Ensemble Partition Sampling (EPS) for Improved Multi-Class Classification
Article
Subjects > Engineering
Europe University of Atlantic > Research > Scientific Production
Fundación Universitaria Internacional de Colombia > Research > Scientific Production
Ibero-american International University > Research > Scientific Production
Ibero-american International University > Research > Scientific Production
Universidad Internacional do Cuanza > Research > Scientific Production
Abierto
Inglés
Classification is a commonly used technique in data mining and is applied in various fields such as sentiment analysis, fraud detection, and fault diagnosis. Multiclass classification, which involves more than two classes, is more complex than binary classification. There are mainly two ways to approach multiclass classification, one is to expand the binary classifier into a multiclass classifier through various strategies and the other is to divide the multiclass classification problem into multiple binary problems (binarization). Two popular approaches for binarization are One vs One (OvO) and One vs All (OvA). It is simpler to aggregate the outputs of all binary classifiers as the number of classifiers decreases. However, it causes an imbalance of positive and negative sample numbers, which affects the classification effect of each binary classifier. In this article, we contribute to the field of ensemble learning and multi-class classification by proposing a new method called Ensemble Partition Sampling (EPS). This article presents a new approach to multiclass classification using an "Ensemble Partition Sampling" method within the "one-vs-all" (OvA) framework. The primary goal of this method is to tackle the problem of data imbalance by incorporating ensemble learning and preprocessing techniques into each binary dataset. The study found that Ensemble Partition Sampling (EPS) is the most effective method for imbalanced and multiclass imbalanced classification, outperforming other methods including OvA, SMOTE, k-means-SMOTE, Bagging-RB, DES-MI, OvO-EASY, and OvO-SMB. The study used CART, Random Forest, and SVM as classifiers, and the results consistently showed that EPS outperformed all other algorithms. The findings suggest that EPS is a highly effective method for improving classification performance in imbalanced and multiclass imbalanced datasets.
metadata
Jabir, Brahim and Díez, Isabel De la Torre and Bautista Thompson, Ernesto and Ramírez-Vargas, Debora L. and Kuc Castilla, Ángel Gabriel
mail
UNSPECIFIED
(2023)
Ensemble Partition Sampling (EPS) for Improved Multi-Class Classification.
IEEE Access.
p. 1.
ISSN 2169-3536
Abstract
Classification is a commonly used technique in data mining and is applied in various fields such as sentiment analysis, fraud detection, and fault diagnosis. Multiclass classification, which involves more than two classes, is more complex than binary classification. There are mainly two ways to approach multiclass classification, one is to expand the binary classifier into a multiclass classifier through various strategies and the other is to divide the multiclass classification problem into multiple binary problems (binarization). Two popular approaches for binarization are One vs One (OvO) and One vs All (OvA). It is simpler to aggregate the outputs of all binary classifiers as the number of classifiers decreases. However, it causes an imbalance of positive and negative sample numbers, which affects the classification effect of each binary classifier. In this article, we contribute to the field of ensemble learning and multi-class classification by proposing a new method called Ensemble Partition Sampling (EPS). This article presents a new approach to multiclass classification using an "Ensemble Partition Sampling" method within the "one-vs-all" (OvA) framework. The primary goal of this method is to tackle the problem of data imbalance by incorporating ensemble learning and preprocessing techniques into each binary dataset. The study found that Ensemble Partition Sampling (EPS) is the most effective method for imbalanced and multiclass imbalanced classification, outperforming other methods including OvA, SMOTE, k-means-SMOTE, Bagging-RB, DES-MI, OvO-EASY, and OvO-SMB. The study used CART, Random Forest, and SVM as classifiers, and the results consistently showed that EPS outperformed all other algorithms. The findings suggest that EPS is a highly effective method for improving classification performance in imbalanced and multiclass imbalanced datasets.
| Item Type: | Article |
|---|---|
| Uncontrolled Keywords: | Ensemble Partition Sampling (EPS); One vs One (OvO); One vs All (OvA); Multi-Class Classification; Imbalanced learning; multiclass imbalanced classification |
| Subjects: | Subjects > Engineering |
| Divisions: | Europe University of Atlantic > Research > Scientific Production Fundación Universitaria Internacional de Colombia > Research > Scientific Production Ibero-american International University > Research > Scientific Production Ibero-american International University > Research > Scientific Production Universidad Internacional do Cuanza > Research > Scientific Production |
| Date Deposited: | 09 May 2023 23:30 |
| Last Modified: | 09 May 2023 23:30 |
| URI: | https://repositorio.unini.edu.mx/id/eprint/7028 |
Actions (login required)
 |
View Item |
en
close
Manuka honey, which is rich in pinocembrin, quercetin, naringenin, salicylic, p-coumaric, ferulic, syringic and 3,4-dihydroxybenzoic acids, has been shown to have pleiotropic effects against colon cancer cells. In this study, potential chemosensitizing effects of Manuka honey against 5-Fluorouracil were investigated in colonspheres enriched with cancer stem cells (CSCs), which are responsible for chemoresistance. Results showed that 5-Fluorouracil increased when it was combined with Manuka honey by downregulating the gene expression of both ATP-binding cassette sub-family G member 2, an efflux pump and thymidylate synthase, the main target of 5-Fluorouracil which regulates the ex novo DNA synthesis. Manuka honey was associated with decreased self-renewal ability by CSCs, regulating expression of several genes in Wnt/β-catenin, Hedgehog and Notch pathways. This preliminary study opens new areas of research into the effects of natural compounds in combination with pharmaceuticals and, potentially, increase efficacy or reduce adverse effects.
