|
| Abstract: |
| In learning and classification problems, feature selection (FS) is critical in finding features that are both meaningful and non-redundant. Today, big data is an integral aspect of all industry sectors. All firms in any industry, such as power, finance, commerce, electronics, communications, and so on, create massive amounts of heterogeneous data that needed to be handled effectively and evaluated correctly. When it comes to big data, feature selection approaches are taken as game-changer since they can assist in minimizing the complexity of genetic data, making it simpler to study and translating it into meaningful information. To enhance classification performance, feature selection is done to remove unnecessary and redundant characteristics from the dataset. In this paper, we presented a novel Grey Wolf Approach based on Quantum leaping neighbor memeplexes which is termed QLGWONM for feature selection and reduction to achieve better classification accuracy. The paper implemented other optimization algorithms such as particle swarm optimization (PSO), slime mould algorithm (SMA), salp swarm algorithm (SSA), artificial butterfly algorithm (ABA), whale optimization (WO), crow search optimization algorithm (CSA), and Jaya models. After the implementation of these algorithms, QLGWONM outperformed other algorithms. The QLGWONM model performed well with an accuracy of 100% for Brain Tumor, CNS, Lung dataset and 97.1% for Ionosphere dataset, and 99% for NSL-KDD. Apart from these, some state-of-art comparisons were also evaluated and QLGWONM gave better results as compared with other existing algorithms. |
| Key words: big data feature extraction machine learning GWO classification |
| DOI：10.11916/j.issn.1005-9113.2022026 |
| Clc Number:TP31 |
| Fund: |
