1. NCCU Academic Hub
  2. Publications
  3. College of Commerce
  4. Theses
  5. Item

Publications-Theses

Article View/Open

Publication Export

Google ScholarTM

NCCU Library

Citation Infomation

Related Publications in TAIR

題名 預測模型中遺失值之選填順序研究
Research of acquisition order of missing values in predictive model
作者 施雲天
貢獻者 唐揆
施雲天
關鍵詞 遺失值
決策樹分類
uncertainty score
decision tree
missing value acquisition
日期 2013
上傳時間 14-Jul-2014 11:26:37 (UTC+8)
摘要 預測模型已經被廣泛運用在日常生活中，例如銀行信用評比、消費者行為或是疾病的預測等等。然而不論在建構或使用預測模型的時候，我們都會在訓練資料或是測試資料中遇到遺失值的問題，因而降低預測的表現。面對遺失值有很多種處理方式，刪除、填補、模型建構以及機器學習都是可以使用的方法；除此之外，直接用某個成本去取得遺失值也是一個選擇。\n本研究著重的議題是用某成本去取得遺失值，並且利用決策樹(因為其在建構時可以容納遺失值)來當作預測模型，希望可以找到用較低的成本的填值方法達到較高的準確率。我們延續過去Error Sampling中Uncertainty Score的概念與邏輯。提出U-Sampling來判斷不同特徵值的「重要性排序」。相較於過去Error Sampling用「受試者」(row-based)的重要性來排序。U-Sampling是根據「特徵值」(column-based)的重要性來排序。\n我們用8組UCI machine Learning Repository的資料進行兩組實驗，分別讓訓練資料以及測試資料含有一定比例的遺失值。再利用U-Sampling、Random Sampling以及過去文獻所提及的Error Sampling作準確率和錯誤減少率的比較。實驗結果顯示在訓練資料有遺失值的情況，U-Sampling在70%以上的檔案表現較佳；而在測試資料有遺失值的情況，U-Sampling則是在87.5%的檔案表現較佳。\n另外，我們也研究了對於不同的遺失比例對於上述方法的效果是否有影響，可以用來判斷哪種情況比較適用哪一種選值方法。希望透過U-Sampling，可以先挑選重要的特徵值來填補，用較少的遺失值取得就得到較高的準確率，也因此可以節省處理遺失值的成本。
參考文獻 1. Allison, P. D. (2001). Missing data. Thousand Oaks, CA: Sage.\n\n2. Alpaydın, E. (2010). Introduction to machine learning. London, England: The MIT Press.\n\n3. Bennett, D. A. (2001). How can I deal with missing data in my study? Australian and New Zealand Journal of Public Health, 25(5), 464–469.\n\n4. Giks, Walter R ; Richardson, Sylvia; Spiegelhalter, David J. (1996). Introducing Markov chain Monte Carlo. In Markov chain Monte Carlo in practice (pp. 1-19). London: Chapman & hall/CRC.\n\n5. Graham, J. W. (2003). Adding missing-data-relevant variables to FIML basedstructural equation models. Structural Equation Modeling, pp. 10, 80–100.\n\n6. Jackson, J. (2002). Overview, data mining: a conceptual. Communications of the Association for Information Systems.\n\n7. Kohavi, R. (1995). A study of Cross-Validation and Bootstrap for Accuracy Estimation and Model Selection. IJCAI, (Vol.14, No.2, pp. 1137-1145).\n\n8. Levin, N., & Zahav, J. (2001, Spring). Predictive modeling using segmentation. Journal of Interactive Marketing, 15(2), 2-22.\n\n9. Melville, P., Saar-Tsechansky, M., Provost, F., & Mooney, R. (2004). Active Feature-Value Acquisition for Classifier Induction. Proceedings of the 4th IEEE International Conference on Data Mining, (pp. 483-486). Brighton, UK.\n\n10. Pallant, J. (2007). SPSS survival manual (3rd ed.). New York, NY: Open University Press.\n\n11. Pedro J. Garcı´a-Laencina Æ Jose´-Luis Sancho-Go´mez Æ, A. R.-V. (2010). Pattern classification with missing data: a review. Neural Comput & Applic.\n\n12. Peng, C. Y. J., Harwell, M., Liou, S.M., & Ehman, L.H. (2006). Advances in missing data methods and implications for educational research. In Real data analysis, 31-78. North Carolina,US : Information Age Publishing.\n\n13. Quinlan, J. R. (1989). Unknown attribute values in induction., In ML (pp. 164-168).\n\n14. Rubin, D. B. (1987). Multiple imputation for non-response in surveys. New York: John Wiley & Sons.\n\n15. Saar-Tsechansky, M., Melville, P., & Provost, F. (2009, 4). Active Feature-Value Acqusition. Management Science, 55(4), 664-684.\n\n16. Schafer, J. L. (1999). Multiple imputation: a primer. Statiscal methods in medical research, 8(1), 3-15.