Models frequently used in management science are deterministic model and stochastic model. The goal of this research was to develop a stochastic forecasting model using time series forecasting method. This research paper is largely composed of two sec...
Models frequently used in management science are deterministic model and stochastic model. The goal of this research was to develop a stochastic forecasting model using time series forecasting method. This research paper is largely composed of two sections. Section 3 proposes a method for estimating fashion market size using sample data, and Sections 4 and 5 proposes a model for forecasting fashion market demand using time series for fashion market size estimated from Section 3.
In today's global competition, plan management technology is the basis of corporate competitiveness. As such, the purpose of this research was to provide a method for decision making based on accurate demand-forecasting data.
This research used the following steps in order to develop a forecasting model that provides accurate demand forecasting data. First, 1,400 samples were collected by survey and interview from 7 metropolitan cities according to proportionate allocation, and fashion market sizes were estimated for each segmented market. Second, 9 exponential smoothing methods were performed on fashion market sizes, which were collected as time series, to develop univariate time series models. Third, 43 variables affecting fashion market size were selected based on previous research, and multivariate time series models were developed by time series regression. Fourth, forecasting accuracies of univariate and multivariate time series models developed in this research were compared.
Fashion market size studied in this research was limited to the total amount of expenditure for fashion clothing products directly purchased by Koreans for wear during 6 months in spring and summer and 6 months in autumn and winter. Total of 8 markets were analyzed, including 7 segmented markets (men's formal wear, women's formal wear, casual wear, sportswear, underwear, and children's wear) and the total market, which is the sum of all segmented markets. Data used for analysis were time series data were collected twice a year for total of 26 time points by the author from 1998 to 2010. Targets of the demand forecasting model were 7 markets excluding outerwear market, which is sensitive to seasonal index, and 21 forecasting models corresponding to three temporal divisions (first half, second half, and whole year) for each segmented market were proposed.
The core goal of this research was to develop a fashion market demand forecasting model using time series data collected for the past 14 years by the author. Forecasting fashion market with high variability using only traditional statistical methods is difficult. However, current research contributes to the research on demand forecasting in fashion and other industries from the following aspects. First, the forecasting model for fashion market size proposed in this research can be used to forecast fashion market size as well as market size of consumer goods other than fashion products. Second, by developing 21 forecasting models for relatively short-term 26 time series data using exponential smoothing, this research proposed a methodology for companies to utilize their various short-term time series databases as useful forecasting data. Third, multivariate time series forecasting model was introduced to complement limitations of time series analysis, thereby improving the forecasting power. This attempt is necessary in order to respond to rapidly changing global economy nowadays. Fourth, temporal correlation analysis of fashion market size and various economic variables showed that there was variance in the duration of the economic variable influence on fashion market size. In other words, it was determined that economic condition immediately influences the fashion market, but the duration of the influence may be short-term or long-term. This result provides insight for companies to devise a long-term strategy to respond to rapidly changing economic environment.
This paper is composed of 6 sections. Section 1 outlines the research, Section 2 reviews the theoretical background and previous research, and Section 3 proposes a method for forecasting fashion market size. Overview of the research target, definition of variables, and the method used for site survey of variables are also presented. Section 4 shows the development of forecasting model using univariate exponential smoothing model and multivariate time series regression model. Section 5 analyzes the result of models from Section 4, and Section 6 presents the conclusions of this research.