Weighted Random k Satisfiability for k=1,2 (r2SAT) in Discrete Hopfield Neural Network

Applied Soft Computing, Journal Year: 2022, Volume and Issue: 126, P. 109312 - 109312

Published: July 16, 2022

Language: Английский

---

A modified reverse-based analysis logic mining model with Weighted Random 2 Satisfiability logic in Discrete Hopfield Neural Network and multi-objective training of Modified Niched Genetic Algorithm

Expert Systems with Applications, Journal Year: 2023, Volume and Issue: 240, P. 122307 - 122307

Published: Oct. 28, 2023

Language: Английский

---

Novel logic mining incorporating log linear approach

Journal of King Saud University - Computer and Information Sciences, Journal Year: 2022, Volume and Issue: 34(10), P. 9011 - 9027

Published: Aug. 28, 2022

Mining the best logical rule from data is a challenging task because not all attribute of dataset will contribute towards optimal representation. Even if correct attributes were selected, wrong connection in formula lead to suboptimal representation datasets. These two factors must be carefully considered creating more robust logic mining method. In this paper, we proposed novel by introducing log-linear analysis select which formulate that embedded into energy-based ANN named Discrete Hopfield Neural Network (DHNN). phase, test association for each carried out where have significant level less than α selected before proceeding phase. By using DHNN, via learned and retrieved induced with classification ability. The hybrid model has been tested various real-life datasets was compared several established methods. Based on findings, winning points dominates 3 metrics 5 average rank. achieve highest rank are Accuracy (1.800), Precision (3.500), Mathews Correlation Coefficient (2.700). accordance experimental result obtained, achieved performance statistically p-value. Hence, these findings an advancement existing statistical

Language: Английский

---

AutoML-GWL: Automated machine learning model for the prediction of groundwater level

Engineering Applications of Artificial Intelligence, Journal Year: 2023, Volume and Issue: 127, P. 107405 - 107405

Published: Nov. 3, 2023

Language: Английский

---

Multi-unit Discrete Hopfield Neural Network for higher order supervised learning through logic mining: Optimal performance design and attribute selection

Journal of King Saud University - Computer and Information Sciences, Journal Year: 2023, Volume and Issue: 35(5), P. 101554 - 101554

Published: April 18, 2023

In the perspective of logic mining, attribute selection, and objective function best is two main factors that identifies effectiveness our proposed mining model. The non-significant attributes selected will cause Discrete Hopfield Neural Network to learned obtain wrong synaptic weight. Thus, this result suboptimal solution. Although we might select correct attributes, conventional limits search space obtained more induced during retrieval phase Network. Therefore, paper proposes a novel by integrating statistical analysis in pre-processing ensure only optimal be selected. Supervised learning approach via correlation implemented for purpose selection. Additionally, permutation operator serves enhance probability higher order satisfiability logical rule satisfied having finite arrangement attributes. During phase, multi-unit which leads efficiency model tested on 15 real-life datasets comparing performance with existing works using five metrics including accuracy, sensitivity, precision, Matthews Correlation Coefficient (MCC) F1 Score. According results, has its own strength dominating most average rank metrics. This demonstrates can differentiate across all domains confusion matrix. p-value based five-performance indicate there significantly difference between since value accuracy (0.000), sensitivity (0.001), precision score (0.000) MCC are less than 0.05. finding statistically prove effective compared mining.

Language: Английский

---

Exploring machine learning algorithms for accurate water level forecasting in Muda river, Malaysia

Heliyon, Journal Year: 2023, Volume and Issue: 9(7), P. e17689 - e17689

Published: June 30, 2023

Accurate water level prediction for both lake and river is essential flood warning freshwater resource management. In this study, three machine learning algorithms: multi-layer perceptron neural network (MLP-NN), long short-term memory (LSTM) extreme gradient boosting XGBoost were applied to develop forecasting models in Muda River, Malaysia. The developed using limited amount of daily meteorological data from 2016 2018. Different input scenarios tested investigate the performance models. results evaluation showed that MLP model outperformed LSTM predicting levels, with an overall accuracy score 0.871 compared 0.865 0.831 XGBoost. No noticeable improvement has been achieved after incorporating into Even though lowest reported was by XGBoost, it faster algorithms due its advanced parallel processing capabilities distributed computing architecture. terms different time horizons, found be more accurate than when 7 days ahead, demonstrating superiority capturing long-term dependencies. Therefore, can concluded each ML own merits weaknesses, differs on case because these depends largely quantity quality available training.

Language: Английский

---

Log-Linear-Based Logic Mining with Multi-Discrete Hopfield Neural Network

Mathematics, Journal Year: 2023, Volume and Issue: 11(9), P. 2121 - 2121

Published: April 30, 2023

Choosing the best attribute from a dataset is crucial step in effective logic mining since it has greatest impact on improving performance of induced logic. This can be achieved by removing any irrelevant attributes that could become logical rule. Numerous strategies are available literature to address this issue. However, these approaches only consider low-order rules, which limit connection clause. Even though some methods produce excellent metrics, incorporating optimal higher-order rules into challenging due large number involved. Furthermore, suboptimal trained an ineffective discrete Hopfield neural network, leads In paper, we propose log-linear analysis during pre-processing phase, multi-unit 3-satisfiability-based reverse with approach. The proposed also integrates network ensure each 3-satisfiability learned separately. context, our employs three unique optimization layers improve final Extensive experiments conducted 15 real-life datasets various fields study. experimental results demonstrated method outperforms state-of-the-art terms widely used metrics.

Language: Английский

---

Dual optimization approach in discrete Hopfield neural network

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

Published: July 7, 2024

Language: Английский

---

S-Type Random k Satisfiability Logic in Discrete Hopfield Neural Network Using Probability Distribution: Performance Optimization and Analysis

Mathematics, Journal Year: 2023, Volume and Issue: 11(4), P. 984 - 984

Published: Feb. 15, 2023

Recently, a variety of non-systematic satisfiability studies on Discrete Hopfield Neural Networks have been introduced to overcome lack interpretation. Although flexible structure was established assist in the generation wide range spatial solutions that converge global minima, fundamental problem is existing logic completely ignores probability dataset’s distribution and features, as well literal status distribution. Thus, this study considers new type termed S-type Random k Satisfiability, which employs creative layer Network, plays significant role identification prevailing attribute likelihood binomial dataset. The goal phase establish logical assign negative literals based two given statistical parameters. performance proposed investigated using comparison metric current state-of-the-art rules; consequently, found models high value parameters efficiently introduce phase. Additionally, by implementing it has observed cost function experiences reduction. A form synaptic weight assessment via methods applied investigate effect structure. Overall, investigation demonstrated controlling good management minima solutions.

Language: Английский

---

MTS-PRO2SAT: Hybrid Mutation Tabu Search Algorithm in Optimizing Probabilistic 2 Satisfiability in Discrete Hopfield Neural Network

Mathematics, Journal Year: 2024, Volume and Issue: 12(5), P. 721 - 721

Published: Feb. 29, 2024

The primary objective of introducing metaheuristic algorithms into traditional systematic logic is to minimize the cost function. However, there a lack research on impact function under different proportions positive literals. In order fill in this gap and improve efficiency algorithm logic, we proposed based mutation tabu search embedded it probabilistic satisfiability discrete Hopfield neural networks. Based algorithm, operators genetic were combined its global ability during learning phase ensure that converged zero at Additionally, further optimization was carried out retrieval enhance diversity solutions. Compared with nine other exhaustive algorithms, superior terms time complexity convergence, showed higher solutions binary space, consolidated phase, significantly improved solution logic.

Language: Английский

---