Article Preview
TopIntroduction
With the vigorous development of the internet and mobile technology, digital advertising is experiencing explosive growth in both scale and complexity (Desai, 2019). No longer confined to simple information delivery, modern digital advertising pursues higher levels of personalization, creativity, and intelligence to better cater to increasingly diverse and discerning user demands. In this wave of digitization, the prediction and analysis of advertising effectiveness have become particularly crucial.
Digital marketing, as a means of brand promotion, market expansion, and customer relationship management through digital technology and media, has become an indispensable component of corporate competitiveness. It boasts advantages such as low cost, high efficiency, quantifiability, and interactivity (Veleva & Tsvetanova, 2020) and is widely applied across various platforms such as the internet, mobile devices, and social media. As digital technology continues to innovate, digital marketing exhibits trends toward diversification, intelligence, and personalization, necessitating more flexible and precise advertising strategies.
In the complex ecosystem of digital marketing, predicting and analyzing advertising effectiveness is a critically important task. This task involves effectively leveraging user data, ad content, and platform features to accurately predict and assess key metrics of ads, such as click-through rates (CTRs), conversion rates, and revenue (Richardson et al., 2007). This not only provides crucial foundations for businesses to formulate advertising strategies and optimize decisions, but also empowers them to better understand and adapt to the ever-changing market environment.
In the core challenges of predicting and analyzing advertising effectiveness, the difficulty lies in delving into the rich information embedded in ad creatives, encompassing various forms such as text, images, and videos. Simultaneously, there is a need to comprehend the relationships between this information and complex factors such as user behavior, platform environment, and market competition. Advancements in this research domain will offer new insights and solutions for the future evolution of digital advertising.
In past studies, scholars have made significant progress in the field of predicting and analyzing advertising effectiveness. They have primarily focused their research efforts on the following areas:
Data Mining Applications
In the early stages of predicting and analyzing advertising effectiveness, researchers actively explored the application of data mining techniques, particularly methods such as association rules and clustering analysis (Duran & Odell, 2013). Through these approaches, early researchers attempted to uncover latent relationships and user behavior patterns in advertising effectiveness data. The use of association rules enabled them to discover implicit associations in advertising placement data, thus identifying mutual influences among advertising elements (Lipianina-Honcharenko et al., 2022). For instance, a study proposed a two-stage ensemble algorithm based on cluster quality (Yan Chen, 2023). This algorithm first partitioned the dataset into several subsets using different clustering algorithms and then integrated the results of the subsets using ensemble learning methods, achieving a more optimal classification effect. This proposal provides valuable insights for our analysis of experimental results using ensemble learning methods.