Butterfly Optimization based Feature Selection with Long-short Term Memory for Web Page Crawling

Info recovery, data removal, manuscript removal, network explore, and connected under-standing all struggle with organising the enormous number of text documents. Due to rapid expansion growth of the World Wide Web, web crawling has become increasingly promi-nent. Website search engines face challenges in efficiently accessing a vast array of online content. A focused crawler is a specialized type of web crawler designed to specifically gather web pages from the Internet. The atmosphere of online sites, however, can readily affect how well the focused crawling works. This research paper presents an automated web scraping technique that blends a feature selection model with deep learning. The pre-processed data is then utilised to train a text-based Convolutional Neural Network (CNN) model, which is then used to extract features. The characteristics are then picked in the most optimal manner using the Butterfly Optimisation Algorithm (BOA) model. way. The chosen traits are then categorised as Long-Short Term Memory (LSTM). In our studies, the suggested model is assessed using the WebKB benchmark dataset. Our experimental findings demonstrated that the suggested model was frequently able to learn the appropriate decision surface using the area under the Precision-Recall Curve as the performance met-ric.