Model for identifying fraud in product sales postings with Scraping on Facebook

Abstract

The article presents a model for identifying fraud in product sales posts in Facebook groups, using scraping and natural language processing (NLP) techniques. With the growth of e-commerce in social networks, fraud cases have increased, which motivates the need for effective solutions. The model starts with data mining of product posts on Facebook buy/sell groups using Python libraries such as BeautifulSoup and Selenium, this data is then processed and analyzed using NLP techniques supported by (gpt-3.5-turbo-instruct) to identify patterns and evaluate the relationship between items and comments. The model employs Cronbach's alpha coefficient to validate the internal consistency of the evaluations and uses anomaly detection to identify unusual patterns that could indicate fraud. The results show that the model is effective in identifying fraud, offering a solution tailored to the specific characteristics of Facebook. The integration of scraping and NLP provides a valuable tool to improve fraud detection accuracy, contributing significantly to the field of business intelligence and strengthening trust in e-commerce on social networks.