Danila Cianciosi mail , Yasmany Armas Diaz mail , José M. Alvarez-Suarez mail , Xiumin Chen mail , Di Zhang mail , Nohora Milena Martínez López mail nohora.martinez@uneatlantico.es, Mercedes Briones Urbano mail mercedes.briones@uneatlantico.es, José L. Quiles mail jose.quiles@uneatlantico.es, Adolfo Amici mail , Maurizio Battino mail maurizio.battino@uneatlantico.es, Francesca Giampieri mail francesca.giampieri@uneatlantico.es,
Cianciosi
<a href="/9698/1/A_Systematic_Survey_of_AI_Models_in_Financial_Market_Forecasting_for_Profitability_Analysis.pdf" class="ep_document_link"><img class="ep_doc_icon" alt="[img]" src="/style/images/fileicons/text.png" border="0"/></a>
en
open
A Systematic Survey of AI Models in Financial Market Forecasting for Profitability Analysis
Artificial intelligence (AI)-based models have emerged as powerful tools in financial markets, capable of reducing investment risks and aiding in selecting highly profitable stocks by achieving precise predictions. This holds immense value for investors, as it empowers them to make data-driven decisions. Identifying current and future trends in multi-class forecasting techniques employed within financial markets, particularly profitability analysis as an evaluation metric is important. The review focuses on examining stud-ies conducted between 2018 and 2023, sourced from three prominent academic databases. A meticulous three-stage approach was employed, encompassing the systematic planning, conduct, and analysis of the se-lected studies. Specifically, the analysis emphasizes technical assessment, profitability analysis, hybrid mod-eling, and the type of results generated by models. Articles were shortlisted based on inclusion and exclusion criteria, while a rigorous quality assessment through ten quality criteria questions, utilizing a Likert-type scale was employed to ensure methodological robustness. We observed that ensemble and hybrid models with long short-term memory (LSTM) and support vector machines (SVM) are being more adopted for financial trends and price prediction. Moreover, hybrid models employing AI algorithms for feature engineering have great potential at par with ensemble techniques. Most studies only employ performance metrics and lack utilization of profitability metrics or investment or trading strategy (simulated or real-time). Similarly, research on multi-class or output is severely lacking in financial forecasting and can be a good avenue for future research.
Bilal Hassan Ahmed Khattak mail , Imran Shafi mail , Abdul Saboor Khan mail , Emmanuel Soriano Flores mail emmanuel.soriano@uneatlantico.es, Roberto García Lara mail , Md. Abdus Samad mail , Imran Ashraf mail ,
Khattak
<a href="/9907/1/sensors-23-09367-v2.pdf" class="ep_document_link"><img class="ep_doc_icon" alt="[img]" src="/style/images/fileicons/text.png" border="0"/></a>
en
open
Internet of Things in Pregnancy Care Coordination and Management: A Systematic Review
The Internet of Things (IoT) has positioned itself globally as a dominant force in the technology sector. IoT, a technology based on interconnected devices, has found applications in various research areas, including healthcare. Embedded devices and wearable technologies powered by IoT have been shown to be effective in patient monitoring and management systems, with a particular focus on pregnant women. This study provides a comprehensive systematic review of the literature on IoT architectures, systems, models and devices used to monitor and manage complications during pregnancy, postpartum and neonatal care. The study identifies emerging research trends and highlights existing research challenges and gaps, offering insights to improve the well-being of pregnant women at a critical moment in their lives. The literature review and discussions presented here serve as valuable resources for stakeholders in this field and pave the way for new and effective paradigms. Additionally, we outline a future research scope discussion for the benefit of researchers and healthcare professionals.
