學術產出-Theses
Article View/Open
Publication Export
-
題名 不平衡資料之數據驅動混合監督式學習方法
Data-driven Hybrid Approach for Imbalanced Data in Supervised Learning作者 劉得心
Liu, Te-Hsin貢獻者 周珮婷
Chou, Pei-Ting
劉得心
Liu, Te-Hsin關鍵詞 不平衡資料
監督式學習
PLR
二元分類問題
Imbalanced data
Supervised learning
PLR
Binary classification日期 2022 上傳時間 1-Jul-2022 16:58:02 (UTC+8) 摘要 不平衡資料意指資料中有特定類別的樣本個數特別少,造成各類別比例懸殊,此資料特性易使監督式學習的分類模型在訓練時,無法有效地學習少數類別的特徵,導致模型預測錯誤。為解決此問題,本研究嘗試對監督式學習方法Pseudo-Likelihood Ratio(PLR)進行兩種不同的調整,並分別提出調整後的分類模型;為了探討兩種分類模型在不同不平衡比例下的分類效能,本研究將調整後的兩個分類模型與原始PLR、KNN、SVM三個模型,對不同不平衡比例的資料集進行分類預測,以此比較五種模型在不同不平衡比例下的分類效能。最後研究顯示,本研究針對PLR所提出之改善方法,在不同資料集中的表現有所不同,但整體而言,對提升原始PLR分類效能是有所成效的。
Imbalanced data means that the number of specific categories in the data is very small, resulting in a disparity in the proportion of each category. This data characteristic easily makes the supervised learning classification model unable to effectively learn the features of a few categories during training, resulting in model prediction error. In order to solve this problem, this study attempts to make two different adjustments to the supervised learning method Pseudo-Likelihood Ratio (PLR), and propose the adjusted classification models respectively; in order to explore the classification accuracy of the two classification models under various imbalance ratios, the adjusted two classification models and the original PLR, KNN, and SVM were put into each imbalanced proportion of the five data sets for classification, so as to compare the classification performance of the five models. The result shows that the improvement methods proposed in this study for PLR have different performances in different data sets. Still, on the whole, it is effective in improving the classification performance of the original PLR.參考文獻 Akbani, R., Kwek, S., & Japkowicz, N. (2004). Applying support vector machines to imbalanced datasets. Paper presented at the European conference on machine learning.Chawla, N. V., Bowyer, K. W., Hall, L. O., & Kegelmeyer, W. P. (2002). SMOTE: synthetic minority over-sampling technique. Journal of artificial intelligence research, 16, pp. 321- 357.Cortes, C., & Vapnik, V. (1995). Support-vector networks. Machine learning, 20(3), pp. 273-297.Cover, T., & Hart, P. (1967). Nearest neighbor pattern classification. IEEE. Transactions on Information Theory, 13(1), pp. 21-27Elizabeth P. Chou & Shan-Ping Yang. (2022). A virtual multi-label approach to imbalanced data classification. Communications in Statistics - Simulation and Computation, DOI: 10.1080/03610918. 2022.2049820Fushing Hsieh, Elizabeth Chou. (2020). Categorical Exploratory Data Analysis: From Multiclass Classification and Response Manifold Analytics Perspectives of Baseball Pitching Dynamics, Entropy, 23(7), pp. 792, 23, 7-792.He, H., & Garcia, E. A. (2009). Learning from imbalanced data. IEEE Transactions. on knowledge and data engineering, 21(9), pp. 1263-1284.Hong, X., Chen, S., & Harris, C. J. (2007). A kernel-based two-class classifier for imbalanced data sets. IEEE Transactions on neural networks, 18(1), pp. 28-41.Seliya, N., Khoshgoftaar, T. M., & Hulse, J. V. (2009). A Study on the Relationships of Classifier Performance Metrics. IEEE International Conference on Tools with Artificial Intelligence, pp. 59-66.Japkowicz, N., & Stephen, S. (2002). The class imbalance problem: A systematic study. Intelligent data analysis, 6(5), pp. 429-449.J. B. Brown. (2018). Classifiers and their Metrics Quantified. Molecular Informatics, 37(1-2), p. 1700127.Kang, P., & Cho, S. (2006). EUS SVMs: Ensemble of under-sampled SVMs for data imbalance problems. Paper presented at the International Conference on Neural Information Processing.Raskutti, B., & Kowalczyk, A. (2004). Extreme re-balancing for SVMs: a case study. ACM Sigkdd Explorations Newsletter, 6(1), pp. 60-69.Lee, H.