Crawling is defined as the automated and systematic collection of useful web pages with an interconnecting link structure. Live web crawling that is dependent on human, expert-nominated seed URIs is employed for the general construction of event collections. In the focused web crawling technique, the crawler is guided by reference content that pertains to a particular event. With this technique, a focused crawler will attempt to predict if a target URL is directed towards a relevant web page prior to retrieving it. However, it can get trapped within a limited web community and overlook relevant web pages outside its track. For web crawling optimization as well as for the selection of more relevant web pages for retrieval by the crawler, a metaheuristic algorithm is employed. A global searching algorithm’s modification can address the problems. The utilization of basic heuristic algorithms and prohibition techniques forms the basis of the Tabu Search (TS). Swarm intelligence is used by the Particle Swarm Optimization (PSO) to identify an optimization problem’s solution. This is done by a generation of better candidate feature subsets as per a given fitness criteria and rapid convergence towards a global optimum within a set number of iterations. Hence, the proposed gradient-based algorithm will present a more rapid convergence for the PSO than the other variants. A proposal for a hybrid Stochastic Gradient Descent-Particle Swarm Optimization has also been presented in this work.