學術產出-學位論文
文章檢視/開啟
書目匯出
-
題名 基於類別異質性結構的監督式學習
Supervised Learning with Potential Label Heterogeneity作者 黃俊嘉
Huang, Jiun-Jia貢獻者 周珮婷
黃俊嘉
Huang, Jiun-Jia關鍵詞 潛在類別
異質性
分類預測
Potential categories
Heterogeneity
Classification prediction日期 2021 上傳時間 1-七月-2021 17:30:54 (UTC+8) 摘要 在過往的研究中,研究人員常常將研究重心放在找出資料的潛在類別,並透過資料的異質性結構定義出新類別,於是本研究提出基於同類別的異質性結構將各個原始類別拆成數個子類別以提高分類預測準確率。本研究以建立標籤內嵌樹的方法進行分類預測,此分類預測方法是先計算類別支配矩陣後,藉由此矩陣進行階層式分群,並由偽概似機率進行分類預測,而本研究比較使用此方法和常見分類器的分類預測表現差異,也比較常見分類器在使用子類別及原始類別的分類預測差異。研究結果顯示所提出的子類別方法,在異質性資料確實會擁有較高的預測正確率。另外,本研究發現在多數的分類器,以子類別預測能提升分類表現,但是需要考慮資料本身是否含有異質性結構。
In previous studies, researchers often focused their research on identifying potential categories of data and defining new categories through the heterogeneous structure of the data. Therefore, in this study, the original categories were divided into sub-categories based on the heterogeneous structure of the same category. Each sub-category is then classified and predicted by the method used in this study. This classification prediction method calculates the label dominance matrix, uses the matrix to group hierarchically, and uses the pseudo-likelihood probability to perform classification prediction. This research will compare the prediction accuracy rates of the common classifiers that use the original categories for classification prediction and this proposed method that uses the subcategories. The research results show that this proposed method will indeed have better results in some datasets. In addition, this study also compared whether the classification prediction using the split into sub-categories and the classification prediction using the original category will increase the accuracy of the prediction of various classifiers. It turns out that most of the classifiers have an improvement. Nevertheless, we need to consider if a heterogeneous structure exists in a category first before applying the proposed method.參考文獻 王郁琮. (2014). 台灣青少年異質性憂鬱發展軌跡之性別差異及與違常行為之關係. 中華心理衛生學刊, 27(1), 97-130.尹霞雲、朱翠英、黎志華、蔡泰生. (2014). 留守兒童情緒和行為問題特徵的潛在類別分析:基於個體為中心的研究視角. 心理科學, 37(2), 329-334.陳杏佳. (2005). 精神分裂症住院病患攻擊行為之異質性與危險因子研究. 臺灣大學護理學研究所學位論文, 1-181.Allwein, E. L., Schapire, R. E., & Singer, Y. (2000). Reducing multiclass to binary: Aunifying approach for margin classifiers. Journal of machine learning research, 1(Dec), 113-141.Allsopp, K., Read, J., Corcoran, R., & Kinderman, P. (2019). Heterogeneity in psychiatric diagnostic classification. Psychiatry research, 279, 15-22.Blanco-Calvo, M., Concha, Á., Figueroa, A., Garrido, F., & Valladares-Ayerbes, M. (2015). Colorectal cancer classification and cell heterogeneity: a systems oncology approach. International journal of molecular sciences, 16(6), 13610-13632.Caliński, T., & Harabasz, J. (1974). A dendrite method for cluster analysis. Communications in Statistics-theory and Methods, 3(1), 1-27.Dahl, K. H., Simonsen, G. S., Olsvik, Ø., & Sundsfjord, A. (1999). Heterogeneity inthe vanB gene cluster of genomically diverse clinical strains of vancomycin-resistant enterococci. Antimicrobial agents and chemotherapy, 43(5), 1105-1110.Farmer, A. E., McGuffin, P., & Spitznagel, E. L. (1983). Heterogeneity in schizophrenia: a cluster-analytic approach. Psychiatry Research, 8(1), 1-12.Fitzpatrick, A. M., Teague, W. G., Meyers, D. A., Peters, S. P., Li, X., Li, H., ... & National Institutes of Health. (2011). Heterogeneity of severe asthma in childhood: confirmation by cluster analysis of