外文翻譯---基于數(shù)據(jù)挖掘的服裝銷售預(yù)測（英文）

1、NTC Project: S01-PH10 (formerly I01-P10) 1Forecasting Women’s Apparel Sales Using Mathematical Modeling Celia Frank*1, Balaji Vemulapalli1, Les M. Sztandera2, Amar Raheja3 1School of Textiles and Materials Technology 2C

2、omputer Information Systems Department, Philadelphia University, Philadelphia, PA 19144 3Department of Computer Science, California State Polytechnic, Pomona, CA 91768 Project web site: http://faculty.philau.edu/frankc/n

3、tc Goal The goal of the work is to demonstrate the effectiveness of soft computing methods like artificial neural networks and fuzzy logic models in apparel sales forecasting. 1. Abstract Sales Forecasting is an integ

4、ral part of apparel supply chain management and very important in order to sustain profitability. Apparel managers require a sophisticated forecasting tool, which can take both exogenous factors like size, price, color

5、, and climatic data, price changes, marketing strategies and endogenous factors like time into consideration. Although models built on conventional statistical forecasting tools are very popular they model sales only o

6、n historic data and tend to be linear in nature. Unconventional artificial intelligence tools like fuzzy logic and ANN can efficiently model sales taking into account both exogenous and endogenous factors and allow ar

7、bitrary non-linear approximation functions derived (learned) directly from the data. In this research, forecasting models were built based on both univariate and multivariate analysis. Models built on multivariate fuz

8、zy logic analysis were better in comparison to those built on other models. The performance of the models was tested by comparing one of the goodness-of-fit statistics, R2, and also by comparing actual sales with the f

9、orecasted sales of different types of garments. Five months sales data (August-December 2001) was used as back cast data in our models and a forecast was made for one month of the year 2002. The performance of the mod

10、els was tested by comparing one of the goodness-of-fit statistics, R2, and also by comparing actual sales with the forecasted sales. An R2 of 0.93 was obtained for multivariate analysis (0.75 for univariate analysis),

11、which is significantly higher than those of 0.90 and 0.75 found for Single Seasonal Exponential Smoothing and Winters’ Three Parameter model, respectively. Yet another model, based on artificial neural network approach

12、, gave an R2 averaging 0.82 for multivariate analysis and 0.92 for univariate analysis. 2. Present Research A multivariate fuzzy model has been built based on important product variables of color, time and size. Thi

13、s model is being extended to include other variables like climate, economic conditions etc., which would be used in building a comprehensive forecasting software package. National Textile Center Annual Report: November

14、2003 NTC Project: S01-PH10 (formerly I01-P10) 36. Fuzzy Logic Model Fuzzy logic allows the representation of human decision and evaluation in algorithmic form. It is a mathematical representation of human logic. The us

15、e of fuzzy sets defined by membership function constitutes fuzzy logic (Von Altrock, 1995). Fuzzy Set: is a set with graded membership over the interval [0, 1]. Membership function: is the degree to which the variabl

16、e is considered to belong to the fuzzy set. A sales fuzzy logic controller is made of: Fuzzification: Linguistic variables are defined for all input variables (color and size) . Fuzzy Inference: rules are compiled fr

17、om the database and based on the rules, the value of the output linguistic variable is determined. Fuzzy inference is made of two components: Aggregation: Evaluation of the IF part of the rules. Composition: Evaluatio

18、n of the THEN part of the rules. Defuzzification: linguistic value(s) of output variable (sales) obtained in the previous stage are converted into a real output value. This can be accomplished by computing typical val

19、ues and the crisp result is found out by balancing out the results. Linguistic value of sales Defuzzification Fuzzy Inference Color, size Linguistic Level Rules compiled from the database Fuzzification Real Level Ling

20、uistic value Sales (Real value) Figure 1: Fuzzy sales controller Fuzzy logic model was applied to grouped data and sales values were calculated for each size- class combination. Total sales value for the whole period was

21、 calculated by summing up the sales values of all the grouped items. Total Sales=∑0n sales (1) Where n? Number of size-color combinations In order to calculate daily sales

22、, two different methods were used: 6.1 Fractional contribution method It was observed that the fraction contribution of each weekday towards total week sales was constant (Garg, 2002). Table 4 and Figure 2 depict the

23、average fractional contribution of a weekday towards total sales of a week, which can be used to forecast the daily sales from the forecasted weekly sales. Table 4: Fractions of Weekly Sales Distributed Among 7 Days Da