國內鋼筋價格波動趨勢預測模型之初步研究

A Preliminary Study on Developing a Forecast Model for Fluctuating Trends inSteel Reinforcing Bars Prices in Taiwan

毛朝瑞(Chao-Ruei Mao)

指導教授 ： 周慧瑜

朝陽科技大學/理工學院/營建工程系/碩士(2013年)

摘要

鋼筋對於營建與土木業來說，是不可或缺的重要大宗建材之一，由於鋼筋佔營建工程專案成本的比重平均高達約一成，鋼筋價格又受到國內外各種市場因素直接或間接的複雜影響，因此對於營造廠商而言，鋼筋價格的波動在專案的損益管控上，一向是相當棘手的風險因素。也因此，如何針對此一風險，發展對實務決策具參考價值的管理工具，是本研究嘗試探討的課題。
為建立適用於國內市場條件之鋼筋價格波動風險監控模型，本研究首先針對3週內之短期價格波動，初步嘗試發展時間序列預測模型，以及運用模型提供近期鋼筋價格波動趨勢參考訊息之可行性。模型經驗證後顯示，本研究所設計之風險監控模型，可找出影響鋼筋價格波動之主要因素，並以此進行短期價格之趨勢研判。不過，由於本研究仍屬於模型發展之初期階段，重點在探討如何應用時間序列分析方法於此一課題上，以及以SPSS 18.0統計分析軟體為介面之細部設定，並初步對於如何應用預測結果進行鋼筋價格波動風險監控建立了驗證方法，也作了實際操作案例的示範。至於模型在預測準確性上的推進，則必須倚賴後續取得更多更完整的潛在變因數據資料，再做進一步的模型測試與實務驗證。

關鍵字

時間序列分析 ； 風險監控 ； 鋼筋價格

並列摘要

Reinforcing steel bars (rebar) are indispensable building materials in the construction industry, accounting for approximately 10% of construction project costs. The price of rebar is directly and indirectly susceptible to the influence of various domestic and international market factors; thus, rebar price fluctuations are a problematic risk factor for construction companies when managing project profits and losses. This study involved developing a management tool that minimizes risks related to rebar price fluctuations and demonstrates practical value as a decision-making reference.
To construct a monitoring model that suits domestic market conditions in Taiwan, a time-series forecasting model was developed for monitoring and controlling the risk of short-term rebar price fluctuations. The results showed that the proposed risk-monitoring model identified the main factors that influenced rebar price fluctuations; thus, the model was applied to assess a 3-week price trend. Because this research was at the preliminary stage of developing the risk-monitoring model, the researchers focused on applying time series analysis to the research topic. Furthermore, the SPSS (version 18.0) statistical software was employed for specific interface settings, a rudimentary verification method was established for applying the model forecasting results to monitor and control the risk of rebar price fluctuations, and a case study was conducted. Regarding the forecasting accuracy of the proposed model, additional comprehensive data regarding the latent factors that affect rebar price fluctuations should be collected to further test and verify the model.

並列關鍵字

risk-monitoring ； time series analysis ； reinforcing bars prices

參考文獻

16. 王百祿，「ARIMA與適應性SVM之混合模型於股市指數預測之研究」，碩士論文，國立成功大學，資訊管理研究所，台南，民國96年。

18. 曾淑惠、王志成，「時間數列ARIMA模式與多變量模糊時間數列模式在預測應用之比較 ─以台灣出口金額之預測為例」，私立致理技術學院國貿系、私立中原大學會計系，國內第四屆全國實證經濟學論文研討會，臺北、桃園，2003年。

1. Marija Norušis, IBM_SPSS_Forecasting 20, Prentice Hall, New Jersey,2011.

2. Marija Norušis, PASW Statistics 18 Core System User''s Guide, Prentice Hall, New Jersey,2007.

3. 金屬中心產業研究組電子報網站：http://www.mirdc.org.tw/publish/Epaper01.aspx?sty=06，查詢日期：2013年4月。