Comparative Analysis of Ad Click Behavior Prediction Using GAN-Augmented Data and Traditional Machine Learning Techniques

Yıl 2024, ERKEN GÖRÜNÜM, 1 - 1

Amel Sulaiman Salıhı , Oktay Yıldız

https://doi.org/10.2339/politeknik.1525138

Öz

In e-commerce, predicting click-through rates (CTR) is crucial to anticipate user behavior. User historical data can be used to extract interests and enhance CTR prediction, leading to higher accuracy. In this study, a Generative Adversarial Network (GAN) has been used to tackle the issue of insufficient dataset for click-through rates. Furthermore, six different machine learning algorithms have been assessed in predicting ad click behavior. For the experimental study, we obtained user demographic and online activity data from Kaggle, along with a binary label indi-cating ad clicks. To enhance the model's performance, we employed a GAN for data augmenta-tion and generated additional training examples. We compared the machine-learning algorithm's outcomes with and without GAN-based data augmentation to evaluate its predicted accuracy. According to the findings, most algorithms have increased sensitivity and specificity after utilis-ing GAN to augment the data, indicating that the generated data has improved their ability to accurately distinguish positive and negative events. GAN-based data augmentation boosted all models to varying degrees, according to the findings.

Anahtar Kelimeler

Click-Through Rate (CTR), Generative Adversarial Network (GAN), Data Augmentation, Machine Learning.

GAN-Artırılmış Veri ve Geleneksel Makine Öğren-imi Teknikleri Kullanılarak Reklam Tıklama Dav-ranışı Tahmininin Karşılaştırmalı Analizi

Öz

E-ticarette, kullanıcı davranışını öngörmek için tıklama oranlarının (TO) tahmin edilmesi önemlidir. Yüksek doğruluklu ilgi alanlarının çıkarılması ve TO tahmini için kullanıcıların geçmiş verileri kullanılabilir. Bu çalışmada, yetersiz ya da dengesiz veri kümelerinde reklam tıklama davranışının tahmini için Üretken Çekişmeli Ağlar (ÜÇA) kullanılmıştır. Çalışmada altı farklı makine öğrenmesi algoritmasının reklam tıklama davranışını tahmin etmedeki etkinliği değerlendirilmiştir. Gerçekleştirilen deneysel çalışmada, Kaggle'dan elde edilen kullanıcı demografik ve çevrimiçi aktivite verileri ve reklam tıklama etiketini gösteren bir veri kümesi kullanılmıştır. Modelin performansını artırmak amacıyla veri artırma yapılmış, bunun için ÜÇA kullanılmıştır. Tahmin doğruluğunu değerlendirmek için makine öğrenimi algoritmalarının ÜÇA temelli veri artırma ve veri artırma olmaksızın elde edilen sonuçlar karşılaştırılmıştır. Elde edilen sonuçlarda, hassasiyet ve özgüllük değerlerinin artığı, oluşturulan verilerin modellerin olumlu ve olumsuz olayları doğru bir şekilde ayırt etme yeteneklerini geliştirdiği gösterilmiştir. Bulgulara göre GAN tabanlı veri artırma, tüm modelleri farklı derecede güçlendirmiştir.

Anahtar Kelimeler

Tıklama Oranı (TO), Üretken Çekişmeli Ağlar (ÜÇA), Veri Arttırma, Makine Öğrenimi.

