請用此 Handle URI 來引用此文件：
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/79248完整後設資料紀錄
| DC 欄位 | 值 | 語言 |
|---|---|---|
| dc.contributor.advisor | 林守德(Shou-De Lin) | |
| dc.contributor.author | Yi-Fu Fu | en |
| dc.contributor.author | 傅羿夫 | zh_TW |
| dc.date.accessioned | 2022-11-23T08:56:39Z | - |
| dc.date.available | 2022-02-21 | |
| dc.date.available | 2022-11-23T08:56:39Z | - |
| dc.date.copyright | 2022-02-21 | |
| dc.date.issued | 2021 | |
| dc.date.submitted | 2022-01-19 | |
| dc.identifier.citation | Diemert Eustache, Betlei Artem, C. Renaudin, and A. MassihReza. A large scale benchmark for uplift modeling. In Proceedings of the AdKDD and TargetAd Work shop, KDD, London,United Kingdom, August, 20, 2018. ACM, 2018. R.Gubela,A.Bequé,S.Lessmann,andF.Gebert.Conversionupliftinecommerce: A systematic benchmark of modeling strategies. International Journal of Information Technology Decision Making, 18(03):747–791, 2019. P. Gutierrez and J.Y. Gérardy. Causal inference and uplift modelling: A review of the literature. In International Conference on Predictive Applications and APIs, pages 1–13. PMLR, 2017. B. Hansotia and B. Rukstales. Incremental value modeling. Journal of Interactive Marketing, 16(3):35, 2002. M. Jaskowski and S. Jaroszewicz. Uplift modeling for clinical trial data. In ICML Workshop on Clinical Data Analysis, volume 46, 2012. G. Ke, Q. Meng, T. Finley, T. Wang, W. Chen, W. Ma, Q. Ye, and T.Y. Liu. Light gbm: A highly efficient gradient boosting decision tree. Advances in neural infor mation processing systems, 30:3146–3154, 2017. S. R. Künzel, J. S. Sekhon, P. J. Bickel, and B. Yu. Metalearners for estimating heterogeneous treatment effects using machine learning. Proceedings of the national academy of sciences, 116(10):4156–4165, 2019. F. Kuusisto, V. S. Costa, H. Nassif, E. Burnside, D. Page, and J. Shavlik. Sup port vector machines for differential prediction. In Joint European Conference on Machine Learning and Knowledge Discovery in Databases, pages 50–65. Springer, 2014. V. S. Lo. The true lift model: a novel data mining approach to response modeling in database marketing. ACM SIGKDD Explorations Newsletter, 4(2):78–86, 2002. I. E. Maksim Shevchenko. User guide for uplift modeling and casual inference. https://www.uplift-modeling.com/en/latest/user_guide/index.html, 2020. N. Radcliffe. Using control groups to target on predicted lift: Building and assessing uplift model. Direct Marketing Analytics Journal, pages 14–21, 2007. W. Zhang, J. Li, and L. Liu. A unified survey of treatment effect heterogeneity modelling and uplift modelling. ACM Computing Surveys (CSUR), 54(8):1–36, 2021. | |
| dc.identifier.uri | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/79248 | - |
| dc.description.abstract | 增益模型是機器學習領域中實用的應用之一，目的是預測特定使用 者對於給予處理下其反應的增益。增益模型最大的問題在於無法同時 觀測到使用者給予處理及不給予處理之反應差異。不少研究均針對此 一問題提出對策。本篇研究中指出現行常見之增益模型策略存在一些 問題。同時本篇研究也提出可以應用於任一現有增益模型上之重標籤 方法。研究中進行之線下及線上實驗均證實使用重標籤方法之增益模 型其表現較原增益模型有顯著進步。 | zh_TW |
| dc.description.provenance | Made available in DSpace on 2022-11-23T08:56:39Z (GMT). No. of bitstreams: 1 U0001-1601202215552400.pdf: 2823098 bytes, checksum: f747fa1d97af2b04f5408724c1fe9433 (MD5) Previous issue date: 2021 | en |
