An integrated method for identifying garment attribute trends through social networks based on machine learning

Purpose This paper proposes an integrated method for predicting the value of a garment’s trendy attributes in the upcoming fashion season. Design/methodology/approach The paper’s methodology includes web crawling and data scraping from the online social network Pinterest to collect opinions on garment attributes. Machine learning models are then employed to classify users as trustworthy based on their opinions. Subsequently, the comments of trustworthy users are analyzed to identify trending garment attributes for the upcoming fashion season. This approach offers valuable insights for fashion designers during the garment design process. Findings The results of the proposed study showed that the proposed method is helpful for fashion designers in the initial stages of the garment design process. Originality/value The paper introduces an integrated method that uses machine learning to predict garment trends based on social media data. It also presents a trust index to identify reliable trendsetters. The paper demonstrates the method’s practical applicability by validating the identified trends with sales data. Additionally, it suggests using the identified trends in decision support systems for garment design.