Kaynakça

• [1] Liu-Thompkins Yuping, "A Decade of Online Advertising Research: What We Learned and What We Need to Know," Journal of Advertising, pp. 1-13, (2018).
• [2] Y., & Zhai, P. Yang, "Click-through rate prediction in online advertising: A literature review," Information Processing & Management, p. 59, (2022).
• [3] Hong, Ziang, Xiong, Jinjie, You, Xiaolin, Wu, Min, Xia Wenxing, "CPIN: Comprehensive present-interest network for CTR prediction," Expert Systems With Applications, (2021).
• [4] Zhao Xudong, Xu Xinying, Han Xiaoxia, Ren Jinchang, Li Xingbing, Xie Jun, "DRIN: Deep Recurrent Interaction Network for click-through," Information Sciences, (2022).
• [5] WeiKang He, Yu Zhu, Jianghu Zhu, Yunpeng Xiao, "A click-through rate model of e-commerce based on user interest and temporal behavior," Expert Systems With Applications, (2022).
• [6] Danqing Zhu, "Advertising Click-Through Rate Prediction Based on CNN-LSTM Neural Network," Computational Intelligence and Neuroscience, (2021).
• [7] Liqing Qiu, Cheng’ai Sun, Qingyu Yang, Caixia Jing, "ICE-DEN: A click-through rate prediction method based on interest contribution extraction of dynamic attention intensity," Knowledge-Based Systems, (2022).
• [8] Y., Wang, S., Huang, Y., Zhao, X., Zhao, W., Duan, Y., & Wang, X. Tang, "Retrieval-Based Factorization Machines for Human Click Behavior Prediction," Computational Intelligence and Neuroscience, (2022).
• [9] J., Ma, C., Zhong, C., Zhao, P., & Mu, X. Zhang, "Multi- scale and multi-channel neural network for click-through rate prediction," Neurocomputing, (2022).
• [10] Dhanani, Keyur Rana Jenish, "Logistic Regression with Stochastic Gradient Ascent to Estimate Click Through Rate," in Information and Communication Technology for Sustainable Development: Proceedings of ICT4SD, Singapore, p. 319,(2018).
• [11] Gharibshah, Xingquan Zhu,Arthur Hainline, Michael Conway Zhabiz, "Deep Learning for User Interest and Response Prediction in Online Display Advertising," Data Science and Engineering, (2020).
• [12] K., Huang, Q., Zhang, F. E., & Lu, J. Song, "Coarse-to- fine: A dual-view attention network for click-through rate prediction," Knowledge-Based Systems, p. 216, (2021).
• [13] K., Huang, Q., Zhang, F. E., Lu, J. Song, "Coarse-to- fine: A dual-view attention network for click-through rate prediction," Knowledge-Based Systems, (2021).
• [14] D., Wang, Z., Zhang, L., Zou, J., Li, Q., Chen, Y., Sheng, W. Zou, "Deep Field Relation Neural Network for click-through rate prediction," Information Sciences, pp. 128-139, (2021).
• [15] D., Xu, R., Xu, X., Xie, Y. Jiang, "Multi-view feature transfer for click-through rate prediction," Information Sciences, pp. 961-976, (2021).
• [16] M., Cai, S., Lai, Z., Qiu, L., Hu, Z., Ding, Y. Liu, "A joint learning model for click-through prediction in display advertising," Neurocomputing, pp. 206-219, (2021).
• [17] D., Hu, B., Chen, Q., Wang, X., Qi, Q., Wang, L., Liu, H. Li, "Attentive capsule network for click-through rate and conversion rate prediction in online advertising," Knowledge-Based Systems, p. 106522, (2021).
• [18] P., Yang, Y., Zhang, C. Zhai, "Causality-based CTR prediction using graph neural networks," Information Processing & Management, p. 103137, (2023).
• [19] Z., Wang, X., He, X., Huang, X., Chua, T. S. Tao, "HoAFM: A High-order Attentive Factorization Machine for CTR prediction," Information Processing and Management, p. 102076, (2020).
• [20] Y., Jiang, D., Wang, X., Xu, R. Xie, "Robust transfer integrated locally kernel embedding for click-through rate prediction," Information Sciences, pp. 190-203, (2019).
• [21] X., Liu, Q., Su, R., Tang, R., Liu, Z., He, X., Yang, J. Yang, "Click-through rate prediction using transfer learning with fine-tuned parameters," Information Sciences, pp. 188-200, (2022).
• [22] A., Shetty, S. D. Jose, "DistilledCTR: Accurate and scalable CTR prediction model through model distillation," Expert Systems with Applications, p. 116474, (2022).
• [23] Alhassan Mumuni and Fuseini Mumuni, "Data augmentation: A comprehensive survey of modern approaches," Array, (2022).
• [24] A., Mittal, M., & Battineni, G. Aggarwal, "Generative adversarial network: An overview of theory and applications.," International Journal of Information Management Data Insights, (2021).
• [25] S Xu et al., "Cardiovascular risk prediction method based on CFS subset evaluation and random forest classification framework," in 2nd international conference on big data analysis, pp. 28–32,(2017).
• [26] C., Csörgő, A., & Martínez-Muñoz, G. Bentéjac, "comparative analysis of gradient boosting algorithms," Artificial Intelligence Review, pp. 1937-1967, (2021).
• [27] K., Patel, H., Sanghvi, D., & Shah, M. Shah, "comparative analysis of logistic regression, random forest and KNN models for the text classification," Augmented Human Research, pp. 1-16, (2020).
• [28] J., & Rana, K. Dhanani, "Logistic Regression with Stochastic Gradient Ascent to Estimate Click Through Rate," Information and Communication Technology for Sustainable Development, pp. 319-326, (2018).
• [29] Tianqi Chen and Carlos Guestrin, "XGBoost: A Scalable Tree Boosting System," in KDD '16: The 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco California, USA, pp. 785–794,(2016).
• [30] C., Csörgő, A., Martínez-Muñoz, G. Bentéjac, "A comparative analysis of gradient boosting algorithms," Artificial Intelligence Review, pp. 1937-1967, (2021).
• [31] Charbuty B. and Abdulazeez A., "Classification Based on Decision Tree Algorithm for Machine Learning," Journal of Applied Science and Technology Trends, pp. 20-28, (2021).
• [32] I. D., Sun, Y., Wang, Z. Mienye, "Prediction performance of improved decision tree-based algorithms: a review," Procedia Manufacturing, pp. 698-703, (2019).
• [33] Shuangjie Li, Kaixiang Zhang, Qianru Chen, Shuqin Wang, and Shaoqiang Zhang, "Feature Selection for High Dimensional Data Using Weighted K-Nearest Neighbors and Genetic Algorithm," IEEE Access, pp. 139512 - 139528, (2020).
• [34] F., Araghinejad, S. Modaresi, "A comparative assessment of support vector machines, probabilistic neural networks, and K-nearest neighbor algorithms for water quality classification," Water resources management, pp. 4095-4111, (2014).
• [35] A. Internet advertising spending worldwide from 2007 to 2024, by format Guttmann. Statista.[Online].https://www.statista.com/statistics/276671/global-internet-advertising-expenditure-by-type/ (2021)