CTR Optimisation for CPC Ad Campaigns Using Hybrid Recommendation System
Loading...
URL
Journal Title
Journal ISSN
Volume Title
Perustieteiden korkeakoulu |
Master's thesis
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.
Authors
Date
2022-06-13
Department
Major/Subject
Data Science
Mcode
SCI3115
Degree programme
Master's Programme in ICT Innovation
Language
en
Pages
69
Series
Abstract
Online advertisers deal with a large amount of historical data, consisting of user interactions with certain ads. Since there is close to none content information, the collaborative filtering approach is applied, which makes recommendations based on similarities between users or items using only historical preference data. Due to this, the most significant problem that these methods have to overcome is the cold-start problem. The thesis covers the applications of recommender systems in the online advertisement domain and investigates the advantages and disadvantages of neural network-based collaborative filtering. Part of the research entails finding ways to extract meaningful data from the users and advertisements and extend the already existing company model with the new content information. The proposed Hybrid method is evaluated in order to measure, if it results in better performance when facing with the cold-start scenario. In order to provide a relative comparison and a way to replicate the achieved results, we test the models on a publicly available dataset. The main experiments, however, are conducted on the company data with additional extracted information. In summary, we investigate the application of neural networks as collaborative filtering systems. Furthermore, we introduce a possible way for data extraction and processing for online advertisements and extend the original collaborative filtering network with the new feature information to create a Hybrid Neural Network-based model.Description
Supervisor
Babbar, RohitThesis advisor
Chernov, SergeyKeywords
collaborative filtering, neural networks, hybrid approach, recommender system, online advertisement