\n\n17. Schlomer, G. L., Bauman, S., & Card, N. A. (2010). Best Practices for Missing Data Management in Counseling Psychology. Journal of Counseling Psychology, 57(1), 1-10.\n\n18. Simon, H. A., & Lea, G. (1974). Problem solving and rule induction: A unified view. Knowledge and cognition. Oxford, England: Lawrence Erlbaum.\n\n19. Simon, H., & Lea, G. (1974). Problem solving and rule induction: A unified view. \n\n20. Turney, P. (2000). Types of Cost in Inductive Concept Learning. Proceedings of the Cost-Sensitive Learning Workshop at the 17th ICML-2000 Conference. Stanford, CA.\n\n\n \n\n21. Vinod, N. C., & Punithavalli, D. M. (2011). Classification of Incomplete Data Handling Techniques-An Overview. International Journal on Computer Science and Engineering, 3(1), 340-344.\n\n22. Zheng, Z., & Padmanabhan, B. (2002). On Active Learning for Data Acquisition. Proceedings of IEEE International Condference on Data Mining, (pp. 562-569).\n\n網路資料\n\n1. UCI machine Learning Repository. (n.d.). Retrieved from http://archive.ics.uci.edu/ml/
描述 碩士
國立政治大學
企業管理研究所
101355055
102
資料來源 http://thesis.lib.nccu.edu.tw/record/#G0101355055
資料類型 thesis
dc.contributor.advisor 唐揆zh_TW
dc.contributor.author (Authors) 施雲天zh_TW
dc.creator (作者) 施雲天zh_TW
dc.date (日期) 2013en_US
dc.date.accessioned 14-Jul-2014 11:26:37 (UTC+8)-
dc.date.available 14-Jul-2014 11:26:37 (UTC+8)-
dc.date.issued (上傳時間) 14-Jul-2014 11:26:37 (UTC+8)-
dc.identifier (Other Identifiers) G0101355055en_US
dc.identifier.uri (URI) https://ah.lib.nccu.edu.tw/item?item_id=75802-
dc.description (描述) 碩士zh_TW
dc.description (描述) 國立政治大學zh_TW
dc.description (描述) 企業管理研究所zh_TW
dc.description (描述) 101355055zh_TW
dc.description (描述) 102zh_TW
dc.description.abstract (摘要) 預測模型已經被廣泛運用在日常生活中，例如銀行信用評比、消費者行為或是疾病的預測等等。然而不論在建構或使用預測模型的時候，我們都會在訓練資料或是測試資料中遇到遺失值的問題，因而降低預測的表現。面對遺失值有很多種處理方式，刪除、填補、模型建構以及機器學習都是可以使用的方法；除此之外，直接用某個成本去取得遺失值也是一個選擇。\n本研究著重的議題是用某成本去取得遺失值，並且利用決策樹(因為其在建構時可以容納遺失值)來當作預測模型，希望可以找到用較低的成本的填值方法達到較高的準確率。我們延續過去Error Sampling中Uncertainty Score的概念與邏輯。提出U-Sampling來判斷不同特徵值的「重要性排序」。相較於過去Error Sampling用「受試者」(row-based)的重要性來排序。U-Sampling是根據「特徵值」(column-based)的重要性來排序。\n我們用8組UCI machine Learning Repository的資料進行兩組實驗，分別讓訓練資料以及測試資料含有一定比例的遺失值。再利用U-Sampling、Random Sampling以及過去文獻所提及的Error Sampling作準確率和錯誤減少率的比較。實驗結果顯示在訓練資料有遺失值的情況，U-Sampling在70%以上的檔案表現較佳；而在測試資料有遺失值的情況，U-Sampling則是在87.5%的檔案表現較佳。\n另外，我們也研究了對於不同的遺失比例對於上述方法的效果是否有影響，可以用來判斷哪種情況比較適用哪一種選值方法。希望透過U-Sampling，可以先挑選重要的特徵值來填補，用較少的遺失值取得就得到較高的準確率，也因此可以節省處理遺失值的成本。zh_TW
dc.description.tableofcontents 目錄\n致謝詞 i\n摘要 ii\nAbstract iii\n表目錄 vi\n圖目錄 vii\n第一章 緒論 1\n第一節 研究背景 1\n第二節 研究動機與目的 2\n第三節 研究架構 3\n第四節 研究結果與貢獻 4\n第五節 論文架構 4\n第二章 文獻回顧 6\n第一節 遺失值 6\n2.1.1 遺失值的數量 6\n2.1.2 遺失值的型態 6\n2.1.3 遺失值處理方式 7\n第二節 決策樹的原理 13\n第三章 研究方法 19\n第一節 研究架構 19\n第二節 U-Sampling說明 20\n3.2.1 研究想法 20\n3.2.2 程式撰寫 22\n3.2.3 假設與限制 22\n第三節 U-Sampling評估 23\n3.3.1 評估指標 23\n3.3.2 檢驗方式 24\n3.3.3 和Random Sampling、Error Sampling的比較 25\n3.3.4 不同資料型態比較 25\n第四章 研究結果 26\n第一節 實驗數據 26\n4.1.1 研究資料介紹 26\n第二節 實驗結果 28\n4.2.1 十折交叉驗證-training data含有遺失值 28\n4.2.2 十折交叉驗證-test data含有遺失值 35\n4.2.4 不同資料型態 40\n第五章 結論與建議 42\n第一節 結論 42\n第二節 研究限制及建議 43\n5.2.1 研究限制 43\n5.2.2 未來建議 43\n參考文獻 44zh_TW
dc.format.extent 1134186 bytes-
dc.format.mimetype application/pdf-
dc.language.iso en_US-
dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0101355055en_US
dc.subject (關鍵詞) 遺失值zh_TW
dc.subject (關鍵詞) 決策樹分類zh_TW
dc.subject (關鍵詞) uncertainty scoreen_US
dc.subject (關鍵詞) decision treeen_US
dc.subject (關鍵詞) missing value acquisitionen_US
