臺灣博碩士論文加值系統

本研究主要是針對台灣 FMCG 產業的銷售資料進行分析，我們使用了便利商店的日
銷售資料來分析並且探討資料背後的隱藏的含意。本研究架構是根據 BADIR 架構來建造
模型，並且比較不同的研究變量從而找出之間的相關連性。
本篇論文使用 ARIMA 模型又稱時間序列模型來建模，根據每日銷售資料來預測未來
銷售量，使用時間序列建模的好處是我們可以清楚知道哪些變量會對銷售量造成影響，並
且對此做出適合的解釋。此研究結果顯示不同的產品會被不同的變量影響，像是某些產品
的銷售量對於氣溫的變量結果是顯著的。我們也發現傳統節日的變量對於某些零食餅乾銷
售量有顯著的影響，而這些零食通常是人們拜拜時會購買的產品。促銷通常能夠增加銷售
額，但並不是越高的促銷比例越能帶來高額銷售。

This research mainly focuses on sales data from the FMCG industry in Taiwan. We use the daily sales data from convenience store products to dig into the hidden story. This paper followed the BADIR framework to build the model and compared different variables to find out correlations.
Basically, this thesis focuses on the ARIMA model which could be called the time series model, based on daily sales data to predict future sales. The benefit of the time series model is we can clearly know which variables could affect the sales, and we can have more explanations and possible assumptions about our results. The result shows different items are affected by other variables, and the significant variables show that most of items are related to temperature. Those products are positively correlated with temperature. We also find out the traditional variables related to some snacks usually people use for worship. Promotions usually increase sales, but the higher ratio of promotions is not the key reason that led to higher sales.

摘 要 i
ABSTRACT ii
ACKNOWLEDGMENTS iii
TABLE OF CONTENTS iv
LIST OF TABLES vi
LIST OF FIGURES vii
CHAPTER 1 INTRODUCTION 1
1.1 General background Information 1
1.2 Research purpose 2
1.3 Research process introduction 2
1.4 Thesis conceptual framework 4
CHAPTER 2 LITERATURE REVIEW 5
2.1 The importance of sales forecasts to business 5
2.2 Fast Moving Consumer Goods (FMCG) global industry overview 6
2.3 Fast Moving Consumer Goods (FMCG) in Taiwan industry overview 7
CHAPTER 3 RESEARCH METHODOLOGY 9
3.1 Introduction 9
3.2 BADIR framework 11
3.3 Business questions and Analysis plan 13
3.4 Data Collection 13
3.4.1 Data review 13
3.4.2 Nomenclature 15
3.5 Data pre-processing 15
3.5.1 Normalized values 16
3.5.2 Correlation Analysis 17
3.5.3 Pearson Correlation Coefficient 17
3.6 Derive Insights 19
3.6.1 Univariate and multivariate testing 19
3.6.2 Variables selected 19
3.7 Time series introduction 19
3.7.1 AutoRegressive (AR) 20
3.7.2 Moving Average (MA) 20
3.7.3 ARMA 21
3.7.4 ARIMA 21
3.7.5 Time series model building 22
3.7.6 Analysis Tool 23
3.8 Evaluation 23
3.8.1 ACF and PACF Criterion 23
3.8.2 AIC and BIC Criterion 24
3.8.3 Evaluation index 25
3.8.4 LOESS 26
CHAPTER 4 RESULTS AND DISCUSSION 27
4.1 Introduction 27
4.2 IBM SPSS Modeler Analysis 28
4.2.1 Univariate and Multivariate Methods Comparison 28
4.2.2 Variables selected 30
4.2.3 Verification of variables 32
4.2.4 Model Verification 38
4.3 Analysis result 39
4.3.1 Model column name explanation 39
4.3.2 Analysis visualization 40
4.3.3 Traditional Holiday, Promotion, variables comparison 55
4.4 Result conclusion 57
4.4.1 Forecast results discussion 59
CHAPTER 5 CONCLUSIONS AND RECOMMENDATIONS 60
5.1 Research Conclusions 60
5.2 Result performance 60
5.3 Research recommendations 61
REFERENCE 62