| dc.description.tableofcontents | "1 Introduction 1 2 Uplift Models 3 2.1 Twomodelapproach ............................ 3 2.2 Singlemodelapproach ........................... 4 2.3 ClassTransformationmethod........................ 4 2.4 Xlearner .................................. 5 3 Pitfalls of Uplift Models 6 3.1 Indirectoptimization ............................ 6 3.2 Unabletousealldatasimultaneously ................... 6 3.3 OverfittingonPartialInformation ..................... 7 3.3.1 Method............................... 7 3.3.2 Result................................ 8 4 Proposed Method 9 4.1 TrainMultipleSubmodels......................... 10 4.2 Relabeling.................................. 10 4.3 FinalModel................................. 11 4.4 AvoidthePitfalls .............................. 11 5 Experiment 12 5.1 Offlineexperiment ............................. 12 5.1.1 Datasets............................... 12 5.1.2 Metric................................ 13 5.1.3 TrainingSetting........................... 14 5.1.4 Result................................ 15 5.2 Onlineexperiment ............................. 17 5.2.1 Metric................................ 18 5.2.2 Trainingsetting........................... 18 5.2.3 Result................................ 19 6 Discussion 6.1 kfold .................................... 20 6.2 recursiverelabel............................... 20 7 Conclusion 22 Bibliography 23 List of Figures 1.1 Data collection through randomized controlled trial. . . . . . . . . . . . . 2 3.1 Label distribution of control and experimental group in Class Transfor mationmethod. ............................... 8 4.1 Flowchartofupliftmodelingwithrelabelprocess. . . . . . . . . . . . . . 9 4.2 Diagramofrelabelprocess.......................... 10 4.3 Label distribution of control and experimental group in relabeled Class Transformationmethod............................ 11 5.1 Derivation of uplift curve from the response curves of control group and experimentalgroup.............................. 15 5.2 OfflineexperimentresultonCriteodataset. . . . . . . . . . . . . . . . . 16 5.3 OfflineexperimentresultonLentadataset.. . . . . . . . . . . . . . . . . 16 5.4 OfflineexperimentresultonX5dataset. .................. 16 5.5 The process of deriving CIR, metric of online performance. . . . . . . . . 18 6.1 Comparisonofrelabelprocesswithdifferentk. . . . . . . . . . . . . . . 21 6.2 Comparisonofusingrecursiverelabelingornot.. . . . . . . . . . . . . . 21 List of Tables 5.1 DistributionofinstancesinCriteoDataset . . . . . . . . . . . . . . . . . 12 5.2 DistributionofinstancesinLentaDataset . . . . . . . . . . . . . . . . . 13 5.3 DistributionofinstancesinX5Dataset................... 13 5.4 DistributionofinstancesinAppierDataset. . . . . . . . . . . . . . . . . 14 5.5 OfflineexperimentresultsonAppierdatasets. . . . . . . . . . . . . . . . 17 5.6 Online experiment results on Appier uplift modeling service. . . . . . . . 19" | |
| dc.language.iso | en | |
| dc.title | 利用重標籤方法於增益模型中進行直接且不偏之優化 | zh_TW |
| dc.title | Relabel Process for Direct and Unbiased Optimization in Uplift Modeling | en |
| dc.date.schoolyear | 110-1 | |
| dc.description.degree | 碩士 | |
| dc.contributor.advisor-orcid | 林守德(0000-0001-9970-1250) | |
| dc.contributor.oralexamcommittee | 林軒田(Chin-Chin Tseng),孫民(Yu-Chin Li) | |
| dc.subject.keyword | 增益模型,條件處理效應,過擬合,重標籤,弱監督式機器學習, | zh_TW |
| dc.subject.keyword | uplift modeling,conditional treatment effect,overfitting,relabel process,weakly-supervised machine learning, | en |
| dc.relation.page | 24 | |
| dc.identifier.doi | 10.6342/NTU202200072 | |
| dc.rights.note | 同意授權(全球公開) | |
| dc.date.accepted | 2022-01-20 | |
| dc.contributor.author-college | 電機資訊學院 | zh_TW |
| dc.contributor.author-dept | 資訊工程學研究所 | zh_TW |
| 顯示於系所單位： | 資訊工程學系 | |
文件中的檔案：
| 檔案 | 大小 | 格式 | |
|---|---|---|---|
| U0001-1601202215552400.pdf | 2.76 MB | Adobe PDF | 檢視/開啟 |
系統中的文件，除了特別指名其著作權條款之外，均受到著作權保護，並且保留所有的權利。