Web Scraping In Machine Learning

Razvojem interneta, povećala se dostupnost ogromne količine raznolikih podataka na webu, no većina tih informacija je nestrukturirana, što otežava njihovu obradu i analizu. Cilj ovog rada je prikazi mogućnosti automatiziranog prikupljanja i pretvaranja nestrukturiranih podataka iz web stranica u strukturiranu formu,tehnikama i metodama web struganja stvarajući kvalitetan skup podataka potreban za razvoj uspješnih modela strojnog učenja. Rezultati ukazuju na prednosti i nedostatke u razvoju web strugača, ali također i na potencijal pristupa koji bi se mogao ostvariti na reprezentativnom skupu podataka.The development of the Internet has increased the availability of vast amounts of diverse data on the web. However, the majority of this information is unstructured, which presents a challenge in its processing and analysis. The goal of this paper is to demonstrate the possibilities of automated data collection and conversion of unstructured web data into structured data using web scraping techniques and methods. This creates a high-quality dataset necessary for the development of successful machine learning models. The results discuss the advantages and disadvantages of web scraping development, as well as the potential approach that can be achieved on a representative dataset