Predicting fashion trend using runway images: application of logistic regression in trend forecasting

Trend forecasting is a challenging job and needs precise prediction based on colour, pattern, and style. Nowadays, researchers are applying machine learning and predictive models to predict the trend. Fashion runways are considered important events by high-street and fast fashion retailers. These events inspire them to design and develop different styles for the mass people. This research presented an approach to predict pattern and outfit based on the images collected from New York Fashion Week Fall/Winter 2019 (NYFW-19) Instagram posts, using logistic regression. The results predicted the patterns that could be used by retailers in the coming season for mass-market consumers. However, it could not predict outfit as a function of colour as there was no relationship between these two variables.