Mohammad Mobarak Hossain mail , Mohammod Abul Kashem mail , Md. Monirul Islam mail , Md. Sahidullah mail , Sumona Hoque Mumu mail , Jia Uddin mail , Daniel Gavilanes Aray mail daniel.gavilanes@uneatlantico.es, Isabel de la Torre Diez mail , Imran Ashraf mail , Md Abdus Samad mail ,
Hossain
<a class="ep_document_link" href="/9908/1/e078815.full.pdf"><img class="ep_doc_icon" alt="[img]" src="/style/images/fileicons/text.png" border="0"/></a>
en
open
Objective The aim was to explore the association of demographic and prehospital parameters with short-term and long-term mortality in acute life-threatening cardiovascular disease by using a hazard model, focusing on elderly individuals, by comparing patients under 75 years versus patients over 75 years of age. Design Prospective, multicentre, observational study. Setting Emergency medical services (EMS) delivery study gathering data from two back-to-back studies between 1 October 2019 and 30 November 2021. Six advanced life support (ALS), 43 basic life support and five hospitals in Spain were considered. Participants Adult patients suffering from acute life-threatening cardiovascular disease attended by the EMS. Primary and secondary outcome measures The primary outcome was in-hospital mortality from any cause within the first to the 365 days following EMS attendance. The main measures included prehospital demographics, biochemical variables, prehospital ALS techniques used and syndromic suspected conditions. Results A total of 1744 patients fulfilled the inclusion criteria. The 365-day cumulative mortality in the elderly amounted to 26.1% (229 cases) versus 11.6% (11.6%) in patients under 75 years old. Elderly patients (≥75 years) presented a twofold risk of mortality compared with patients ≤74 years. Life-threatening interventions (mechanical ventilation, cardioversion and defibrillation) were also related to a twofold increased risk of mortality. Importantly, patients suffering from acute heart failure presented a more than twofold increased risk of mortality. Conclusions This study revealed the prehospital variables associated with the long-term mortality of patients suffering from acute cardiovascular disease. Our results provide important insights for the development of specific codes or scores for cardiovascular diseases to facilitate the risk of mortality characterisation.
Carlos del Pozo Vegas mail , Daniel Zalama-Sánchez mail , Ancor Sanz-Garcia mail , Raúl López-Izquierdo mail , Silvia Sáez-Belloso mail , Cristina Mazas Pérez-Oleaga mail cristina.mazas@uneatlantico.es, Irma Dominguez Azpíroz mail irma.dominguez@unini.edu.mx, Iñaki Elío Pascual mail inaki.elio@uneatlantico.es, Francisco Martín-Rodríguez mail ,
del Pozo Vegas
<a class="ep_document_link" href="/9931/1/fimmu-14-1289032.pdf"><img class="ep_doc_icon" alt="[img]" src="/style/images/fileicons/text.png" border="0"/></a>
en
open
Introduction: Rotavirus infection is a major cause of mortality among children under 5 years in Bangladesh. There is lack of integrated studies on rotavirus prevalence and genetic diversity during 1973 to 2023 in Bangladesh. Methods: This meta-analysis was conducted to determine the prevalence, genotypic diversity and seasonal distribution of rotavirus during pre-vaccination period in Bangladesh. This study included published articles on rotavirus A, rotavirus B and rotavirus C. We used Medline, Scopus and Google Scholar for published articles. Selected literatures were published between 1973 to 2023. Results: This study detected 12431 research articles published on rotavirus. Based on the inclusion criteria, 29 of 75 (30.2%) studies were selected. Molecular epidemiological data was taken from 29 articles, prevalence data from 29 articles, and clinical symptoms from 19 articles. The pooled prevalence of rotavirus was 30.1% (95% CI: 22%-45%, p = 0.005). Rotavirus G1 (27.1%, 2228 of 8219) was the most prevalent followed by G2 (21.09%, 1733 of 8219), G4 (11.58%, 952 of 8219), G9 (9.37%, 770 of 8219), G12 (8.48%, 697 of 8219), and G3 (2.79%, 229 of 8219), respectively. Genotype P[8] (40.6%, 2548 of 6274) was the most prevalent followed by P[4] (12.4%, 777 of 6274) and P[6] (6.4%, 400 of 6274), respectively. Rotavirus G1P[8] (19%) was the most frequent followed by G2P [4] (9.4%), G12P[8] (7.2%), and G9P[8], respectively. Rotavirus infection had higher odds of occurrence during December and February (aOR: 2.86, 95% CI: 2.43-3.6, p = 0.001). Discussion: This is the first meta-analysis including all the studies on prevalence, molecular epidemiology, and genetic diversity of rotavirus from 1973 to 2023, pre-vaccination period in Bangladesh. This study will provide overall scenario of rotavirus genetic diversity and seasonality during pre-vaccination period and aids in policy making for rotavirus vaccination program in Bangladesh. This work will add valuable knowledge for vaccination against rotavirus and compare the data after starting vaccination in Bangladesh.
Nadim Sharif mail , Nazmul Sharif mail , Afsana Khan mail , Irma Dominguez Azpíroz mail irma.dominguez@unini.edu.mx, Raquel Martínez Díaz mail raquel.martinez@uneatlantico.es, Isabel De la Torre Díez mail , Anowar Khasru Parvez mail , Shuvra Kanti Dey mail ,
Sharif