance of the Autoencoder algorithm for dimensionality reduction on these features. The analysis was conducted on 569 cases from the Wisconsin Diagnostic Breast Cancer (WDBC) dataset, accessible via an online repository provided by the University of Wisconsin–Madison. The dataset included quantitative features for each cell nucleus, specifically radius, smoothness, compactness, and concavity. The Autoencoder algorithm was applied to the entire dataset to reduce dimensionality while preserving relevant information. To illustrate its operation concretely, four primary features from five randomly selected observations were used, demonstrating the algorithm’s performance on a small, non-linear subset of the data. For comparison, commonly used dimensionality reduction techniques, including Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA), were also applied. Results indicate that the flexible architecture of Autoencoders effectively captures the most informative features, supporting their applicability to clinical datasets and potential integration into computer-aided diagnostic workflows. This approach provides a reliable foundation for analyzing complex biomedical data and assessing algorithm performance in real-world clinical contexts.
Results: This study focused on a comparative analysis of four key variables radius mean, smoothness mean, compactness mean, and concavity mean using both the original dataset and the dataset reconstructed through an Autoencoder model. In the original dataset, the mean and standard deviation values of these variables were calculated as 14.13 ± 3.52, 0.10 ± 0.01, 0.10 ± 0.05, and 0.09 ± 0.08, respectively. At the output layer, the Autoencoder successfully reconstructed the input features, preserving their mean values and yielding corresponding mean ± standard deviation values of 14.13 ± 2.38, 0.10 ± 0.01, 0.10 ± 0.05, and 0.09 ± 0.07. The reduction in standard deviations in the reconstructed dataset, particularly for the radius mean and concavity mean variables, indicates decreased variability and suggests that the model produced a more compact representation while retaining the essential characteristics of the data. The primary objective of the Autoencoder is to ensure that the output closely resembles the original input by utilizing a hidden layer (h) that captures the essential structure of the data. Aligned with this purpose, the algorithm effectively compressed the four-dimensional input into a more compact latent representation while preserving key characteristics. The analyses showed that the hidden layer representations were highly consistent with the original data and were optimized successfully. Consequently, the dimensionality of the dataset was reduced from four variables to a lower-dimensional representation, enabling a more efficient and informative encoding of the data.
Conclusion: This study evaluated the performance of the Autoencoder algorithm for dimensionality reduction using quantitative features of cell nuclei obtained from fine-needle aspiration (FNA) samples of breast masses. The analysis was conducted on a dataset of 569 cases, and to illustrate the algorithm’s operation, data from four key features (radius, smoothness, compactness, and concavity) of five randomly selected observations were used as examples. This approach allowed for the demonstration of the Autoencoder’s performance on small and non-linear subsets of the data. The findings indicate that dimensionality reduction plays a significant role in clinical data analysis and that the Autoencoder algorithm also reduces computational costs. These results confirm the potential of Autoencoders as a reliable and effective tool for dimensionality reduction. Consequently, the use of Autoencoders can enable faster, more accurate, and more efficient processing of healthcare data, thereby enhancing the effectiveness of clinical decision support systems
From \u27Sexiest Job\u27 to \u27Most Responsible Role\u27: The Evolution of Data Scientists
No full textThis opinion article explores the evolving responsibilities of data scientists in the current data-driven landscape, in which ethical, privacy, and governance standards have grown considerably in importance. Although the job of data scientist initially attracted attention for its allure and high earning potential, in recent years, it has become associated with a particularly high level of responsibility, requiring practitioners to balance their technical skills with a commitment to social impact and accountability. This article examines the essential qualifications and criteria of a responsible data scientist, including a robust ethical awareness, an understanding of privacy safeguards and transparency, and a commitment to continuous learning. This article also discusses hiring practices that prioritize these qualities and outlines strategies for fostering a data-driven culture grounded in responsibility and trust. In the current landscape, responsible data scientists not only analyze data but also protect ethical data practices, which is crucial to building a transparent, fair, and sustainable digital world. This article also provides a framework and guidelines for identifying and recruiting responsible data scientists
The Role of AI in Early Detection of Alzheimer\u27s and Parkinson\u27s Diseases: A Literature Survey
