Brand evaluation system for vehicles based on multi-source heterogeneous data and CatBoost model
Huiying Hu, Jing Yang, Fan Zhang, Zhao Yishu · 2024
In order to comprehensively evaluate the brand equity of different new energy vehicles and further promote the development of the new energy vehicle industry, this paper introduces a new energy vehicle brand equity evaluation system based on multi-source heterogeneous data and CatBoost model. Firstly, this paper constructs a two-layer multi-source heterogeneous data index system based on market performance parameter indicators, consumer sentiment indicators and product parameter indicators. Secondly, according to the types and formats of different index data, different mapping methods were used for data preprocessing, and the subjective and objective comprehensive variation weighting methods were used to calculate the mixed weights. Then, based on the advantages and characteristics of the CatBoost model, the input data was iteratively cumulatively calculated. Finally, this paper analyzes the case studies of new energy vehicles based on joint venture brands and independent brands. The brand equity of China's own brand of new energy vehicles is 1.9 points higher than that of joint venture brands, mainly due to the overtaking of China's new energy vehicles. It can be seen that this paper provides an efficient evaluation system for the real-time calculation of brand equity of new energy vehicles in the new energy vehicle industry, and the emergence of this evaluation system makes up for the gap in the calculation of brand equity in the new energy vehicle industry to a certain extent.