Click-Through Rate Prediction of Online Banners Featuring Multimodal Analysis

Abstract

As the online advertisement industry continues to grow, it is predicted that online advertisement will account for about 45% of global advertisement spending by 2020.a Thus, predicting the click-through rates (CTRs) of advertisements is increasingly crucial for the advertisement industry. Many studies have already addressed the CTR prediction. However, most studies tried to solve the problem using only metadata such as user id, URL of the landing page, business category, device type, etc., and did not include multimedia contents such as images or texts. Using these multimedia features with deep learning techniques, we propose a method to effectively predict CTRs for online banners, a popular form of online advertisements. We show that multimedia features of advertisements are useful for the task at hand. In our previous work [ 1 ], we proposed a CTR prediction model, which outperformed the state-of-the-art method that uses the three features mentioned above, and also we introduced an attention network for visualizing how much each feature affected the prediction result. In this work, we introduce another text analysis technique and more detailed metadata. As a result, we have achieved much better performance as compared to our previous work. Besides, for better analyzing of our model, we introduce another visualization technique to show regions in an image that make its CTR better or worse. Our prediction model gives us useful suggestions for improving design of advertisements to acquire higher CTRs.

Attractiveness Computing in Image Media

“…For the analysis of online advertising, we are conducting joint research with SEPTENI CO., LTD., one of the major agencies of online advertisement in Japan. By our own deep learning architecture, we have succeeded in predicting CTR values in advance with a correlation coefficient of 0.82 [27], [28] for still image advertisements and 0.70 [29] for video advertisements. By predicting advertising effectiveness with high accuracy in advance, it is possible to arrange advertisements efficiently and effectively without conducting AB tests, etc.…”

A Deep Learning-Based Decision Support System for Mobile Performance Marketing

“…In online advertising, there are two main prediction tasks: Click Through Rate (CTR), 11,8 predicting if an ad link is clicked when a user views a webpage or app; and CVR, 7 estimating if there will be a sale when an ad is viewed. Under the analyzed MPM domain, the CTR probability is always 100% since users need to click an ad before accessing the publisher content (e.g., news portal).…”

Predicting Online Video Advertising Effects with Multimodal Deep Learning

“…Although many studies have worked on the effects of image advertising [1]- [3], there are seldom research on video advertising. In particular, to our knowledge, research on predicting the CTR for online video ads has not been conducted extensively.…”

“…These methods dealt primarily with metadata. Chen et al [10] showed the effectiveness of using features extracted from images by convolutional neural networks (CNNs) for prediction; Iwazaki [1] proposed a method to use features from text extracted by CNNs in addition to images and metadata; Xia et al [2], [3] showed that extracting features from embedded text improved prediction accuracy. In addition, they visualized the contribution of images, metadata, Fig.…”