跳到主要內容

臺灣博碩士論文加值系統

(44.192.38.248) 您好!臺灣時間:2022/11/27 00:20
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

我願授權國圖
: 
twitterline
研究生:桑米高
研究生(外文):Miguel Sandoval
論文名稱:An Application of Support Vector Regression inForecasting Newspaper Business Demand
論文名稱(外文):An Application of Support Vector Regression inForecasting Newspaper Business Demand
指導教授::蔡 啟 揚 博士
指導教授(外文):Dr. Chi-Yang Tsai
口試委員:Chen, Yee-MingChi-Tai Wang
口試委員(外文):Chen, Yee-MingChi-Tai Wang
口試日期:2015-06-17
學位類別:碩士
校院名稱:元智大學
系所名稱:工業工程與管理學系
學門:工程學門
學類:工業工程學類
論文種類:學術論文
畢業學年度:103
語文別:英文
論文頁數:84
中文關鍵詞:84sales predictionnewspaper circulationsupport vector regression
外文關鍵詞:sales predictionnewspaper circulationsupport vector regression
相關次數:
  • 被引用被引用:0
  • 點閱點閱:80
  • 評分評分:
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
ABSTRACT
This thesis deals with forecasting newspaper for a local newspaper business. Current
methodology is based on classical statistical methods such as Exponential Smoothing.
However most of these classical statistical approaches are mainly focused on building
regression models based only on sample data sets. In the last years new methods based on
Learning Machines are being employed for forecasting problems. These methods understand
universal approximations of nonlinear functions, thus resulting more able to model complex
nonlinear phenomena. This study suggest the use of a predictive data mining technique called
“Support Vector Regression” for forecasting daily newspaper sales for the major newspaper
business in Nicaragua. In order to improve the forecasting accuracy, this thesis quantifies some
factors that affect the forecasts of sales of newspaper such as: promotion, performance per
outlet, weather condition, which are more suitable for this type of product. Our numerical
experiments show that forecasting sales per outlet through SVR and the input variables
considered outperform the current methodology employed by the company and other classical
statistical approaches based on common prediction accuracy measures such as the mean
absolute percentage error “MAPE”. In addition a practical model implemented in Ms Excel
3
based on genetic algorithm for variable selection and the SVR’s parameter is adapted. Our
results prove that it is advantageous to apply SVM to forecast daily newspaper demand.
Table of Contents
ABSTRACT ......................................................................................................... 2
TABLE OF CONTENTS .................................................................................... 4
LIST OF FIGURES ............................................................................................ 6
LIST OF TABLES ............................................................................................... 7
ACKNOWLEDGMENT .................................................................................... 8
CHAPTER 1 INTRODUCTION ....................................................................... 9
1.1 RESEARCH PROBLEM .................................................................................................. 10
1.2 RESEARCH OBJECTIVE ................................................................................................ 11
1.3 RESEARCH METHODOLOGY........................................................................................ 12
CHAPTER 2 BACKGROUND THEORY ..................................................... 14
2.1 FORECASTING METHODS ............................................................................................. 14
2.1.1 Na#westeur048#ve forecast method ..................................................................................... 15
2.1.2 Moving average ............................................................................................... 15
2.1.3 Simple exponential smoothing ......................................................................... 16
2.1.4 Advanced exponential smoothing .................................................................... 17
2.2 BASIC IDEA OF SUPPORT VECTOR MACHINE .............................................................. 19
2.2.1 Empirical Risk Minimization (ERM) .............................................................. 22
2.2.2 Structural Risk Minimization (SRM)............................................................... 23
2.3 SUPPORT VECTOR CLASSIFICATION ........................................................................... 25
2.4 KERNEL FUNCTIONS .................................................................................................... 29