No full textEarly detection of neurodegenerative diseases like Alzheimer’s and Parkinson’s is crucial for improving patient care and enabling timely interventions. Artificial intelligence (AI) offers innovative approaches to analyzing complex medical datasets, revolutionizing the detection of these diseases at early stages. This review discusses key AI methodologies, including machine learning (ML), deep learning (DL), natural language processing (NLP), and reinforcement learning (RL), and their applications in early diagnosis. ML models excel in predicting disease risk and classifying imaging and biometric data, while DL techniques, such as convolutional and recurrent neural networks, are effective in processing unstructured data like images and speech. NLP facilitates extracting critical insights from clinical notes and patient narratives, and RL enhances decision-making in diagnostic workflows. Integrating multimodal data—such as genomics, neuroimaging, wearable device metrics, and electronic health records—further strengthens diagnostic precision. Despite its promise, the widespread implementation of AI faces challenges, including the need for standardized data, ethical considerations, and clinical validation. Overcoming these obstacles is essential for AI to transform early detection and management of neurodegenerative diseases. This review emphasizes the significance of interdisciplinary efforts and sustained research to unlock AI’s full potential in medical applications
Prediction Liver Diseases based on Machine Learning and Deep Learning Techniques: A Review
No full textVarious critical issues in liver diseases include cirrhosis, hepatitis, and liver cancer, which can be fatal. They indeed require early diagnosis with appropriate diagnosis of the disease. The different conventional diagnostic methods generally can\u27t identify these diseases during their early stages; consequently, prognosis is not always good. Recently, in subsequence to improve this gap, ML/DL has emerged as the tool for transformation. It gives an overview of different ML and DL models used for predicting liver diseases, including supervised, unsupervised, semi-supervised learning, and reinforcement learning, and emphasizes the better performance that deep learning models like Convolutional Neural Networks (CNNs) and Bidirectional Long Short-Term Memory (Bi-LSTM) networks are providing in handling complex medical data. These DL models perform significantly better in diagnostic accuracy when compared to the traditional ML methods, hence holding tremendous potential in their medical applications. Besides, hybrid and ensemble methods, which are combined models, are emphasized for their ability to overcome the limitations of individual algorithms and enhance diagnostic precision and robustness. This study further underlines the need to develop more advanced DL methodologies for the early detection and intervention in liver diseases, which is necessary to reduce the global burden and improve patient outcomes
Comparative Analysis of AI-powered Outbound Dialer Campaigns vs. Legacy Outbound Dialer Campaigns (Contact Center)
No full textAim: Scope of this work aims to explore the effectiveness of AI-powered outbound dialer campaigns with legacy outbound dialer campaigns by evaluating their technological capabilities, operational efficiencies, compliance features, customer experience impact, cost-effectiveness, and scalability.
Study Design: This is a comparative analysis study evaluating two types of outbound dialer systems used in customer engagement campaigns.
Place and Duration of Study: This study is based on a review of industry practices and integration strategies in contact centers across various organizations, focusing on solutions implemented between 2018 and 2024.
Methodology: This study employed a comparative analysis methodology to evaluate AI-powered and legacy outbound dialer campaigns. The research was conducted through secondary data collection from industry reports, case studies, and real-world deployment insights over a study period spanning 2018 to 2024. Key performance metrics assessed included call connection rates, compliance adherence, personalization capabilities, operational costs, and scalability. Data was synthesized using a structured comparative framework to highlight technological, operational, and customer experience differences. Qualitative evaluations of agent productivity and customer satisfaction were also incorporated to provide a holistic analysis.
Results: AI-powered outbound dialers demonstrated superior efficiency with call connection rates increasing by up to 30% compared to legacy systems. Automated regulatory monitoring strengthened compliance adherence, cutting violation risks by 40%. Personalization capabilities drove a 25% increase in customer satisfaction scores, while automation initiatives reduced operational costs by 20%. In contrast, legacy dialers faced higher abandonment rates and lacked real-time adaptability. Scalability and omnichannel integration were also more seamless in AI-powered systems, supporting modern customer engagement strategies.
Conclusion: AI-powered outbound dialers outperform legacy systems in nearly all critical aspects, including efficiency, compliance, personalization, and cost-effectiveness. They provide a scalable, adaptive solution for businesses aiming to optimize customer engagement. While legacy systems may still be viable in smaller or less dynamic operations, AI-driven technologies are emerging as the preferred choice for future-proof outbound campaigns. Further adoption of AI-powered systems will continue to redefine industry standards