Development of a Machine Learning Model for Prediction of Colour Trends in Fashion Industry

In the fashion industry, good trend prediction is the key to success for both manufacturers and retailers. Currently, the speed at which the industry is working is faster than ever before. The Colour forecasting process in many fashion organizations is not visible to the public. A forecasting method is proposed in this study to provide colour trends to industries in advance. Predicting colour trends would allow retailers to improve their logistics for the storage/shipping of clothes. This research proposes a systematic prediction model to forecast colour quickly and cost-effectively in real-time data. This research examines the colour forecasting process, its methodology, and how it is presented and used in the fashion industry. This study used Machine Learning (ML) to examine image data from the latest fashion trends to collect data via web-scraping images, extracting colours from images using k-means algorithms, and assessing the most trending 45 colours. We feed trendy clothing colours from different Pakistani brand websites to the algorithm. The algorithm predicts the frequency of each colour to either be positive or negative in the future. This research uses a forecasting model like ARIMA to forecast colour trends. Furthermore, the mean squared error is quite low, at 0.025. As a result, the current intelligent prediction system meets the criteria for capturing colour in trends for organizations. It would be a good development tool to maximize the profits of fashion companies and minimize the loss. Furthermore, it enables industries to make decisions for selecting colour trends.