2.5 SUPPORT VECTOR REGRESSION ................................................................................. 31
CHAPTER 3 PARAMETERS OF SVR AND IMPLEMENT ATION ......... 35
3.1 PARAMETERS OF THE SVR .......................................................................................... 35
3.1.1 Exhaustive Grid Search.................................................................................... 36
3.1.2 Cherkassky’s Approach .................................................................................... 36
3.1.3 A combination of both (Cherkassky’s + Grid search approach) ...................... 37
3.1.4 Heuristic search (Genetic Algorithms) ............................................................. 37
5
3.2 IMPLEMENTATION ....................................................................................................... 39
3.2.1 Steps for obtaining the models ......................................................................... 43
3.2.2 Experiments Results......................................................................................... 44
CHAPTER 4 NEWSPAPER FORECASTING CASE OF STUDY ............. 46
4.1 DATA SET (TRAINING, VALIDATION AND TEST) ............................................................ 46
4.2 THE SVR MODEL ......................................................................................................... 48
4.3 RESULTS AND ANALYSIS .............................................................................................. 52
CHAPTER 5 CONCLUSION .......................................................................... 62
5.1 CONTRIBUTIONS .......................................................................................................... 62
5.2 FUTURE STUDIES .......................................................................................................... 63
REFERENCE .................................................................................................... 64
APPENDIX ........................................................................................................ 67
REFERENCE
Akbaş, B., Karahoca, D., Karahoca, A., &; G#westeur061#ng#westeur055#r, A. (2014). Predicting newspaper sales by
using data mining techniques. Paper presented at the Proceedings of the 15th
International Conference on Computer Systems and Technologies.
Alon, I., Qi, M., &; Sadowski, R. J. (2001). Forecasting aggregate retail sales:: a comparison of
artificial neural networks and traditional methods. Journal of Retailing and Consumer
Services, 8(3), 147-156. doi: http://dx.doi.org/10.1016/S0969-6989(00)00011-4
Bishop, C. M. (2006). Pattern recognition and machine learning (Vol. 1): springer New York.
Boser, B. E., Guyon, I. M., &; Vapnik, V. N. (1992). A training algorithm for optimal margin
classifiers. Paper presented at the Proceedings of the fifth annual workshop on
Computational learning theory.
Box, G. E., Jenkins, G. M., &; Reinsel, G. C. (2013). Time series analysis: forecasting and
control: John Wiley &; Sons.
Brandimarte, P., &; Zotteri, G. (2007). Introduction to distribution logistics (Vol. 21): John
Wiley &; Sons.
Burges, C. J. (1998). A tutorial on support vector machines for pattern recognition. Data mining
and knowledge discovery, 2(2), 121-167.
Cardoso, G., &; Gomide, F. (2007). Newspaper demand prediction and replacement model
based on fuzzy clustering and rules. Information Sciences, 177(21), 4799-4809. doi:
http://dx.doi.org/10.1016/j.ins.2007.05.009
Carter, A. E., &; Ragsdale, C. T. (2002). Scheduling pre-printed newspaper advertising inserts
using genetic algorithms. Omega, 30(6), 415-421. doi:
http://dx.doi.org/10.1016/S0305-0483(02)00059-2
Chen, K.-Y. (2007). Forecasting systems reliability based on support vector regression with
genetic algorithms. Reliability Engineering &; System Safety, 92(4), 423-432. doi:
http://dx.doi.org/10.1016/j.ress.2005.12.014
Chen, K.-Y., &; Wang, C.-H. (2007). Support vector regression with genetic algorithms in
forecasting tourism demand. Tourism Management, 28(1), 215-226. doi:
http://dx.doi.org/10.1016/j.tourman.2005.12.018
Cherkassky, V., &; Ma, Y. (2004). Practical selection of SVM parameters and noise estimation
for SVM regression. Neural Networks, 17(1), 113-126. doi:
http://dx.doi.org/10.1016/S0893-6080(03)00169-2
Chervonenkis., V. V. a. A. (1974). Structural risk minimization. from
http://www.svms.org/srm/
Chih-Wei Hsu, C.-C. C., and Chih-Jen Lin. ( April 15, 2010). A Practical Guide to Support
Vector Classification. Techinical Report.
連結至畢業學校之論文網頁點我開啟連結
註: 此連結為研究生畢業學校所提供,不一定有電子全文可供下載,若連結有誤,請點選上方之〝勘誤回報〞功能,我們會盡快修正,謝謝!
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top