dc.title (題名) 預測模型中遺失值之選填順序研究zh_TW
dc.title (題名) Research of acquisition order of missing values in predictive modelen_US
dc.type (資料類型) thesisen
dc.relation.reference (參考文獻) 1. Allison, P. D. (2001). Missing data. Thousand Oaks, CA: Sage.\n\n2. Alpaydın, E. (2010). Introduction to machine learning. London, England: The MIT Press.\n\n3. Bennett, D. A. (2001). How can I deal with missing data in my study? Australian and New Zealand Journal of Public Health, 25(5), 464–469.\n\n4. Giks, Walter R ; Richardson, Sylvia; Spiegelhalter, David J. (1996). Introducing Markov chain Monte Carlo. In Markov chain Monte Carlo in practice (pp. 1-19). London: Chapman & hall/CRC.\n\n5. Graham, J. W. (2003). Adding missing-data-relevant variables to FIML basedstructural equation models. Structural Equation Modeling, pp. 10, 80–100.\n\n6. Jackson, J. (2002). Overview, data mining: a conceptual. Communications of the Association for Information Systems.\n\n7. Kohavi, R. (1995). A study of Cross-Validation and Bootstrap for Accuracy Estimation and Model Selection. IJCAI, (Vol.14, No.2, pp. 1137-1145).\n\n8. Levin, N., & Zahav, J. (2001, Spring). Predictive modeling using segmentation. Journal of Interactive Marketing, 15(2), 2-22.\n\n9. Melville, P., Saar-Tsechansky, M., Provost, F., & Mooney, R. (2004). Active Feature-Value Acquisition for Classifier Induction. Proceedings of the 4th IEEE International Conference on Data Mining, (pp. 483-486). Brighton, UK.\n\n10. Pallant, J. (2007). SPSS survival manual (3rd ed.). New York, NY: Open University Press.\n\n11. Pedro J. Garcı´a-Laencina Æ Jose´-Luis Sancho-Go´mez Æ, A. R.-V. (2010). Pattern classification with missing data: a review. Neural Comput & Applic.\n\n12. Peng, C. Y. J., Harwell, M., Liou, S.M., & Ehman, L.H. (2006). Advances in missing data methods and implications for educational research. In Real data analysis, 31-78. North Carolina,US : Information Age Publishing.\n\n13. Quinlan, J. R. (1989). Unknown attribute values in induction., In ML (pp. 164-168).\n\n14. Rubin, D. B. (1987). Multiple imputation for non-response in surveys. New York: John Wiley & Sons.\n\n15. Saar-Tsechansky, M., Melville, P., & Provost, F. (2009, 4). Active Feature-Value Acqusition. Management Science, 55(4), 664-684.\n\n16. Schafer, J. L. (1999). Multiple imputation: a primer. Statiscal methods in medical research, 8(1), 3-15.\n\n17. Schlomer, G. L., Bauman, S., & Card, N. A. (2010). Best Practices for Missing Data Management in Counseling Psychology. Journal of Counseling Psychology, 57(1), 1-10.\n\n18. Simon, H. A., & Lea, G. (1974). Problem solving and rule induction: A unified view. Knowledge and cognition. Oxford, England: Lawrence Erlbaum.\n\n19. Simon, H., & Lea, G. (1974). Problem solving and rule induction: A unified view. \n\n20. Turney, P. (2000). Types of Cost in Inductive Concept Learning. Proceedings of the Cost-Sensitive Learning Workshop at the 17th ICML-2000 Conference. Stanford, CA.\n\n\n \n\n21. Vinod, N. C., & Punithavalli, D. M. (2011). Classification of Incomplete Data Handling Techniques-An Overview. International Journal on Computer Science and Engineering, 3(1), 340-344.\n\n22. Zheng, Z., & Padmanabhan, B. (2002). On Active Learning for Data Acquisition. Proceedings of IEEE International Condference on Data Mining, (pp. 562-569).\n\n網路資料\n\n1. UCI machine Learning Repository. (n.d.). Retrieved from http://archive.ics.uci.edu/ml/zh_TW