A CTR Prediction Approach for Text Advertising Based on the SAE-LR Deep Neural Network

다국어 초록 (Multilingual Abstract)

For the autoencoder (AE) implemented as a construction component, this paper uses the method of greedylayer-by-layer pre-training without supervision to construct the stacked autoencoder (SAE) to extract theabstract features of the original input data, which is regarded as the input of the logistic regression (LR)model, after which the click-through rate (CTR) of the user to the advertisement under the contextualenvironment can be obtained. These experiments show that, compared with the usual logistic regressionmodel and support vector regression model used in the field of predicting the advertising CTR in theindustry, the SAE-LR model has a relatively large promotion in the AUC value. Based on the improvement ofaccuracy of advertising CTR prediction, the enterprises can accurately understand and have cognition for theneeds of their customers, which promotes the multi-path development with high efficiency and low costunder the condition of internet finance.

참고문헌 (Reference)

1 "scikit-learn 0.15.2"

2 "Word2vec"

3 G. Santamaria-Bonfil, "Wind speed forecasting for wind farms : a method based on support vector regression" 85 : 790-809, 2016

4 T. Graepel, "Web-scale Bayesian click-through rate prediction for sponsored search advertising in Microsofts Bing search engine" 13-20, 2010

5 "Theano"

6 T. Akimova, "Standardization, evaluation, and areaunder-curve analysis of human and murine Treg suppressive function" 1371 : 43-78, 2016

7 P. Vincent, "Stacked denoising autoencoders : learning useful representations in a deep network with a local denoising criterion" 11 : 3371-3408, 2010

8 K. P. Lin, "Solar power output forecasting using evolutionary seasonal decomposition leastsquare support vector regression" 134 (134): 456-462, 2016

9 R. Trivedi, "Simultaneous prediction of blast-induced flyrock and fragmentation in opencast limestone mines using back propagation neural network" 7 (7): 237-252, 2016

10 "Setting Up Scientific Python"

11 Z. Jiang, "Research on CTR prediction for contextual advertising based on deep architecture model" 18 (18): 11-19, 2016

12 P. Rebentrost, "Quantum support vector machine for big data classification" 113 : 2014

13 "Python sklearn.svm.SVR Examples"

14 K. Dave, "Predicting the click-through rate for rare/new Ads" Centre for Search and Information Extraction Lab International Institute of Information Technology 2010

15 H. Cheng, "Personalized click prediction in sponsored search" 351-360, 2010

16 H. Cheng, "Multimedia features for click prediction of new ads in display advertising" 777-785, 2012

17 O. Chapelle, "Modeling delayed feedback in display advertising" 1097-1105, 2014

18 K. J. Grimm, "Model selection in finite mixture models : a k-fold cross-validation approach" 24 (24): 246-252, 2017

19 "Logistic Regression in Python"

20 Y. Bengio, "Learning deep architectures for AI" 2 (2): 1-127, 2009

21 H. I. Suk, "Latent feature representation with stacked auto-encoder for AD/MCI diagnosis" 220 (220): 841-859, 2015

22 "KDD Cup 2012, Track 2"

23 A. Jimenez-Valverde, "Insights into the area under the receiver operating characteristic curve(AUC)as a discrimination measure in species distribution modelling" 21 (21): 498-507, 2012

24 Y. Hu, "Incentive problems in performance-based online advertising pricing : cost per click vs. cost per action" 62 (62): 2022-2038, 2015

25 K. Weinberger, "Feature hashing for large scale multitask learning" 1113-1120, 2009

26 J. Gehring, "Extracting deep bottleneck features using stacked autoencoders" 3377-3381, 2013

27 Taher Osman, "Driving factors of urban sprawl in Giza governorate of the Greater Cairo Metropolitan Region using a logistic regression model" 도시과학연구원 20 (20): 206-225, 2016

28 A. J. Talabani, "Clinical diagnostic accuracy of acute colonic diverticulitis in patients admitted with acute abdominal pain, a receiver operating characteristic curve analysis" 32 (32): 41-47, 2017

29 Y. Tagami, "CTR prediction for contextual advertising : learning-to-rank approach" 2013

30 X. Deng, "An improved method to construct basic probability assignment based on the confusion matrix for classification problem" 340-341 : 250-261, 2016

31 Y. Bengio, "Advances in Neural Information Processing Systems" MIT Press 153-160, 2006

32 H. B. McMahan, "Ad click prediction : a view from the trenches" 1222-1230, 2013

33 Z. Jiang, "A feature vector representation approach for short text based on RNNLM and pooling computation" 15 (15): 6-14, 2017

34 G. E. Hinton, "A fast learning algorithm for deep belief nets" 18 (18): 1527-1554, 2006