利用模糊資料探勘改善專案前期成本估算效能

Applying Fuzzy Data Mining to Improve Project Cost Estimation at Early Stage

鄭堯升(Yau-Sheng Jeng)

指導教授 ： 張慶寶

崑山科技大學/資訊科技學院/數位生活科技研究所/碩士(2012年)

https://doi.org/10.6828/KSU.2012.00051

摘要

本論文研製之專案成本估算(Cost Estimation)為專案管理重要的活動，也是所有軟體開發公司的重要課題，而準確專案成本的估算不但可以提供軟體發展廠商承接專案的參考，亦可作為專案風險評估的重要依據，然大部分專案成本估算皆著重在專案規劃後，開發廠商常忽略在承包專案前進行準確性評估成本的程序，也存在著未知的變數與人為因素，因而伴隨著潛藏的風險，因此在決定承接專案前便需要評估專案成本以為承接與否的依據，在此承包專案之前的評估階段，目前國內較少提出在此階段之評估。因此本研究提出一種以模糊資探勘(Fuzzy Data Mining)為基礎的專案成本評估方式利用專案，蒐集專案前期有限的資訊來評估專案成本，以便作為承接專案的參考，本研究利用所蒐集的專案資料來建立評估模型與評估模型驗證，歷史資料進行前置處理，將數據資料進行定義與規則。影響成本的因素不只是軟體大小(Size)與需求的複雜度(Complexity)，風險影響範圍的程度也應納入校正考量的因素為主要關鍵，蒐集歷史專案資料的數據越多，對於成本估計的準確性也越高，本研究的結果顯示所提出的專案成本評估方式可以用於專案前期的成本評估並獲致相當不錯的結果。

關鍵字

成本估算 ； 模糊資料探勘 ； 專案管理

並列摘要

A procedure is Project cost estimate is an important project management activities, but for all software development an important issue facing the company, and accurate project cost estimates not only provide software development company to undertake the project reference, but also as a project risk assessment important basis, but now most of the project cost estimates are focused on project planning, the so manufacturers often overlooked in development before the project contract, the cost of the procedure for assessing the accuracy, of course, there are unknown variables and human factors , but also because of potential risks associated with these, unfinished projects and failed to develop a small number of manufacturers have not, so the decision to undertake the project before the project costs will need to assess whether or not to undertake the basis that, in this assessment before the project contract stage, the current domestic focus at this stage make less, but after the project plan for the spindle. Therefore, this study proposes a fuzzy data mining project cost estimates based on the use of ad hoc, limited pre-project to collect information to assess the project costs in order to undertake the project as a reference, this study collected project data to establish assessment models and evaluation of model validation, historical data pre-processing, data definition and data rules. Factors affecting the cost of software is not just the size and the complexity of demand, the degree of risk affected areas should be included in consideration of the correction factors for the key, the data collected historical project data more accurate cost estimates for is also higher, this study showed that the cost of the proposed project pre-assessment methods can be used to project the costs of assessment and was to a very good result.

並列關鍵字

cost estimates ； fuzzy data mining ； project management

參考文獻

[4] Baydoun, M., "Risk management of large-scale development projects in developing countries: Cases from MDI's projects,"International Journal of Technology Management & Sustainable Development, Vol.9, Issue 3, pp. 237-249, 2011.

[5] Genuchten, M., "Why is software late?an empirical study of reasons for delay insoftware development,"IEEE Trans. Softw. Eng. Vol.17, 1991.

[6] Nurmuliani, N., Zowghi, D., Fowell, S., "Analysis of requirements volatility duringsoftware development life cycle,"2004 Australian Softw. Eng. Conf.(ASWEC’04),Australian, 2004.

[8] MolOkken, K., JOrgensen, M., "A comparison of software project

[9] Jorgensen,M.,"Estimation of Software Development Work Effort: Evidence on Expert Judgement and Formal Models," Int. Journal of Forecasting, Vol.23, No.3, pp.449-462,2007.