Comparison of machine learning algorithms of internet publications classification

Barsolevskaya Anna F., Kondrashkin Dmitry A., Samoylov Vyacheslav E., Tsaregorodtsev Anatoliy V.

The Internet is a vast source of information on Earth. You can find almost everything you want online. Information on the Internet is displayed in various formats and types, including text documents, videos and photos. Nonetheless, gathering useful information without using some web-based utilities can be a daunting task. In this scenario, web mining may come in handy. This method provides a tool that simplifies the extraction process of the required data from Internet resources. Many studies have focused on the issue of highly accurate classification of web pages and publications. This study evaluates several supervised learning algorithms to identify categories and classify social media posts. During the course of the research, following machine learning algorithms to compare the effectiveness of solving social network user publication classification problems were used: Random Forest, Neural Networks, Dimensionality Reduction, AdaBoost.

Keywords:classification of Internet user publications; social networks; classification of web pages; data analysis; random forest; neural networks; dimensionality reduction; AdaBoost.