children in the National Institutes of Health/National Heart, Lung, and Blood Institute Severe Asthma Research Program. Journal of allergy and clinical immunology, 127(2), 382-389.Hastie, T., & Tibshirani, R. (1998). Classification by pairwise coupling. Paper presented at the Advances in neural information processing systems.Hsieh, F., & Chou, E. P. (2021). Categorical Exploratory Data Analysis: From Multiclass Classification and Response Manifold Analytics Perspectives of Baseball Pitching Dynamics. Entropy, 23(7), 792.Krzanowski, W. J., & Lai, Y. T. (1988). A criterion for determining the number of groups in a data set using sum-of-squares clustering. Biometrics, 23-34.Krijthe, J.H. & Loog, M. (2015). Implicitly Constrained Semi-Supervised Least Squares Classification. In E. Fromont, T. de Bie, & M. van Leeuwen, eds. 14th International Symposium on Advances in Intelligent Data Analysis XIV (Lecture Notes in Computer Science Volume 9385). Saint Etienne. France, pp. 158-169.Lanza, S. T., & Rhoades, B. L. (2013). Latent class analysis: an alternative perspective on subgroup analysis in prevention and treatment. Prevention Science, 14(2), 157-168.Leisch F, Dimitriadou E (2021). mlbench: Machine Learning Benchmark Problems. R package version 2.1-3.MacQueen, J. (1967, June). Some methods for classification and analysis of multivariate observations. In Proceedings of the fifth Berkeley symposium on mathematical statistics and probability (Vol. 1, No. 14, pp. 281-297).Milligan, G. W., & Cooper, M. C. (1985). An examination of procedures for determining the number of clusters in a data set. Psychometrika, 50(2), 159-179.Rifkin, R., & Klautau, A. (2004). In defense of one-vs-all classification. Journal of machine learning research, 5(Jan), 101-141.Reser, M. P., Allott, K. A., Killackey, E., Farhall, J., & Cotton, S. M. (2015). Exploring cognitive heterogeneity in first-episode psychosis: What cluster analysis can reveal. Psychiatry Research, 229(3), 819-827.Shaw, S. Y., Shah, L., Jolly, A. M., & Wylie, J. L. (2008). Identifying heterogeneity among injection drug users: a cluster analysis approach. American Journal of Public Health, 98(8), 1430-1437.Ward Jr, J. H. (1963). Hierarchical grouping to optimize an objective function. Journal of the American statistical association, 58(301), 236-244.Wanberg, C. R., & Marchese, M. C. (1994). Heterogeneity in the Unemployment Experience: A Cluster Analytic Investigation 1. Journal of Applied Social Psychology, 24(6), 473-488. 描述 碩士
國立政治大學
統計學系
108354013資料來源 http://thesis.lib.nccu.edu.tw/record/#G0108354013 資料類型 thesis dc.contributor.advisor 周珮婷 zh_TW dc.contributor.author (Authors) 黃俊嘉 zh_TW dc.contributor.author (Authors) Huang, Jiun-Jia en_US dc.creator (作者) 黃俊嘉 zh_TW dc.creator (作者) Huang, Jiun-Jia en_US dc.date (日期) 2021 en_US dc.date.accessioned 1-七月-2021 17:30:54 (UTC+8) - dc.date.available 1-七月-2021 17:30:54 (UTC+8) - dc.date.issued (上傳時間) 1-七月-2021 17:30:54 (UTC+8) - dc.identifier (Other Identifiers) G0108354013 en_US dc.identifier.uri (URI) http://nccur.lib.nccu.edu.tw/handle/140.119/135930 - 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) 108354013 zh_TW dc.description.abstract (摘要) 在過往的研究中,研究人員常常將研究重心放在找出資料的潛在類別,並透過資料的異質性結構定義出新類別,於是本研究提出基於同類別的異質性結構將各個原始類別拆成數個子類別以提高分類預測準確率。本研究以建立標籤內嵌樹的方法進行分類預測,此分類預測方法是先計算類別支配矩陣後,藉由此矩陣進行階層式分群,並由偽概似機率進行分類預測,而本研究比較使用此方法和常見分類器的分類預測表現差異,也比較常見分類器在使用子類別及原始類別的分類預測差異。研究結果顯示所提出的子類別方法,在異質性資料確實會擁有較高的預測正確率。