Research on advertising recommendation method based on classification model

Research on advertising recommendation method based on classification model

Beijing Jiaotong University Zhe

The main work of the paper is as follows. First, we implemented a visual statistical and analytical tool for advertising log data provided by an Internet company, using the Hadoop platform to analyze the data and discover the dependencies between features and advertising. Secondly, an improved method of using the single tag classification model based on the non advertising feature and the advertising feature dependency is proposed, which utilizes the mutual information to select the combination feature to join the dependency relationship between the features. Thirdly, this paper proposes a new method to improve the multi-tag classification model, which uses a heuristic method to construct the classifier chain to make better use of the relationship between ads. Finally, the experiments on the Hadoop platform show that both of the improved methods can get better results than those before the improved method. Therefore, whether it is the relationship between the characteristics or advertising, is the advertising recommendation problem can not be neglected to affect the factors.

Research on advertising recommendation method based on classification model