Analysis of the application of abstract symbols in advertising design based on cluster analysis method

Abstract

This paper constructs an abstract symbol integration system based on big data applied in advertising design. Firstly, the abstract symbols are regarded as high-dimensional data, and the abstract symbols are collected using the local differential privacy algorithm in the big data algorithm. Secondly, the collected data are clustered, which is the process of big data processing. Finally, data classification is achieved by calculating the information gain rate and information entropy using the C4.5 decision tree. To verify the effectiveness of the built platform, this paper uses the platform to filter the most appropriate abstract symbols, and audiences of different age groups rate the ads and derive the online and offline sales of the products involved in the ads before and after the application of the abstract symbols. The results show that the ads’ quality improved significantly after applying abstract symbols. The post-90s raised 43 points for the prominent theme of the ads. Meanwhile, the online sales of products increased by 1.08 million units, and offline store sales increased by 840,000 units after applying abstract symbols. This shows that applying abstract symbols in advertising design can attract consumers to buy products and promote consumption.

Keywords

• local differencing
• cluster analysis
• c4.5 decision tree
• information entropy
• 65d17