What is it about?
It is widely acknowledged that data analytics models operate on a "Garbage In – Garbage Out" (GI-GO) principle. That is why selecting the relevant features as part of the data analytics process significantly contributes to the model performance. We introduce a hybrid metaheuristic which use both local search algorithm and genetic algorithms for feature subset selection problem.
Featured Image
Why is it important?
In the latest developments, hybrid metaheuristics have demonstrated superior performance compared to traditional standalone metaheuristics across multiple domains. We created a novel hybrid genetic local search algorithm wrapper and evaluated its performance against several existing algorithms using various benchmark datasets. Additionally, we applied this algorithm to an innovation management dataset, showcasing its advantages in a real-world setting for senior management
Perspectives
Wrappers out performs simple filters for the feature subset selection problem. Of course the trade off is the computational time. That's why hybrid wrappers might reduce the computational requirement yet provide good results.
Prof. Kemal Kilic
Sabanci Universitesi
Read the Original
This page is a summary of: A novel Hybrid Genetic Local Search Algorithm for feature selection and weighting with an application in strategic decision making in innovation management, Information Sciences, September 2017, Elsevier,
DOI: 10.1016/j.ins.2017.04.009.
You can read the full text:
Contributors
The following have contributed to this page
