Trend forecast and collection management in apparel retail

Loading...
Thumbnail Image

Date

2022

Journal Title

Journal ISSN

Volume Title

Publisher

Kadir Has Üniversitesi

Research Projects

Organizational Units

Journal Issue

Abstract

This study addresses the new methods and some existing methods with a different approach for trend forecasting and using new trends in the collections in apparel retail industry. There are several approaches to determine the potential of fashion trends. This study describes several approaches for trend forecasting and develops methods for measuring the potential of new fashion trends with unknown potential and without sales data. Firstly, merchandise testing focuses on the process of testing products with new trends. It describes the test store selection, forecasting methods and analyze the accuracy of forecasting with real data. Secondly, Sales-Based Store Network of Stores model is presented to examine cross-store sales similarity and establishes a store network using Collaborative Filtering method as in recommendation systems. A clustering method like K-means is studied to cluster the stores using store network data. Moreover, Distribution of Collection into Store method focuses on distributing the main collection made for a category into each stores using some constraints such as capacity of stores, rates of product attributes in the main collection. Integer programming is used to distribute the collection. The sales potential of the new planned products is crucial. It is necessary to choose the products with highest potential among the hundreds of products. Prediction of products’ demand based on stores addresses a prediction model using sales data containing store features and product attributes with different forecasting methods with different parameters. Furthermore, store-based forecasts are used in Distribution of collection into stores method while selecting the best products for the stores.

Description

Keywords

Apparel Retail, Fashion Trends, Merchandise Testing, Forecast, Clustering, K-means, Integer Programming, Collaborative Filtering

Turkish CoHE Thesis Center URL

Citation

WoS Q

Scopus Q

Source

Volume

Issue

Start Page

End Page

Collections