Chi-Leung Hui

Fast fashion sales forecasting with limited data and time

Fast fashion is a commonly adopted strategy in fashion retailing. Under fast fashion, operational decisions have to be made with a tight schedule and the corresponding forecasting method has to be completed with very limited data within a limited time duration. Motivated by fast fashion business practices, in this paper, an intelligent forecasting algorithm, which combines tools such as the extreme learning machine and the grey model, is developed. Our real data analysis demonstrates that this newly derived algorithm can generate reasonably good forecasting under the given time and data constraints. Further analysis with an artificial dataset shows that the proposed algorithm performs especially well when either (i) the demand trend slope is large, or (ii) the seasonal cycles variance is large. These two features fit the fast fashion demand pattern very well because the trend factor is significant and the seasonal cycle is usually highly variable in fast fashion. The results from this paper lay the foundation which can help to achieve real time sales forecasting for fast fashion operations in the future. Some managerial implications are also discussed.

An intelligent fast sales forecasting model for fashion products

Research highlights? We devise a fast and efficient sales forecasting model for fashion products. ? We propose a systematic framework to determine the suitable parameters of the model. ? We illustrate the features and the efficiency of the model with two sets of real data. Sales forecasting is crucial in fashion business because of all the uncertainty associated with demand and supply. Many models for forecasting fashion products are proposed in the literature over the past few decades. With the emergence of artificial intelligence models, artificial neural networks (ANN) are widely used in forecasting. ANN models have been revealed to be more efficient and effective than many traditional statistical forecasting models. Despite the reported advantages, it is relatively more time-consuming for ANN to perform forecasting. In the fashion industry, sales forecasting is challenging because there are so many product varieties (i.e., SKUs) and prompt forecasting result is needed. As a result, the existing ANN models would become inadequate. In this paper, a new model which employs both the extreme learning machine (ELM) and the traditional statistical methods is proposed. Experiments with real data sets are conducted. A comparison with other traditional methods has shown that this ELM fast forecasting (ELM-FF) model is quick and effective.

Sales Forecasting for Fashion Retailing Service Industry: A Review

Sales forecasting is crucial for many retail operations. It is especially critical for the fashion retailing service industry in which product demand is very volatile and product’s life cycle is short. This paper conducts a comprehensive literature review and selects a set of papers in the literature on fashion retail sales forecasting. The advantages and the drawbacks of different kinds of analytical methods for fashion retail sales forecasting are examined. The evolution of the respective forecasting methods over the past 15 years is revealed. Issues related to real-world applications of the fashion retail sales forecasting models and important future research directions are discussed.

Color Trend Forecasting of Fashionable Products with Very Few Historical Data

In time-series forecasting, statistical methods and various newly emerged models, such as artificial neural network (ANN) and grey model (GM), are often used. No matter which forecasting method one would apply, it is always a huge challenge to make a sound forecasting decision under the condition of having very few historical data. Unfortunately, in fashion color trend forecasting, the availability of data is always very limited owing to the short selling season and life of products. This motivates us to examine different forecasting models for their performances in predicting color trend of fashionable product under the condition of having very few data. By employing real sales data from a fashion company, we examine various forecasting models, namely ANN, GM, Markov regime switching, and GM+ANN hybrid models, in the domain of color trend forecasting with a limited amount of historical data. Comparisons are made among these models. Insights on the appropriate choice of forecasting models are generated.

An Intelligent Quick Prediction Algorithm With Applications in Industrial Control and Loading Problems

The Artificial Neural Network (ANN) and its variations have been well-studied for their applications in the prediction of industrial control and loading problems. Despite showing satisfactory performance in terms of accuracy, the ANN models are notorious for being slow compared to, e.g., the traditional statistical models. This substantially hinders ANN models real-world applications in control and loading prediction problems. Recently a novel learning approach of ANN called Extreme Learning Machine (ELM) has emerged and it is proven to be very fast compared with the traditional ANN. In this paper, an Intelligent Quick Prediction Algorithm (IQPA), which employs an extended ELM (ELME) in producing fast, stable, and accurate prediction results for control and loading problems, is devised. This algorithm is versatile in which it can be used for short, medium to long-term predictions with both time series and non-time series data. Publicly available power plant operations and aircraft control data are employed for conducting analysis with this proposed novel model. Experimental results show that IQPA is effective and efficient, and can finish the prediction task with accurate results within a prespecified time limit.

