Estimation of HbA1c for DMT2 risk prediction on the Mexican population based in Artificial Neural Networks DOI Creative Commons
Alexis Alonso-Bastida,

Marisol Cervantes-Bobadilla,

Dolores Azucena Salazar-Piña

et al.

Journal of King Saud University - Computer and Information Sciences, Journal Year: 2023, Volume and Issue: 36(1), P. 101905 - 101905

Published: Dec. 31, 2023

In this paper, the main objective is to estimate percentage of glycosylated hemoglobin through an easily accessible computational platform risk generating type 2 diabetes mellitus in Mexican population. The estimation tool developed artificial neural network model, which was trained and validated according a population sample 1120 people between 18 59 years old. model inputs were gender, age, body mass index, waist circumference, weekly food consumption, family history, whether person suffers from any chronic degenerative disease other than T2DM. We used as output, estimated dynamic glucose model. results present coefficient determination 99%, demonstrating acceptable performance aid for health personnel, seeks generate first approximation glycemic status those communities with high marginalization index prevention strategies.

Language: Английский

In search of the best fitness function for optimum generation of trajectories for Automated Guided Vehicles DOI Creative Commons
Eduardo Bayona, Jesús Enrique Sierra-García, Matilde Santos

et al.

Engineering Applications of Artificial Intelligence, Journal Year: 2024, Volume and Issue: 133, P. 108440 - 108440

Published: April 19, 2024

This paper presents an offline optimization method designed for use with industrial robots in environments static obstacles. It is particularly useful industry where stability and predictability are crucial to meeting expected timelines automated guided vehicle operations. The main methodological contribution of this work lies the integral process used define effective fitness function that guides search optimal solutions. cost plays a critical role effectiveness trajectory tracking algorithm by quantifying path quality allowing comparisons between design poses challenges including accuracy, suitability, minimization length, avoiding or reducing collisions. To achieve objectives address some issues such as sensitivity parameter scaling risk premature convergence, different approaches can be used. proposes incorporate constraints into function, adjust parameters reflect conditions problem, based on prior knowledge accurate representation goals. three relevant contributions planning routes following. Firstly, development mathematical model trajectories Frenet curves considers occupancy map environment. Second, strategy generate safe paths. Finally, towards solutions considering resolution parameters. study exhaustive analysis functions obtained, each one evaluated key metrics length trajectory, average minimum distance map, number collisions along path. results show obtained successfully avoids environment all scenarios consistently remains largest obstacles, at least 50% higher than other study.

Language: Английский

Citations

7
Dual optimization approach in discrete Hopfield neural network DOI
Yueling Guo, Nur Ezlin Zamri, Mohd Shareduwan Mohd Kasihmuddin

et al.

Applied Soft Computing, Journal Year: 2024, Volume and Issue: 164, P. 111929 - 111929

Published: July 7, 2024

Language: Английский

Citations

6
Combined analysis of thermofluids and electromagnetism using physics-informed neural networks DOI
Yeonhwi Jeong, Junhyoung Jo,

Tonghun Lee

et al.

Engineering Applications of Artificial Intelligence, Journal Year: 2024, Volume and Issue: 133, P. 108216 - 108216

Published: March 12, 2024

Language: Английский

Citations

5
Higher order Weighted Random <i>k</i> Satisfiability ($k = 1, 3$) in Discrete Hopfield Neural Network DOI Creative Commons
Xiaoyan Liu, Mohd Shareduwan Mohd Kasihmuddin, Nur Ezlin Zamri

et al.

AIMS Mathematics, Journal Year: 2025, Volume and Issue: 10(1), P. 159 - 194

Published: Jan. 1, 2025

<p>Researchers have explored various non-systematic satisfiability approaches to enhance the interpretability of Discrete Hopfield Neural Networks. A flexible framework for has been developed investigate diverse logical structures across dimensions and improved lack neuron variation. However, logic phase this approach tends overlook distribution characteristics literal states, ratio negative literals not mentioned with higher-order clauses. In paper, we propose a new named Weighted Random $k$ Satisfiability ($k = 1, 3$), which implements in The proposed logic, integrated into Network, established structure by incorporating during phase. This enhancement increased network's storage capacity, improving its ability handle complex, high-dimensional problems. advanced was evaluated learning metrics. When values were $r 0.2$, 0.4, 0.6, 0.8, demonstrated potential better performances smaller errors. Furthermore, performance positive impact on management synaptic weights. results indicated that optimal global minimum solutions are achieved when set 0.8$. Compared state-of-the-art structures, novel more significant achieving solutions, particularly terms literals.</p>

Language: Английский

Citations

0
Learning characteristic parameters and dynamics of centrifugal pumps under multiphase flow using physics-informed neural networks DOI
Felipe de Castro Teixeira Carvalho, Kamaljyoti Nath, Alberto Luiz Serpa

et al.

Engineering Applications of Artificial Intelligence, Journal Year: 2024, Volume and Issue: 138, P. 109378 - 109378

Published: Oct. 2, 2024

Language: Английский

Citations

3
Logic mining method via hybrid discrete hopfield neural network DOI
Yueling Guo, Mohd Shareduwan Mohd Kasihmuddin, Nur Ezlin Zamri

et al.

