Unlocking Fashion Trends: How Pusan National University Researchers Are Leveraging AI for Accurate Predictions

Fashion trend forecasting is crucial for brands to anticipate what styles will resonate with consumers in the upcoming seasons. Traditionally, this relied heavily on the insights of fashion experts. But now, with the inclusion of big data analytics, companies can gain a deeper understanding of consumer behavior. However, accessing these advanced tools can be challenging for students and smaller brands.

Enter artificial intelligence. Tools like ChatGPT can simplify data analysis, making it more accessible. These large language models draw from extensive cultural and societal information, which can help in predicting fashion trends. Despite their potential, they also have limitations, such as generating incorrect information, making it essential to verify their findings carefully.

A recent study by Yoon Kyung Lee, an assistant professor at Pusan National University, explored how ChatGPT can aid in predicting fashion trends. “Instead of asking, ‘What will be popular?’ we designed a systematic strategy to get more detailed and accurate answers,” Lee said. This innovative method contrasts traditional forecasting, which often lacks structure.

Lee and her team applied a new technique called Top-Down Prompting (TDP), inspired by brainstorming methods. They began with a broad question about future fashion trends, then broke it down into specific topics like silhouette, materials, and colors.

Using this approach, they predicted men’s fashion trends for fall/winter 2024 and compared the results to reports from the Official Fashion Trend Information Company (OFTIC). While they found ChatGPT’s predictions largely echoed established trends, it also highlighted emerging themes like gender fluidity, which is becoming increasingly prominent in the fashion world.

“Even though the accuracy of ChatGPT’s predictions was low, it offered insights into new trends not captured before,” Lee noted. This suggests that AI could sense cultural shifts and create new pathways in fashion design.

Although ChatGPT might not replace expert analysis, it can be a valuable tool for emerging designers and students, allowing them to gain a more nuanced understanding of trends. By combining AI with expert insights, the study proposes a more integrated approach to fashion forecasting.

This research reflects a growing trend in various industries where AI tools are transforming traditional methods. As AI continues to evolve, it may become an indispensable part of fashion forecasting, making it more systematic and accessible for all parties involved.

For further insights on this topic, check the detailed findings published in the Clothing and Textiles Research Journal.

Source link