| Sumario: | With the huge upsurge in the volume of information accessible on the World Wide Web (WWW) nowadays, and the development need for an above method to access this information, there was a powerful resurgence of interest in web mining research. Web mining is a dangerous problem in data mining and other information process methods to determine valuable patterns. With the upsurge in the amount of websites and web users, the necessity for the classification of website increases attraction. The classification of the website based on URLs only plays a vital role, as the contents of web pages are not necessarily attained for classification. Nowadays, Machine learning (ML) and Deep learning (DL) can be significant in finding known and novel malicious URLs. Various types of research are performed on malicious URL classification and detection utilizing different ML techniques. In this manuscript, we present a Website Classification using the Bidirectional Gated Attention Recurrent Neural Network with Improved Coati Optimization (WCBGARNNICO) technique. The key intention of the WCBGARNN-ICO method is to identify websites in web mining. At first, the WCBGARNN-ICO model applies the quality of text preprocessing with different levels to attain clear data and extract significant data. For the extraction of the feature process, the bidirectional encoder representations from the transformers (BERT) method can be exploited. Besides, the attention-based bidirectional gated recurrent neural network (A-BGRNN) model is employed for the website classification process. At last, the improved coati optimization algorithm (ICOA) is deployed to fine-tune the hyperparameter of the A-BGRNN technique. The experimental study of the WCBGARNN-ICO algorithm is tested on a benchmark database and the findings are measured with numerous measures. The experimental findings highlighted the development of the WCBGARNN-ICO system over other approaches.
|
|---|