資訊管理學報
韓文銘;
頁: 185-205
日期: 2014/04
摘要: 實務者需要可靠的規模度量技術以便追蹤範疇遞增與工作量預估，然而，使用不適當的功能性規模度量方法論來度量軟體規模是沒有意義的。本研究提出一個決策模式來協助經驗不足的管理者選擇一個相對適當的軟體功能性規模度量方法。這個以CHAID決策樹為基礎的模式是經由使用包含14個系統特徵值的537筆ISBSG專案資料所建立，接著再使用累積增益圖與10等份交叉驗證法來檢視與呈現模式決策績效。共有4個系統特徵值（分散式資料處理、容易修改、複雜邏輯與多個站點）被保留於最後的決策模式中以便支援一個無前例可循系統是否適合用功能點分析來度量。因此這個結構化且可追蹤的決策模式可以用來支援軟體功能性規模方法的選擇問題，此外，本研究亦發現四個可確保管理資訊系統專案成功的關鍵因素。
關鍵字: 功能性規模度量;功能性規模度量方法論;功能點分析;軟體度量;
Abstract: Practitioners require realistic size measurement techniques to track scope creep and to estimate effort. However, using an unsuitable functional size measurement method (FSMM) to measure software size is ineffective. This study proposes a decision support approach to assist managers who lack adequate software engineering knowledge and expertise to perform functional size measurements when selecting a suitable FSMM among candidate FSMMs. A chi-squared automatic interaction detection (CHAID) model was developed based on 14 information system characteristics obtained from 537 International Software Benchmarking Standards Group (ISBSG) projects. A gains chart and 10-fold cross-validation showed the performance of the proposed approach. Four IS characteristics were retained in the final approach, including distributed data processing, facilitate change, complex processing, and multiple sites, thereby enabling meaningful decisions to be made on whether function point analysis can be used to quantify the functional requirements of unprecedented software. This study proposes a structured and traceable method rather than a rule-of-thumb method to support decision-making for FSMM selection problems. In addition, our findings revealed crucial features for management information system (MIS) projects that should be carefully controlled to ensure success.
Keywords: Functional size measurement;FSM;Function point analysis;Software metrics;ISO/IEC 14143;
瀏覽次數: 15020 下載次數: 797
引用 導入Endnote
頁: 185-205
日期: 2014/04
摘要: 實務者需要可靠的規模度量技術以便追蹤範疇遞增與工作量預估，然而，使用不適當的功能性規模度量方法論來度量軟體規模是沒有意義的。本研究提出一個決策模式來協助經驗不足的管理者選擇一個相對適當的軟體功能性規模度量方法。這個以CHAID決策樹為基礎的模式是經由使用包含14個系統特徵值的537筆ISBSG專案資料所建立，接著再使用累積增益圖與10等份交叉驗證法來檢視與呈現模式決策績效。共有4個系統特徵值（分散式資料處理、容易修改、複雜邏輯與多個站點）被保留於最後的決策模式中以便支援一個無前例可循系統是否適合用功能點分析來度量。因此這個結構化且可追蹤的決策模式可以用來支援軟體功能性規模方法的選擇問題，此外，本研究亦發現四個可確保管理資訊系統專案成功的關鍵因素。
關鍵字: 功能性規模度量;功能性規模度量方法論;功能點分析;軟體度量;
支援軟體功能性規模度量方法論選擇之決策模式
Abstract: Practitioners require realistic size measurement techniques to track scope creep and to estimate effort. However, using an unsuitable functional size measurement method (FSMM) to measure software size is ineffective. This study proposes a decision support approach to assist managers who lack adequate software engineering knowledge and expertise to perform functional size measurements when selecting a suitable FSMM among candidate FSMMs. A chi-squared automatic interaction detection (CHAID) model was developed based on 14 information system characteristics obtained from 537 International Software Benchmarking Standards Group (ISBSG) projects. A gains chart and 10-fold cross-validation showed the performance of the proposed approach. Four IS characteristics were retained in the final approach, including distributed data processing, facilitate change, complex processing, and multiple sites, thereby enabling meaningful decisions to be made on whether function point analysis can be used to quantify the functional requirements of unprecedented software. This study proposes a structured and traceable method rather than a rule-of-thumb method to support decision-making for FSMM selection problems. In addition, our findings revealed crucial features for management information system (MIS) projects that should be carefully controlled to ensure success.
Keywords: Functional size measurement;FSM;Function point analysis;Software metrics;ISO/IEC 14143;
瀏覽次數: 15020 下載次數: 797
引用 導入Endnote