A new population initialisation method based on the Pareto 80/20 rule for meta‐heuristic optimisation algorithms

Hasanzadeh, Mohammad; Keynia, Farshid

doi:10.1049/sfw2.12025

Cited by 9 publications

(2 citation statements)

References 77 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…If the initial population of GA is better, then the probability of obtaining good solutions is high (5) . Randomly selected initial population takes more time to converge but heuristics based initial population converges rapidly (6) . Zhou et al (7) have experimented an automatic clustering of low-dimensional data using k-means algorithm where GA based initial seed pattern is used for the clustering.…”

Section: Introductionmentioning

confidence: 99%

Genetic Algorithm Based Over-Sampling with DNN in Classifying the Imbalanced Data Distribution Problem

Karthikeyan¹,

Kathirvalavakumar²

2023

IJST

View full text Add to dashboard Cite

Objective: Data imbalance exists in many real-life applications. In the imbalanced datasets, the minority class data creates a wrong inference during the classification that leads to more misclassification. More research has been done in the past to solve this issue, but as of now there is no global working solution found to do efficient classification. After analyzing various existing literatures, it is proposed to minimize the misclassification through genetic based oversampling and deep neural network (DNN) classifier. Method: In the proposed oversampling method synthetic samples are generated based on genetic algorithm. Initial populations for the genetic algorithm are generated using Gaussian weight initialization technique and the fittest individual from the population are selected by Euclidean distance for further processing to generate synthetic data in double the minority class size and the dataset is classified with the DNN. Findings: The performance of the oversampled training data with DNN Classifier is compared with C4.5 and Support Vector Machine (SVM) classifiers and found that the DNN classifier outperforms the other two classifiers. The data generated using SMOTE and ADASYN are considered for comparison. It is found that the proposed approach outperforms the other approaches. It is also proved from the experiment that misclassification is reduced and the proposed method is statistically significant and is comparatively better. Novelty: Initial population generation by Gaussian weight initialization, the fittest sample selection by Euclidean distance measure, synthetic samples with double the minority class size and DNN for classification to reduce the misclassification is novelty in this work.

show abstract

Section: Introductionmentioning

confidence: 99%

Genetic Algorithm Based Over-Sampling with DNN in Classifying the Imbalanced Data Distribution Problem

Karthikeyan¹,

Kathirvalavakumar²

2023

IJST

View full text Add to dashboard Cite

show abstract

“…Initialisation in meta-heuristic optimisers is usually done randomly based on the principle of uniform search space coverage [8][9][10][11][12][13][14]. The results of some studies have shown that different initialisation methods lead to different accuracies for different problems [15][16][17]. Exploration in optimisation refers to the process of extensively exploring promising spaces from the possible search space, and exploitation refers to the search capability around promising areas of the exploration phase.…”

Section: Introductionmentioning

confidence: 99%

A new index‐based hyper‐heuristic algorithm for global optimisation problems

2022

Self Cite

View full text Add to dashboard Cite

In this research study, a new combination search algorithm, based on indexing its constituent processes, is proposed to solve global optimisation problems. As optimisation problems become more complex, especially real‐world problems, the use of higher‐performance algorithms has become essential. One of the techniques that lead to design of an algorithm with stronger strategies in exploration and exploitation, as well as a better balance between these two strategies, is the appropriate combination of parent algorithm processes. The proposed algorithm was developed using a new innovation with the help of the supply‐demand‐based optimisation and the Harris hawks optimisation algorithms processes as parent algorithms. In this algorithm, the local and global search sections of its parent algorithms, are separated, and then based on a new indexing method in each iteration, according to the current population indexing, a global search and a local search are selected from the processes of its parent algorithms, and then the current population is updated with two selected sections. The performance and effectiveness of the proposed algorithm in solving well‐known standard benchmark problems and in solving real‐world engineering problems have been tested and validated by statistical tools. The results of the research study show that the proposed algorithm can provide very effective results compared to other competing algorithms as well as its parent algorithms in many tests. The results show that the proper combination of optimisation algorithm processes can be used as a technique to design more powerful algorithms to solve global optimisation problems, especially complex real‐world problems.

show abstract

A Multi-population-Based Algorithm with Different Ways of Subpopulations Cooperation

Cpałka

Łapa

Rutkowski

2023

Lecture Notes in Computer Science

View full text Add to dashboard Cite

A new population initialisation method based on the Pareto 80/20 rule for meta‐heuristic optimisation algorithms

Cited by 9 publications

References 77 publications

Genetic Algorithm Based Over-Sampling with DNN in Classifying the Imbalanced Data Distribution Problem

Genetic Algorithm Based Over-Sampling with DNN in Classifying the Imbalanced Data Distribution Problem

A new index‐based hyper‐heuristic algorithm for global optimisation problems

A Multi-population-Based Algorithm with Different Ways of Subpopulations Cooperation

Contact Info

Product

Resources

About