Different mutation and crossover set of genetic programming in an automated machine learning

Masrom, Suraya; Mohamad, Mohd Saberi; Hatim, Shahirah Mohamed; Baharun, Norhayati; Omar, Nasiroh; Rahman, Abdullah Sani Abd

doi:10.11591/ijai.v9.i3.pp402-408

Cited by 6 publications

(5 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Generation is the number of iterations for the optimization search while population is the number of maximum individual (pipelines) to be randomly selected in the optimization search. The findings in [29] suggested that default configuration by TPOT itself was able to generate good prediction results when tested on some common benchmark problems used in the study. The higher of population numbers is expected to provide more probability for the TPOT to select achieve better optimal models but it is depending the tested dataset, which should be observed in this research.…”

Section: Methodsmentioning

confidence: 85%

Comparisons of automated machine learning (AutoML) in predicting whistleblowing of academic dishonesty with demographic and theory of planned behavior

Rahman,

Masrom,

Mohamad

et al. 2023

MethodsX

View full text Add to dashboard Cite

Section: Methodsmentioning

confidence: 85%

Comparisons of automated machine learning (AutoML) in predicting whistleblowing of academic dishonesty with demographic and theory of planned behavior

Rahman,

Masrom,

Mohamad

et al. 2023

MethodsX

View full text Add to dashboard Cite

“…Genetic algorithm (GA) was introduced by John Holland in 1975 [101] mimics the natural concepts, which are genetic to represent the solution and selection, crossover, mutation to perform its operation. At each phase, GAs employ three distinct sorts of rules to generate the next generation from the present population: selection, crossover, and mutation [93], [95], [102]- [113]. Selection rules determine which individuals, referred to as parents, contribute to the population in the following generation.…”

Section: Genetic Algorithmmentioning

confidence: 99%

Comparison of meta-heuristic algorithms for fuzzy modelling of COVID-19 illness’ severity classification

Aseri

Ismail

Fakharudin

et al. 2022

IJ-AI

View full text Add to dashboard Cite

The world health organization (WHO) proclaimed the COVID-19, commonly known as the coronavirus disease 2019, was a pandemic in March 2020. When people are in close proximity to one another, the virus spreads mostly through the air. It causes some symptoms in the affected person. COVID-19 symptoms are quite variable, ranging from none to severe sickness. As a result, the fuzzy method is seen favourably as a tool for determining the severity of a person’s COVID-19 sickness. However, when applied to a large situation, manually generating a fuzzy parameter is challenging. This could be because of the identification of a large number of fuzzy parameters. A mechanism, such as an automatic procedure, is consequently required to identify the right fuzzy parameters. The metaheuristic algorithm is regarded as a viable strategy. Five meta-heuristic algorithms were analyzed and utilized in this article to classify the severity of COVID-19 sickness data. The performance of the five meta-heuristic algorithms was evaluated using the COVID-19 symptoms dataset. The COVID-19 symptom dataset was created in accordance with WHO and the Indian ministry of health and family welfare criteria. The findings provide the average classification accuracy for each approach.

show abstract

“…Lim et al [32] discusses both exploitation and exploration zones for crossover operators and their effects on GA's performance. Masrom et al [33] discusses various parametric settings and concludes that a higher crossover rate provides better results.…”

Section: Figure 1 Pseudo Code For Genetic Algorithmmentioning

confidence: 99%

Analysis and comparison of a proposed mutation operator and its effects on the performance of genetic algorithm

Ullah

Salam²,

Masood³

2022

IJEECS

View full text Add to dashboard Cite

Genetic algorithms (GAs) are dependent on various operators and parameters. The most common evolutionary operators are parent selection, crossover, and mutation. Each operator has broad implementations with its pros and cons. A successful GA is highly dependent on genetic diversity which is the main driving force that steers a GA towards an optimal solution. Mutation operator implements the idea of exploration to search for uncharted areas and introduces diversity in a population. Thus, increasing the probability of GA to converge to a globally optimum solution. In this paper, a new variant of mutation operator is proposed, and its functions are studied and compared with the existing operators. The proposed mutation operator as well as others such as m-mutation, shuffle, swap, and inverse are tested for their ability to introduce diversity in population and hence, their effects on the performance of GA. All these operators are applied to Max one problem. The results concluded that the proposed variant is far more superior to the existing operators in terms of introducing diversity and hence early convergence to an optimum solution.

show abstract

Different mutation and crossover set of genetic programming in an automated machine learning

Cited by 6 publications

References 21 publications

Comparisons of automated machine learning (AutoML) in predicting whistleblowing of academic dishonesty with demographic and theory of planned behavior

Comparisons of automated machine learning (AutoML) in predicting whistleblowing of academic dishonesty with demographic and theory of planned behavior

Comparison of meta-heuristic algorithms for fuzzy modelling of COVID-19 illness’ severity classification

Analysis and comparison of a proposed mutation operator and its effects on the performance of genetic algorithm

Contact Info

Product

Resources

About