Abstract: Systems and methods are provided can be implemented using various ones of the algorithms, methodologies, predictions, preprocessing and/or models, among other things, in an initial size fit indicator process. This initial size fit indicator process, in a preferred embodiment, is directed to online sales, whereby for a given size of a given garment, a relative size indicator is provided, which relative size indicator is preferably chosen from two, three or five different values, but in any event is preferably a small set of values.

Abstract: Systems and methods are provided that analyze and extract implicit fit preference expressions from retailer transactional data, e.g. purchases and returns, and product data. Algorithms are provided that, among other things, extract such fit preference information. Such data may be processed offline to generate fit profiles for shoppers and then presented to such shoppers, e.g. in a friendly user interface such as on the retailer's product pages. Returning shoppers receive automatic fit predictions while first-time shoppers may receive fit predictions with a few clicks. Fit predictions may include sizes that fit a shopper best as well as other suitable garment, e.g. accessories. For example, when the system identifies the shopper as someone who has shopped at any of the online or mobile stores of the retailer before, embodiments automatically calculate fit predictions for the customer to preselect and present the size that fits her best.