-j., & Cho, S. (2006). The novelty detection approach for different degrees of. class imbalance. Paper presented at the International conference on neural information processing.Liu, Y., An, A., & Huang, X. (2006). Boosting prediction accuracy on imbalanced datasets with SVM ensembles. Paper presented at the Pacific-Asia Conference on Knowledge Discovery and Data Mining.Chawla, N.V., Japkowicz, N., & Kolcz, A. (2004). Editorial: Special Issue on. Learning from Imbalanced Data Sets. ACM SIGKDD Explorations Newsletter, 6(1), pp. 1-6.Tang, Y., & Zhang, Y.-Q. (2006). Granular SVM with repetitive undersampling for highly imbalanced protein homology prediction. Paper presented at the 2006 IEEE International Conference on Granular Computing.Wang, B. X., & Japkowicz, N. (2008). Boosting support vector machines for imbalanced data sets. Paper presented at the International Symposium on Methodologies for Intelligent Systems.Fan, W., Stolfo, S.J., Zhang, J., & Chan, P.K. (1999). AdaCost: Misclassification Cost-Sensitive Boosting. Proc. Int’l Conf. Machine Learning, pp. 97-105.Zou, K. H., O’Malley, A. J., & Mauri, L. (2007). Receiver-operating characteristic analysis for evaluating diagnostic tests and predictive models. Circulation, 115(5), pp. 654-657. 描述 碩士
國立政治大學
統計學系
109354020資料來源 http://thesis.lib.nccu.edu.tw/record/#G0109354020 資料類型 thesis dc.contributor.advisor 周珮婷 zh_TW dc.contributor.advisor Chou, Pei-Ting en_US dc.contributor.author (Authors) 劉得心 zh_TW dc.contributor.author (Authors) Liu, Te-Hsin en_US dc.creator (作者) 劉得心 zh_TW dc.creator (作者) Liu, Te-Hsin en_US dc.date (日期) 2022 en_US dc.date.accessioned 1-Jul-2022 16:58:02 (UTC+8) - dc.date.available 1-Jul-2022 16:58:02 (UTC+8) - dc.date.issued (上傳時間) 1-Jul-2022 16:58:02 (UTC+8) - dc.identifier (Other Identifiers) G0109354020 en_US dc.identifier.uri (URI) http://nccur.lib.nccu.edu.tw/handle/140.119/140753 - dc.description (描述) 碩士 zh_TW dc.description (描述) 國立政治大學 zh_TW dc.description (描述) 統計學系 zh_TW dc.description (描述) 109354020 zh_TW dc.description.abstract (摘要) 不平衡資料意指資料中有特定類別的樣本個數特別少,造成各類別比例懸殊,此資料特性易使監督式學習的分類模型在訓練時,無法有效地學習少數類別的特徵,導致模型預測錯誤。為解決此問題,本研究嘗試對監督式學習方法Pseudo-Likelihood Ratio(PLR)進行兩種不同的調整,並分別提出調整後的分類模型;為了探討兩種分類模型在不同不平衡比例下的分類效能,本研究將調整後的兩個分類模型與原始PLR、KNN、SVM三個模型,對不同不平衡比例的資料集進行分類預測,以此比較五種模型在不同不平衡比例下的分類效能。最後研究顯示,本研究針對PLR所提出之改善方法,在不同資料集中的表現有所不同,但整體而言,對提升原始PLR分類效能是有所成效的。 zh_TW dc.description.abstract (摘要) Imbalanced data means that the number of specific categories in the data is very small, resulting in a disparity in the proportion of each category. This data characteristic easily makes the supervised learning classification model unable to effectively learn the features of a few categories during training, resulting in model prediction error. In order to solve this problem, this study attempts to make two different adjustments to the supervised learning method Pseudo-Likelihood Ratio (PLR), and propose the adjusted classification models respectively; in order to explore the classification accuracy of the two classification models under various imbalance ratios, the adjusted two classification models and the original PLR, KNN, and SVM were put into each imbalanced proportion of the five data sets for classification, so as to compare the classification performance of the five models. The result shows that the improvement methods proposed in this study for PLR have different performances in different data sets. Still, on the whole, it is effective in improving the classification performance of the original PLR. en_US dc.description.tableofcontents 摘要 iAbstract ii目 次 iii表次 v圖次 vii第一章 緒論 1第一節 研究背景 1第二節 研究目的 1第二章 文獻回顧 3第一節 分類模型 3第二節 評估指標 5第三節 研究方法 6第三章 研究方法與過程 7第一節 分類模型 7第二節 Pseudo-Likelihood Ratio 9第三節 研究方法 10第四節 評估指標 12第四章 研究結果與分析 16第一節 資料介紹 16第二節 不平衡資料模擬 17第三節 實驗流程 18第四節 實驗結果與分析 18第五章 結論與建議 28第一節 結論 28第二節 建議 29參考文獻 30附錄 32附錄ㄧ 不平衡資料模擬結果 32附錄二 Ionosphere資料集實驗結果 35附錄三 Abalone資料集實驗結果 40附錄四 Epileptic Seizure Recognition資料集實驗結果 45附錄五 Wave Database Generator資料集實驗結果 50附錄六 Credit Card資料集實驗結果 55 zh_TW dc.format.extent 1280001 bytes - dc.format.mimetype application/pdf - dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0109354020 en_US dc.subject (關鍵詞) 不平衡資料 zh_TW dc.subject (關鍵詞) 監督式學習 zh_TW dc.subject (關鍵詞) PLR zh_TW dc.subject (關鍵詞) 二元分類問題 zh_TW dc.subject (關鍵詞) Imbalanced data en_US dc.subject (關鍵詞) Supervised learning en_US dc.subject (關鍵詞) PLR en_US dc.subject (關鍵詞) Binary classification en_US dc.title (題名) 不平衡資料之數據驅動混合監督式學習方法 zh_TW dc.title (題名) Data-driven Hybrid Approach for Imbalanced Data in Supervised Learning en_US dc.type (資料類型) thesis en_US dc.relation.reference (參考文獻) Akbani, R., Kwek, S., & Japkowicz, N. (2004). Applying support vector machines to imbalanced datasets. Paper presented at the European conference on machine learning.Chawla, N. V., Bowyer, K. W., Hall, L. O., & Kegelmeyer, W. P. (2002). SMOTE: synthetic minority over-sampling technique. Journal of artificial intelligence research, 16, pp. 321- 357.Cortes, C., & Vapnik, V. (1995). Support-vector networks. Machine learning, 20(3), pp. 273-297.Cover, T., & Hart, P. (1967). Nearest neighbor pattern classification. IEEE. Transactions on Information Theory, 13(1), pp. 21-27Elizabeth P. Chou & Shan-Ping Yang. (2022). A virtual multi-label approach to imbalanced data classification. Communications in Statistics - Simulation and Computation, DOI: 10.1080/03610918. 2022.2049820Fushing Hsieh, Elizabeth Chou. (2020). Categorical Exploratory Data Analysis: From Multiclass Classification and Response Manifold Analytics Perspectives of Baseball Pitching Dynamics, Entropy, 23(7), pp. 792, 23, 7-792.He, H., & Garcia, E. A. (2009). Learning from imbalanced data. IEEE Transactions. on knowledge and data engineering, 21(9), pp. 1263-1284.Hong, X., Chen, S., & Harris, C. J. (2007). A kernel-based two-class classifier for imbalanced data sets. IEEE Transactions on neural networks, 18(1), pp. 28-41.Seliya, N., Khoshgoftaar, T. M., & Hulse, J. V. (2009). A Study on the Relationships of Classifier Performance Metrics. IEEE International Conference on Tools with Artificial Intelligence, pp. 59-66.Japkowicz, N., & Stephen, S. (2002). The class imbalance problem: A systematic study. Intelligent data analysis, 6(5), pp. 429-449.J. B. Brown. (2018). Classifiers and their Metrics Quantified. Molecular Informatics, 37(1-2), p. 1700127.Kang, P., & Cho, S. (2006). EUS SVMs: Ensemble of under-sampled SVMs for data imbalance problems. Paper presented at the International Conference on Neural Information Processing.Raskutti, B., & Kowalczyk, A. (2004). Extreme re-balancing for SVMs: a case study. ACM Sigkdd Explorations Newsletter, 6(1), pp. 60-69.Lee, H.-j., & Cho, S. (2006). The novelty detection approach for different degrees of. class imbalance. Paper presented at the International conference on neural information processing.Liu, Y., An, A., & Huang, X. (2006). Boosting prediction accuracy on imbalanced datasets with SVM ensembles. Paper presented at the Pacific-Asia Conference on Knowledge Discovery and Data Mining.Chawla, N.V., Japkowicz, N., & Kolcz, A. (2004). Editorial: Special Issue on. Learning from Imbalanced Data Sets. ACM SIGKDD Explorations Newsletter, 6(1), pp. 1-6.Tang, Y., & Zhang, Y.-Q. (2006). Granular SVM with repetitive undersampling for highly imbalanced protein homology prediction. Paper presented at the 2006 IEEE International Conference on Granular Computing.Wang, B. X., & Japkowicz, N. (2008). Boosting support vector machines for imbalanced data sets. Paper presented at the International Symposium on Methodologies for Intelligent Systems.Fan, W., Stolfo, S.J., Zhang, J., & Chan, P.K. (1999). AdaCost: Misclassification Cost-Sensitive Boosting. Proc. Int’l Conf. Machine Learning, pp. 97-105.Zou, K. H., O’Malley, A. J., & Mauri, L. (2007). Receiver-operating characteristic analysis for evaluating diagnostic tests and predictive models. Circulation, 115(5), pp. 654-657. zh_TW dc.identifier.doi (DOI) 10.6814/NCCU202200484 en_US