Time Series Forecasting Utilizing Automated Machine Learning (AutoML): A Comparative Analysis Study on Diverse Datasets

Information, Journal Year: 2024, Volume and Issue: 15(1), P. 39 - 39

Published: Jan. 11, 2024

Automated Machine Learning (AutoML) tools are revolutionizing the field of machine learning by significantly reducing need for deep computer science expertise. Designed to make ML more accessible, they enable users build high-performing models without extensive technical knowledge. This study delves into these in context time series analysis, which is essential forecasting future trends from historical data. We evaluate three prominent AutoML tools—AutoGluon, Auto-Sklearn, and PyCaret—across various metrics, employing diverse datasets that include Bitcoin COVID-19 The results reveal performance each tool highly dependent on specific dataset its ability manage complexities thorough investigation not only demonstrates strengths limitations but also highlights criticality dataset-specific considerations analysis. Offering valuable insights both practitioners researchers, this emphasizes ongoing research development specialized area. It aims serve as a reference organizations dealing with guiding framework academic enhancing application

Language: Английский

---

Secure Internet Financial Transactions: A Framework Integrating Multi-Factor Authentication and Machine Learning

AI, Journal Year: 2024, Volume and Issue: 5(1), P. 177 - 194

Published: Jan. 10, 2024

Securing online financial transactions has become a critical concern in an era where services are becoming more and digital. The transition to digital platforms for conducting daily exposed customers possible risks from cybercriminals. This study proposed framework that combines multi-factor authentication machine learning increase the safety of transactions. Our methodology is based on using two layers security. first layer incorporates factors authenticate users. second utilizes component, which triggered when system detects potential fraud. employs facial recognition as decisive factor further protection. To build model, four supervised classifiers were tested: logistic regression, decision trees, random forest, naive Bayes. results showed accuracy each classifier was 97.938%, 97.881%, 96.717%, 92.354%, respectively. study’s superiority due its methodology, integrates embedded address usability, efficacy, dynamic nature various e-commerce platform features. With evolving landscape, continuous exploration datasets enhance adapt security measures will be considered future work.

Language: Английский

---

Data Pipeline Training: Integrating AutoML to Optimize the Data Flow of Machine Learning Models

Published: March 1, 2024

Data Pipeline plays an indispensable role in tasks such as modeling machine learning and developing data products. With the increasing diversification complexity of sources, well rapid growth volumes, building efficient has become crucial for improving work efficiency solving complex problems. This paper focuses on exploring how to optimize flow through automated methods by integrating AutoML with Pipeline. We will discuss leverage technology enhance intelligence Pipeline, thereby achieving better results tasks. By delving into automation optimization flows, we uncover key strategies constructing pipelines that can adapt ever-changing landscape. not only accelerates process but also provides innovative solutions problems, enabling more significant outcomes increasingly intricate domains.

Language: Английский

---

A Technical Comparative Heart Disease Prediction Framework Using Boosting Ensemble Techniques

Computation, Journal Year: 2024, Volume and Issue: 12(1), P. 15 - 15

Published: Jan. 16, 2024

This paper addresses the global surge in heart disease prevalence and its impact on public health, stressing need for accurate predictive models. The timely identification of individuals at risk developing cardiovascular ailments is paramount implementing preventive measures interventions. World Health Organization (WHO) reports that diseases, responsible an alarming 17.9 million annual fatalities, constitute a significant 31% mortality rate. intricate clinical landscape, characterized by inherent variability complex interplay factors, poses challenges accurately diagnosing severity cardiac conditions predicting their progression. Consequently, early emerges as pivotal factor successful treatment heart-related ailments. research presents comprehensive framework prediction leveraging advanced boosting techniques machine learning methodologies, including Cat boost, Random Forest, Gradient boosting, Light GBM, Ada boost. Focusing “Early Heart Disease Prediction using Boosting Techniques”, this aims to contribute development robust models capable reliably forecasting health risks. Model performance rigorously assessed substantial dataset illnesses from UCI library. With 26 feature-based numerical categorical variables, encompasses 8763 samples collected globally. empirical findings highlight AdaBoost preeminent performer, achieving notable accuracy 95% excelling metrics such negative predicted value (0.83), false positive rate (0.04), (0.01). These results underscore AdaBoost’s superiority overall compared alternative algorithms, contributing valuable insights field prediction.

Language: Английский

---

Benchmarking Automated Machine Learning (AutoML) Frameworks for Object Detection

Information, Journal Year: 2024, Volume and Issue: 15(1), P. 63 - 63

Published: Jan. 21, 2024

Automated Machine Learning (AutoML) is a subdomain of machine learning that seeks to expand the usability traditional methods non-expert users by automating various tasks which normally require manual configuration. Prior benchmarking studies on AutoML systems—whose aim compare and evaluate their capabilities—have mostly focused tabular or structured data. In this study, we systems task object detection curating three commonly used datasets (Open Images V7, Microsoft COCO 2017, Pascal VOC2012) in order benchmark different frameworks—namely, Google’s Vertex AI, NVIDIA’s TAO, AutoGluon. We reduced only include images with single instance understand effect class imbalance, as well dataset size. metrics average precision (AP) mean (mAP). Solely terms accuracy, our results indicate AutoGluon best-performing framework, mAP 0.8901, 0.8972, 0.8644 for VOC2012, Open V7 datasets, respectively. NVIDIA TAO achieved 0.8254, 0.8165, 0.7754 those same while VertexAI scored 0.855, 0.793, 0.761. found size had an inverse relationship across all frameworks, there was no between imbalance accuracy. Furthermore, discuss each framework’s relative benefits drawbacks from standpoint ease use. This study also points out issues examined labels subset dataset. Labeling errors appear have substantial negative accuracy not resolved larger datasets. Overall, provides platform future development research nascent field learning.