Computers & Industrial Engineering, Journal Year: 2025, Volume and Issue: unknown, P. 111200 - 111200

Published: May 1, 2025

Language: Английский

Citations

0
Data mining algorithm in the identification of accounting fraud by smart city information technology DOI Creative Commons
Xinyi Zheng, Mohamad Ali Abdul Hamid,

Y. Hou

et al.

Heliyon, Journal Year: 2024, Volume and Issue: 10(9), P. e30048 - e30048

Published: April 27, 2024

The identification of accounting fraud is an important measure to safeguard the interests stakeholders and ensure long-term development company. current traditional methods for identifying rely on manual review judgment, lacking objectivity accuracy. In order improve accuracy identification, efficiency objectivity, this article combines smart city information technology conduct in-depth research data mining algorithms identification. This first provides a brief overview cities technology, then introduces basic theory finally implements through k-means clustering algorithm. divided into k clusters, abnormal clusters are identified by checking characteristics attributes each cluster. Compared with rule-based pattern based methods, approach can more flexibly adapt different types forms fraud, discover unknown patterns fraud. experiment, used electronic collection, analysis, retrieval systems websites Shanghai Stock Exchange Shenzhen collect 641 annual reports financial from 62 listed companies that engaged in statement 84 were not reported have 2012 2021 as test samples. results tested analyzed several aspects, including number misjudgments, misjudgment rate, ROC curve. final show compared comprehensive rate has decreased 3 %. conclusion indicates identify help audit effectiveness.

Language: Английский

Citations

2
Unsupervised logic mining with a binary clonal selection algorithm in multi-unit discrete Hopfield neural networks via weighted systematic 2 satisfiability DOI Creative Commons
Nurul Atiqah Romli, Nur Fariha Syaqina Zulkepli, Mohd Shareduwan Mohd Kasihmuddin

et al.

AIMS Mathematics, Journal Year: 2024, Volume and Issue: 9(8), P. 22321 - 22365

Published: Jan. 1, 2024

<p>Evaluating behavioral patterns through logic mining within a given dataset has become primary focus in current research. Unfortunately, there are several weaknesses the research regarding models, including an uncertainty of attribute selected model, random distribution negative literals logical structure, non-optimal computation best logic, and generation overfitting solutions. Motivated by these limitations, novel model incorporating mechanism to control literal systematic Satisfiability, namely Weighted Systematic 2 Satisfiability Discrete Hopfield Neural Network, is proposed as structure represent behavior dataset. For we used ratio <italic>r</italic> structures prevent solutions optimize synaptic weight values. A new computational approach considering both true false classification values learning system was applied this work preserve significant Additionally, unsupervised techniques such Topological Data Analysis were ensure reliability attributes model. The comparative experiments models utilizing 20 repository real-life datasets conducted from repositories assess their efficiency. Following results, dominated all metrics for average rank. ranks each metric Accuracy (7.95), Sensitivity (7.55), Specificity (7.93), Negative Predictive Value (7.50), Mathews Correlation Coefficient (7.85). Numerical results in-depth analysis demonstrated that consistently produced optimal induced represented performance study.</p>

Language: Английский

Citations

2
Optimizing Artificial Neural Network Learning Using Improved Reinforcement Learning in Artificial Bee Colony Algorithm DOI Creative Commons
Taninnuch Lamjiak, Booncharoen Sirinaovakul,

Siriwan Kornthongnimit

et al.

Applied Computational Intelligence and Soft Computing, Journal Year: 2024, Volume and Issue: 2024(1)

Published: Jan. 1, 2024

Artificial neural networks (ANNs) are widely used machine learning techniques with applications in various fields. Heuristic search optimization methods typically to minimize the loss function ANNs. However, these can lead network become stuck local optima, limiting performance. To overcome this challenge, study introduces an improved approach, improvement of reinforcement artificial bee colony (improved R‐ABC) algorithm, enhance process for The proposed method aims limitations heuristic and improve efficiency weight adjustment This new approach enhances discovery phase traditional R‐ABC by including parameters neighboring food sources, augmenting capabilities finding optimal solution. performance was compared ANNs utilizing backpropagation stochastic gradient descent (SGD) Adam optimizers, as well other swarm intelligence (SI) such particle (PSO) R‐ABC. results showed that both PSO continuously solutions across all benchmark datasets. In iris dataset, SI approaches consistently achieved F 1‐scores exceeding 0.94, outperforming SGD Adam. For datasets, generally outperformed methods. indicate when is applied ANNs, it outperforms optimization, especially size expands. Although faster execution times TensorFlow, suggests using model accuracy efficiency. Advanced increase ability obtain solutions. Enhanced algorithms significantly ANN training efficiency, complex high‐dimensional

Language: Английский

Citations

2
J-type random 2,3 satisfiability: a higher-order logical rule in discrete hopfield neural network DOI
Xiaofeng Jiang, Mohd Shareduwan Mohd Kasihmuddin, Yueling Guo

et al.

Evolutionary Intelligence, Journal Year: 2024, Volume and Issue: 17(5-6), P. 3317 - 3336

Published: April 24, 2024

Language: Английский

Citations

1