REFERENCE
ArunKumar, K. E., Kalaga, D. V., Kumar, C. M. S., Chilkoor, G., Kawaji, M., & Brenza, T. M. (2021). Forecasting the dynamics of cumulative COVID-19 cases (confirmed, recovered and deaths) for top-16 countries using statistical machine learning models: Auto-Regressive Integrated Moving Average (ARIMA) and Seasonal Auto-Regressive Integrated Moving Average (SARIMA). Applied Soft Computing, 103. doi:10.1016/j.asoc.2021.107161
Aryng. (2011). Data Science + Decision Science. Retrieved from https://aryng.com/aryng-BADIR-advantage
Braun, M., Bernard, T., Piller, O., & Sedehizade, F. (2014). 24-Hours Demand Forecasting Based on SARIMA and Support Vector Machines. Procedia Engineering, 89, 926-933. doi:https://doi.org/10.1016/j.proeng.2014.11.526
Chen, P. C. (2018). Using Service Data Decision Framework to Construct a Sales Forecasting System of Perishable Commodities in CVS. (Master), National Kaohsiung University of Science and Technology, Kaohsiung.
Chou, J.-S., & Tran, D.-S. (2018). Forecasting energy consumption time series using machine learning techniques based on usage patterns of residential householders. Energy, 165, 709-726. doi:https://doi.org/10.1016/j.energy.2018.09.144
Deb, C., Zhang, F., Yang, J., Lee, S. E., & Shah, K. W. (2017). A review on time series forecasting techniques for building energy consumption. Renewable and Sustainable Energy Reviews, 74, 902-924. doi:https://doi.org/10.1016/j.rser.2017.02.085
Huan-Su, W. (2017). Regression calibration estimation of measurement errors in negative binomial regression models. (Master), National Chung Hsing University, Taichung City. Retrieved from https://hdl.handle.net/11296/emtrjs
IBM Corp. (1994). About IBM SPSS Modeler. Retrieved from https://www.ibm.com/docs/en/spss-modeler/18.2.1?topic=tutorial-about-spss-modeler
Jain, P., & Sharma, P. (2014). Behind every good decision: How anyone can use business analytics to turn data into profitable insight: Amacom.
Kwok-Wang, N. (2011). Cycle-Time-Aware Semiconductor Manufacturing Yield Analysis using Stepwise Regression with Pearson Correlation Approximation and Piecewise Linear Modeling
(Master), College of Engineering Yuan Ze University, Taoyuan.
Lewis, C. D. (1982). Industrial and Business Forecasting Methods: Butterworth-Heinemann
Ou, T. Y. (2010). Constructing a Retailing Sales Forecasting Model Base on Data Mining Framework: An Empirical Study on Fresh Food in Chain Store. (PhD), National Tsing Hua University, Hsinchu City.
Pan, I. H. (2018). Using Data Mining to Forecast Drug Consumption in Pharmaceutical Logistics Center. (Master), National Formosa University, Yunlin. Retrieved from https://hdl.handle.net/11296/9s6wub
Yang, Y.-R. (2018). A Study on the Correlation between in FMCG
Industry Business Performance and Brand Channels in Taiwan
－Case Study of Body Care Hand Care Products.
Yi-Nung, Y. (2011). Time series analysis in economics and finance. Taiper: YEH YEH BOOK GALLERY LTD.
Zeng, M.-F. (2019). Application of ARIMA Model inStock Price Forecasting. (Master), National Taiwan Ocean University, Keelung. Retrieved from https://hdl.handle.net/11296/vq83np
Box, G. E., Jenkins, G. M., Reinsel, G. C., & Ljung, G. M. (1976). Time series analysis: forecasting and control: John Wiley & Sons.
Hsu, C.-Y. (2013). Competitive Strategy of FMCG Industry in Taiwan：A Case Study of P&G. (Master), Department of Business Management National Sun Yat-sen University,