Evaluation of the effectiveness of the use of NLP tools and AI systems for the analysis of advertisements in the Internet advertising exchange systems

Abstract

The study determines the effectiveness of using natural language processing tools and artificial intelligence systems for analyzing advertising campaigns in the Internet advertising exchange systems. The article discusses which tools can be used to detect keywords in ad text, as well as how these tools can be combined with custom machine learning models to detect fraudulent and malicious information in ad exchange web servers. The article illustrates which metrics can be used to evaluate ad content for unwanted content using modern artificial intelligence systems. An analysis of existing tools and their results is conducted using a real high-risk advertisement. A detailed report is provided according to different evaluation metrics. The expediency of integrating the above technologies into the business logic of advertising networks is determined.

Keywords

• detection
• online advertisements
• malicious content
• fraudulent content
• the internet advertising exchange systems
• advertising campaigns