머신러닝을 이용한 여성 골프웨어 판매량 예측 -기상요인과 요일을 중심으로- (Prediction of Women's Golf Wear Sales Using Machine Learning -Focused on Weather Factors and Days of the Week-)

서지정보

· 발행기관 : 한국복식학회
· 수록지 정보 : 복식 / 71권 / 1호 / 17 ~ 33페이지
· 저자명 : 한기향

초록

Despite the overall low growth of fashion industry due to growing interest in sports and increasing demand for sportswear, golf wear showed the biggest growth in 2020. Although the launch of new brands is continuing due to the growth of golf wear, as the competition is overheating and the consumers' need is diversing, the issue of inventory has become one factor that hinders the growth of the fashion industry. The increase in big data and the use of AI due to Industry 4.0 are already widely used throughout the industry, and various attempts are already made in the fashion industry. The purpose of this study is to identify factors that can affect consumers' purchases and propose strategies to solve inventory problems and maximize profit margins through predictive planning and production near to the point of sale. The authors analyzed the data utilizing SPSS Modeler 18.0, in particular its Neural Networks, CHAID, C&RT, GenLin, and Ensemble algorithms. The data used daily sales data and seven types of weather factor and day information provided by the Korea Meteorological Administration over the past three years (January 1, 2017 to December 31, 2019) of the domestic golf wear brand 'A'. The analysis shows the highest predictive power in the Ensemble model algorithm, with an accuracy of 82.7% of the prediction. The factor that has the greatest impact on sales forecasting was shown as category, followed by season and month as important variables in sales forecasting. An analysis of data separately for detailed analysis determined that the category was an important factor for prediction in all four seasons. In spring and fall, weather requirements such as average relativity, minimal air temperature, and average temperature have affected sales, while in summer and winter, days of the week have been identified as important factors for forecasting sales.nt factors for forecasting sales.

영어초록

Despite the overall low growth of fashion industry due to growing interest in sports and increasing demand for sportswear, golf wear showed the biggest growth in 2020. Although the launch of new brands is continuing due to the growth of golf wear, as the competition is overheating and the consumers' need is diversing, the issue of inventory has become one factor that hinders the growth of the fashion industry. The increase in big data and the use of AI due to Industry 4.0 are already widely used throughout the industry, and various attempts are already made in the fashion industry. The purpose of this study is to identify factors that can affect consumers' purchases and propose strategies to solve inventory problems and maximize profit margins through predictive planning and production near to the point of sale. The authors analyzed the data utilizing SPSS Modeler 18.0, in particular its Neural Networks, CHAID, C&RT, GenLin, and Ensemble algorithms. The data used daily sales data and seven types of weather factor and day information provided by the Korea Meteorological Administration over the past three years (January 1, 2017 to December 31, 2019) of the domestic golf wear brand 'A'. The analysis shows the highest predictive power in the Ensemble model algorithm, with an accuracy of 82.7% of the prediction. The factor that has the greatest impact on sales forecasting was shown as category, followed by season and month as important variables in sales forecasting. An analysis of data separately for detailed analysis determined that the category was an important factor for prediction in all four seasons. In spring and fall, weather requirements such as average relativity, minimal air temperature, and average temperature have affected sales, while in summer and winter, days of the week have been identified as important factors for forecasting sales.nt factors for forecasting sales.

참고자료

· 없음