Intelligent Fashion Forecasting Systems: Models and Applications

Forecasting is a crucial function for companies in the fashion industry, but for many real-life forecasting applications in the, the data patterns are notorious for being highly volatile and it is very difficult, if not impossible, to analytically learn about the underlying patterns. As a result, many traditional methods (such as pure statistical models) will fail to make a sound prediction. Over the past decade, advances in artificial intelligence and computing technologies have provided an alternative way of generating precise and accurate forecasting results for fashion businesses. Despite being an important and timely topic, there is currently an absence of a comprehensive reference source that provides up-to-date theoretical and applied research findings on the subject of intelligent fashion forecasting systems. This three-part handbook fulfills this need and covers materials ranging from introductory studies and technical reviews, theoretical modeling research, to intelligent fashion forecasting applications and analysis. This book is suitable for academic researchers, graduate students, senior undergraduate students and practitioners who are interested in the latest research on fashion forecasting.

Quick Response Healthcare Apparel Supply Chains: Value of RFID and Coordination

Radio frequency identification (RFID) is an important tool for enhancing the performance of inventory management. Motivated by our real-world observations in local hospitals, we study the case where a hospital, which is using a bar-coding system, considers switching to an RFID system because the RFID technology can potentially enhance the hospitals inventory management under the quick response system (QRS). In order to examine the value of RFID, we develop a formal analytical Bayesian model for the information updating process. We derive the expected value of information of the RFID system, and reveal the conditions in which the RFID system outperforms the bar-coding system. In addition, we evaluate the impacts of the QRS toward the expected profit and level of risk of the hospital, the supplier, and the whole supply chain (SC) system. We further propose two policies to help achieve SC coordination. Numerical analyses are reported and important insights are generated.

A New and Efficient Intelligent Collaboration Scheme for Fashion Design

Technology-mediated collaboration process has been extensively studied for over a decade. Most applications with collaboration concepts reported in the literature focus on enhancing efficiency and effectiveness of the decision-making processes in objective and well-structured workflows. However, relatively few previous studies have investigated the applications of collaboration schemes to problems with subjective and unstructured nature. In this paper, we explore a new intelligent collaboration scheme for fashion design which, by nature, relies heavily on human judgment and creativity. Techniques such as multicriteria decision making, fuzzy logic, and artificial neural network (ANN) models are employed. Industrial data sets are used for the analysis. Our experimental results suggest that the proposed scheme exhibits significant improvement over the traditional method in terms of the time-cost effectiveness, and a company interview with design professionals has confirmed its effectiveness and significance.

Towards the objective evaluation of garment appearance

As a first step towards objective evaluation of garment appearance, the present work considered seams on three‐dimensional surfaces which simulate actual garment surfaces. The geometric profiles of the 3‐D seams were scanned using a laser scanner. 1‐D and 2‐D digital filters were applied to obtain pucker signals from the geometric profiles by removing “high frequency” components due to fabric surface texture and “low frequency” components due to garment silhouette and drape. The advantages and disadvantages of the 1‐D and 2‐D digital filters are compared. Four physical parameters are examined to see their relevance to the subjective pucker grade. It was found that log(σ2), i.e. the logarithm of the variance of the heights of pucker signals, is the best set of physical parameters for the objective evaluation of seam pucker. In addition, latest attempts at capturing and analyzing 3D garment image using a Cyberware laser scanner and Surfacer software are reported.

Learning‐based fuzzy colour prediction system for more effective apparel design

Purpose – This paper aims to design and develop a learning‐based fuzzy colour prediction system for providing more effective apparel design in computer‐aided design system.Design/methodology/approach – In this study, we propose using a fuzzy system integrated with preliminary knowledge of colour prediction for facilitating apparel design. The performance of the proposed system is evaluated in terms of its computational efficiency and robustness. In addition, the proposed system is evaluated by target group of customers.Findings – It was found that the performance of the proposed system is better than the traditional approach.Research limitations/implications – Although the proposed system has some limitations, the outcome of this study could be used to produce a future breakthrough in providing an intelligent computer‐aided design system for apparel product.Originality/value – Using such an approach, an apparel designer could predict the favourite colours of garment for a target group of customers. The sy...