Combination of 1D CNN and 2D CNN to Evaluate the Attractiveness of Display Image Advertisement and CTR Prediction
Autor: | Wee Lorn Jhinn, Hiang-Kwang Chua, Poo Kuan Hoong |
---|---|
Rok vydání: | 2019 |
Předmět: |
Attractiveness
Computer science Digital data Margin of error Advertising 02 engineering and technology Space (commercial competition) Image (mathematics) Metric (mathematics) 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Mass audience ComputingMilieux_MISCELLANEOUS Digital advertising |
Zdroj: | Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing ISBN: 9783030264277 |
Popis: | With the explosion of digital data nowadays, it has catapulted the usage of data analytic in the emergence of digital advertising space. One of the digital advertising giants, Facebook has accelerated the growth of this digital data volume as they are the most common used platforms for advertisers to advertise and deliver advertising messages to the mass audience online. However, this phenomenon has increased the challenges faced by advertisers to further attract audiences attentions to look at the digital advertisements when they are shown advertisements in Facebook platforms. Hence, in this paper, we proposed a method to evaluate and analyze the elements of attractiveness within the display advertisement in Facebook Advertisement platform by applying the 2D CNN on the display advertisement images while 1D CNN on the click metric data respectively. Based on our experiment results, we are able to predict the display CTR with a reasonable margin of error. |
Databáze: | OpenAIRE |
Externí odkaz: |