Language: Английский

---

Leveraging Azure Automated Machine Learning and CatBoost Gradient Boosting Algorithm for Service Quality Prediction in Hospitality

Computers, Journal Year: 2025, Volume and Issue: 14(2), P. 32 - 32

Published: Jan. 22, 2025

The importance of measuring service quality for business performance has been widely recognized in marketing literature due to its pivotal influence on customer satisfaction and long-term impact loyalty. SERVQUAL model, comprising five dimensions—reliability, assurance, tangibility, empathy, responsiveness—provides a measurable framework evaluating the overall satisfaction. This study endeavors ascertain whether all dimensions carry equal weight their effect estimate based various input features. To achieve this, questions were framed assess variables such as gender, age, marital status, highest level education, frequency hotel stays. each feature relative was investigated using machine learning models, specifically, CatBoost Microsoft Azure Automated Machine Learning (AutoML) studio. revealed that both AutoML identified stays age group dominant predictors quality. Additionally, highlighted status more significant factor, suggesting potential preferences. comparative modeling results demonstrated strong alignment between derived from AutoML, enabling decision-makers identify which are influenced by specific focus targeted improvements.

Language: Английский

---

Temporal Data Simulation and Drift-Resilient Machine Learning in Cardiovascular Disease Management: A Technical Analysis

International Journal of Scientific Research in Computer Science Engineering and Information Technology, Journal Year: 2025, Volume and Issue: 11(1), P. 1211 - 1218

Published: Jan. 28, 2025

Cardiovascular diseases remain a leading cause of death globally, necessitating advanced tools for effective prediction, prevention, and management. Machine learning has emerged as transformative approach in healthcare, offering solutions risk assessment, disease progression modeling, personalized treatment recommendations. However, the performance ML models often deteriorates over time due to data drift—shifts distributions, relationships between variables, or diagnostic thresholds—posing significant challenges dynamic healthcare environments. This article explores methods simulating temporal designing machine infrastructures resilient drift, focusing on their applications CVD The examines techniques including Autoregressive Integrated Moving Average, Hidden Markov Models, adaptive strategies modeling evolving trends cardiovascular metrics. To address paper highlights detecting mitigating its effects model through comprehensive monitoring frameworks validation protocols. Additionally, integrating simulated into pipelines, automated retraining workflows continual systems that maintain robustness, are reviewed. These approaches applied predict cardiac events, optimize plans, manage hospital resources. Ethical considerations, such fairness datasets, privacy protection, practical implementation challenges, also discussed.

Language: Английский

---

An AutoGluon-enabled robust machine learning model for concrete tensile and compressive strength forecast

International Journal of Construction Management, Journal Year: 2025, Volume and Issue: unknown, P. 1 - 12

Published: Feb. 4, 2025

Language: Английский

---

Assessing Innovation, Sustainability, and Market Performance in Taiwan's Semiconductor Sector: Insights From ESG‐Driven Analysis

Business Strategy and the Environment, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 11, 2025

ABSTRACT This study examines the relationship between ESG sub‐indicators and performance of Taiwan's semiconductor industry from 2016 to 2020. Using a combination data envelopment analysis, truncated regression, classification regression trees, research evaluates influence 12 factors on innovation, sustainability, market performance. The findings reveal that midstream manufacturers lead in innovation performance, while upstream excel sustainability. Corporate governance transparency emerges as most critical factor driving overall followed by employee management, product quality, stakeholder treatment. Conversely, greenhouse gas emissions waste management have limited impact due high costs regulatory challenges. highlights need for firms balance strategies with goals, emphasizing energy key levers competitiveness. provides practical insights into optimizing implementation enhance across value chain.

Language: Английский

---

Evaluating automated machine learning platforms for use in healthcare

JAMIA Open, Journal Year: 2024, Volume and Issue: 7(2)

Published: April 8, 2024

Abstract Objective To describe development and application of a checklist criteria for selecting an automated machine learning (Auto ML) platform use in creating clinical ML models. Materials Methods Evaluation Auto suited to needs local health district were developed 3 steps: (1) identification key requirements, (2) market scan, (3) assessment process with desired outcomes. Results The final comprising 21 functional 6 non-functional was applied vendor submissions heparin dosing model as case. Discussion A team clinicians, data scientists, stakeholders which can be adapted healthcare organizations, the case providing relevant example. Conclusion An evaluative platforms requires validation larger multi-site studies.

Language: Английский

---