另外,本研究發現在多數的分類器,以子類別預測能提升分類表現,但是需要考慮資料本身是否含有異質性結構。 zh_TW dc.description.abstract (摘要) In previous studies, researchers often focused their research on identifying potential categories of data and defining new categories through the heterogeneous structure of the data. Therefore, in this study, the original categories were divided into sub-categories based on the heterogeneous structure of the same category. Each sub-category is then classified and predicted by the method used in this study. This classification prediction method calculates the label dominance matrix, uses the matrix to group hierarchically, and uses the pseudo-likelihood probability to perform classification prediction. This research will compare the prediction accuracy rates of the common classifiers that use the original categories for classification prediction and this proposed method that uses the subcategories. The research results show that this proposed method will indeed have better results in some datasets. In addition, this study also compared whether the classification prediction using the split into sub-categories and the classification prediction using the original category will increase the accuracy of the prediction of various classifiers. It turns out that most of the classifiers have an improvement. Nevertheless, we need to consider if a heterogeneous structure exists in a category first before applying the proposed method. en_US dc.description.tableofcontents 第一章 緒論 8第二章 文獻回顧 10第三章 研究方法 14第一節 標籤內嵌樹 14一、 建立Triplet以及矩陣H 14二、 計算矩陣H內之機率值 14三、 透過矩陣H判斷距離並且進行新資料點分類 15第二節 資料處理及研究流程 16一、 研究工具 16二、 研究流程 17第四章 資料介紹 19一、 Two Half-moon Dataset 19二、 Two Spiral Dataset 20三、 Mixture of normal dataset 20四、 Multivariate Normal Distribution Dataset 21五、 Speaker Accent Recognition Dataset 22六、 Iris Dataset 22第五章 研究結果 23第一節 比較子類別預測與其他分類器原始類別預測 23一、 Two Half-moon Dataset 23二、 Two Spiral Dataset 27三、 Mixture of normal dataset 30四、 Multivariate Normal Distribution Dataset 33五、 Speaker Accent Recognition Dataset 35六、 Iris Dataset 39第二節 比較子類別分類預測與原始類別分類預測 41一、 Two Half-moon Dataset 42二、 Two Spiral Dataset 42三、 Mixture of normal dataset 43四、 Multivariate Normal Distribution Dataset 44五、 Speaker Accent Recognition Dataset 44六、 Iris Dataset 45第六章 結論與建議 46第七章 參考文獻 47 zh_TW dc.format.extent 1980620 bytes - dc.format.mimetype application/pdf - dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0108354013 en_US dc.subject (關鍵詞) 潛在類別 zh_TW dc.subject (關鍵詞) 異質性 zh_TW dc.subject (關鍵詞) 分類預測 zh_TW dc.subject (關鍵詞) Potential categories en_US dc.subject (關鍵詞) Heterogeneity en_US dc.subject (關鍵詞) Classification prediction en_US dc.title (題名) 基於類別異質性結構的監督式學習 zh_TW dc.title (題名) Supervised Learning with Potential Label Heterogeneity en_US dc.type (資料類型) thesis en_US dc.relation.reference (參考文獻) 王郁琮. (2014). 台灣青少年異質性憂鬱發展軌跡之性別差異及與違常行為之關係. 中華心理衛生學刊, 27(1), 97-130.尹霞雲、朱翠英、黎志華、蔡泰生. (2014). 留守兒童情緒和行為問題特徵的潛在類別分析:基於個體為中心的研究視角. 心理科學, 37(2), 329-334.陳杏佳. (2005). 精神分裂症住院病患攻擊行為之異質性與危險因子研究. 