企業面臨瞬息萬變的競爭環境，紛紛導入企業資源規劃 (Enterprise Resource Planning；ERP)，ERP能將市場訊息與顧客需求等即時資訊快速回覆給企業管理階層，藉以增進企業營運效率。對於ERP的開發廠商而言，必須持續投入新的研發資源以達成企業管理的需求，並將開發風險降至最低，以獲取最大的利潤。
　　而這些軟體開發專案最終是否執行，在軟體的開發過程中雖然必須同時考慮時程、成本與品質三項因素，但一般而言，軟體專案管理者會基於市場先機的取得，排除因顧客需求或其他整合型計劃延遲等因素，而將開發時程列為最優先的考量。因此，開發時數是重要的關鍵決定要素之一，本研究直接以開發時數代表軟體專案之開發成本，除能真實反應公司內部投入專案的成本，也不會因為專案對外的報價折扣不一讓內部投入成本失真。
　　然而在專案初期因為資訊不充足，以及軟體產品具有客製化與客戶服務的特性，更增加了初始估算的困難性，因此，如何準確估算開發時數，是非常重要的任務，若專案成本估算過高會減低專案執行的可能性，相反的若估算低於實際開發成本，將侵蝕公司的營運績效。本研究以國內某大型ERP軟體公司近年開發之資訊系統為資料庫，應用線性迴歸、支援向量迴歸與類神經網路等資料探勘技術，以分類及推估方法建構並比較多種階層式時程推估模式，期在合理的準確度下，有效率地估算專案開發所需時數。本研究所提出之階層式推估模式，其估算能力比單一推估模式準確率更佳，可作為評估專案初期成本與價格的另一項參考依據，若再搭配其他專案評估方法一起運用，將有助於決策者於軟體專案評估時做更詳細之判斷。

　　Enterprises import ERP (Enterprise Resource Planning) one after another due to facing an intense competitive environment. ERP can rapidly react to instant information such as the market information and customer demand for the management hierarchy of enterprises to improve the efficiency of business operations. In order to maximize profits, ERP system developers must persistently invest new R&D resources to achieve business management needs and minimize development risks.
　　However, software development project implementation ultimately implemented depends on several factors, including development time, cost and quality. But in general, software project managers’ consideration is based on market opportunities, thus set the development time as the highest priority consideration, and exclude customer demand or other factors such as integrated plans to delay. Therefore, the development time is one of the important key determinants, this study directly substituted the development time for software project development costs. It not only can truly reflect the project development costs but also have no distortions of the internal costs while offering different discounts for external clients.
　　The inadequate information in the early stages of the project, and software has features with customized and a customer service orientation, increasing the difficulty of the initial estimation. Therefore, how to estimate project development time during the early stage of development thus is an important task for software developers. If the estimated development cost of a project exceeds the expected cost, the probability of the project implementation is reduced; the opposite if the estimated cost is far significantly below the actual cost, the project will negatively impact firm’ business performance. This study employed information systems recently developed by a large domestic DRP developer as a database. Using Linear Regression and data mining techniques such as Support Vector Regression and Neural Network, this study applied classification and estimation methods to construct a hierarchical time-estimation model for software development projects and compare it with other models, in the hope of estimating project development time efficiently and accurately. Comparing with a single estimation model, the study proposed hierarchical time-estimation model has better estimation ability. The estimation result provides an important reference for project decision makers, it also increases their decision making quality.

Jay Heizer and Barry Render (2001), “Operations Management,” Prentice Hall, sixth edition, 292-294.
Kale, V., (2000) “Implementing SAP R/3: The Guide for Business and Technology Managers.” SAMS Publishing, Indianapolis.
Ian Sommerville (2001), “Software Engineering,” 6th Edition, Addison-Wesley.
Royce, W. (1970). “Managing the Development of Large Software Systems.” Proceedings of IEEE WESCON 26 (August), 1-9.
Bally, L., Brittan, J., and Wagner, K. H. (1977). “A Prototype Approach to Information System Design and Development.” Information and Management, 1, 21-26.
Mills, H. (1971). “Top-down programming in large systems.” Debugging Techniques in Large Systems, 41-55.
Boehm (1988). “A Spiral Model of Software Development and Enhancement.” TRW Defense Syst. Group, Redondo Beach, CA, 61-72.
James Martin (1991), “Rapid Application Development.” Macmillan Publishing Company, New York.
Frawley, W. J., G. Piatesky-Shapire, and GCJ Matheus (1991). “Knowledge Discovery in Databases : An Overview, Menlo Park.” CA:AAAI Press/The MIT Press.
Fayyad, U. M., Piatesky, S. G. and Smyth, P. (1996). “From Data Mining to Knowledge Discovery in Databases.” American Association for Artificial Intelligence, 37-54.
Berry, M. J. A. and Linoff, G. (1997). “Data Mining Techniques: for Marking, Sales, and Customer Support.” NY: John Wiley & Sons Inc.
Jiawei Han and Micheline Kamber (2001), “Data Mining Concepts and Techniques.” Fraser University, Moran Kaufmann publishers, 45-61
Vapnik, V. (1995), “The nature of statistic learning theory.” New York, Springer-Verlag.
Vapnik, V., S. Golowich and A. Smola (1997), Support Vector Machines for function approximation, regression estimation, and single processing, Advances in Neural Information Processing Systems, 9, 281-287.
R. Agarval, et.al. (2001), “Estimating software projects.” Software Engineering Notes, 60-67.
Aamodt, A. & Plaza, E. (1994), “Case-based reasoning: foundational issues, methodological variations, and system approaches.” Artificial Intelligence Communications, Vol. 7, 39-59.
Turney, P. B. (1991). Common cents: the ABC performance breakthrough, Cost Technology, Hillsboro, Oregon.
Albrecht, A. J. (1979). “Measuring Application Development Productivity.” Proceedings of the Joint SHARE/GUIDE/IBM Application Development Symposium, 83-92.
IFPUG (2009). “Function Point Counting Practices Manual Release 4.3.” International Function Point Users Group.
David Longstreet (2002), “Fundamentals of Function Point Analysis.” www.SoftwareMetrics.Com。
Thomas Fetcke, Alain Abran and Tho-Hau Nguyen (1998), “Mapping the OO-Jacobson Approach into Function Point Analysis.” IEEE.
Garmus and Herron (2001), “Function Point Analysis: Measurement Practices for Successful Software Projects.” Addison-Wesley.
Knowledge PMBOK Guide (2004), “Project Management Institute.” Newtown Square, P. A, USA.
Werbos, P. J. (1974), Beyond regression: New tools for prediction and analysis in the behavioral sciences. Ph.D. Thesis, Harvard University.
Rumelhart, D. E., Hinton, G. E. & Williams, R. J. (1986), in Parallel Distributed Processing: Explorations in the Microstructure of Cognition. Vol. 1: Foundations (eds. Rumelhart, D. E. & McClelland, J. L.) 318−362 (MIT, Cambridge)
炬見工作室 （2006），ERP企業資源規劃－導論與個案，博碩文化股份有限公司。
葉怡成 (1999)，類神經網路模式與應用實作，儒林書局。
羅華強 (2001)，類神經網路：MATLAB的應用，清蔚科技。
吳丞傑 (2010)，ERP系統開發時程影響因子探討與預測模型建構：以D公司為例，國立中正大學企業管理研究所碩士論文。