Abstract

AbstractA procedure of demand forecasting using data mining techniques is proposed to forecast the sales amount of new short life-cycle products for an actual food processing enterprise. The enterprise annually produces 100∼150 kinds of new items with short life-cycle between one week and three months to supply 260 convenience stores in the region of jurisdiction. Based on the previous delivery data in the first selling week, sales amount in the second, and the third selling weeks can be forecasted for their new products. Especially, some effective association rules about hot items and cold items are obtained by using data mining technologies for new short life-cycle products.

Highlights

  • In recent years, the retailing environment is undergoing a significant change

  • Rie Gaku mining technique is proposed in this study to analyze the demand trends for those very short life-cycle new products, based on the actual delivery data at the initial stage of the sales

  • This study focuses on new short life-cycle delicatessen products sold in the convenience chain stores

Read more

Summary

Introduction

The retailing environment is undergoing a significant change. Since the 1980s, consumption patterns take on a diversified tendency, the life-cycle of products put on the market is gradually shortening. The research shows that to maintain efficient production, advanced information techniques for the analysis of marketing demand should be developed to transform available sales data into knowledge and into effective production action For this food manufacturing enterprise, it is critical to develop an advanced procedure to analyze the new product demand trend with a relatively short life-cycle. Rie Gaku mining technique is proposed in this study to analyze the demand trends for those very short life-cycle new products, based on the actual delivery data at the initial stage of the sales. A hierarchical neural network model receives as input the actual sales data of the new short life-cycle products during the first selling week so as to forecast the sales amount for the second and the third selling weeks. Characterization of shortened life-cycle products in this study and proposed demand forecasting procedure

Characterization of products in this study
Proposed demand forecasting procedure for shortened life-cycle products
Definition of objective variables and explanatory variables
Formatting the datasets for learning and verification mechanism
Neural network model generation for the learning mechanism
Verification Process
Extracting the sales association rules
Findings
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call