臺灣大學護理學研究所學位論文, 1-181.Allwein, E. L., Schapire, R. E., & Singer, Y. (2000). Reducing multiclass to binary: Aunifying approach for margin classifiers. Journal of machine learning research, 1(Dec), 113-141.Allsopp, K., Read, J., Corcoran, R., & Kinderman, P. (2019). Heterogeneity in psychiatric diagnostic classification. Psychiatry research, 279, 15-22.Blanco-Calvo, M., Concha, Á., Figueroa, A., Garrido, F., & Valladares-Ayerbes, M. (2015). Colorectal cancer classification and cell heterogeneity: a systems oncology approach. International journal of molecular sciences, 16(6), 13610-13632.Caliński, T., & Harabasz, J. (1974). A dendrite method for cluster analysis. Communications in Statistics-theory and Methods, 3(1), 1-27.Dahl, K. H., Simonsen, G. S., Olsvik, Ø., & Sundsfjord, A. (1999). Heterogeneity inthe vanB gene cluster of genomically diverse clinical strains of vancomycin-resistant enterococci. Antimicrobial agents and chemotherapy, 43(5), 1105-1110.Farmer, A. E., McGuffin, P., & Spitznagel, E. L. (1983). Heterogeneity in schizophrenia: a cluster-analytic approach. Psychiatry Research, 8(1), 1-12.Fitzpatrick, A. M., Teague, W. G., Meyers, D. A., Peters, S. P., Li, X., Li, H., ... & National Institutes of Health. (2011). Heterogeneity of severe asthma in childhood: confirmation by cluster analysis of children in the National Institutes of Health/National Heart, Lung, and Blood Institute Severe Asthma Research Program. Journal of allergy and clinical immunology, 127(2), 382-389.Hastie, T., & Tibshirani, R. (1998). Classification by pairwise coupling. Paper presented at the Advances in neural information processing systems.Hsieh, F., & Chou, E. P. (2021). Categorical Exploratory Data Analysis: From Multiclass Classification and Response Manifold Analytics Perspectives of Baseball Pitching Dynamics. Entropy, 23(7), 792.Krzanowski, W. J., & Lai, Y. T. (1988). A criterion for determining the number of groups in a data set using sum-of-squares clustering. Biometrics, 23-34.Krijthe, J.H. & Loog, M. (2015). Implicitly Constrained Semi-Supervised Least Squares Classification. In E. Fromont, T. de Bie, & M. van Leeuwen, eds. 14th International Symposium on Advances in Intelligent Data Analysis XIV (Lecture Notes in Computer Science Volume 9385). Saint Etienne. France, pp. 158-169.Lanza, S. T., & Rhoades, B. L. (2013). Latent class analysis: an alternative perspective on subgroup analysis in prevention and treatment. Prevention Science, 14(2), 157-168.Leisch F, Dimitriadou E (2021). mlbench: Machine Learning Benchmark Problems. R package version 2.1-3.MacQueen, J. (1967, June). Some methods for classification and analysis of multivariate observations. In Proceedings of the fifth Berkeley symposium on mathematical statistics and probability (Vol. 1, No. 14, pp. 281-297).Milligan, G. W., & Cooper, M. C. (1985). An examination of procedures for determining the number of clusters in a data set. Psychometrika, 50(2), 159-179.Rifkin, R., & Klautau, A. (2004). In defense of one-vs-all classification. Journal of machine learning research, 5(Jan), 101-141.Reser, M. P., Allott, K. A., Killackey, E., Farhall, J., & Cotton, S. M. (2015). Exploring cognitive heterogeneity in first-episode psychosis: What cluster analysis can reveal. Psychiatry Research, 229(3), 819-827.Shaw, S. Y., Shah, L., Jolly, A. M., & Wylie, J. L. (2008). Identifying heterogeneity among injection drug users: a cluster analysis approach. American Journal of Public Health, 98(8), 1430-1437.Ward Jr, J. H. (1963). Hierarchical grouping to optimize an objective function. Journal of the American statistical association, 58(301), 236-244.Wanberg, C. R., & Marchese, M. C. (1994). Heterogeneity in the Unemployment Experience: A Cluster Analytic Investigation 1. Journal of Applied Social Psychology, 24(6), 473-488. zh_TW dc.identifier.doi (DOI) 10.6814/NCCU202100555 en_US