A car recommendation method and device

Abstract

Embodiments of the invention disclose a car recommendation method and apparatus. The method comprises the steps of obtaining dimensional data of car models in car vertical websites, wherein the dimensional data comprises dimension evaluation scores; preprocessing the dimensional data to enable the evaluation scores to be distributed in a preset interval range; clustering the preprocessed dimensional data to obtain clustering clusters of different car models, wherein a clustering cluster number is determined according to dimensional information in the obtained dimensional data of the car models; obtaining dimensional information specified by a user and according to the dimensional information specified by the user, screening the obtained clustering clusters; and sorting all the car models in the clustering clusters in the selected clustering clusters, and displaying a sorting result to the user. Convenient and efficient car product recommendation can be provided for the user.

Description

Technical field [0001] The embodiment of the present invention relates to the technical field of data processing, in particular to a method and device for recommending a car. Background technique [0002] At present, there are two main traditional methods for recommending the purchase of cars. The first is the recommendation method based on professionals. It collects user needs, such as price, brand, model and other information, and then becomes familiar with the car. Market professionals give several candidates based on their personal experience. The second method is for users to learn by themselves. The users of car buyers will compare and observe on the websites of various car verticals and different car 4s stores according to their own needs, and then choose the models they think are reasonable. [0003] Since Volkswagen did not know much about automobiles in the past few years, the first recommendation method based on professionals became the mainstream method for obtaining t...

AI Technical Summary

Problems solved by technology

Another problem is that due to the characteristics of personal preferences, different professionals will lead to the same user needs, and the candidate models given by different professionals are different, which cannot give users efficient recommendations.

[0004] For the second method of learning by users, although with the rapid development of the information age, the threshold for users to learn new domain knowledge has become lower and lower, but there are still certain technical barriers between different domain knowledge. The field of science and engineering is even worse. Because the automotive field involves many fields, which contains a lot of professional and technical knowledge, the learning cost of ordinary people without background is relatively high

Secondly, due to the energy and time constraints of each user, the car models that can be learned, the materials that can be learned, and the data that can be observed in 4S stores are all limited, so there will be some incomplete data samples, resulting in some high-quality The model is not shortlisted, that is, the selection made on the limited sample data sometimes ignores the high-quality models outside the sample

Images