Font Size: a A A

Data Mining Based On The 21st Chengdu International Automobile Exhibition

Posted on:2020-01-22Degree:MasterType:Thesis
Country:ChinaCandidate:C YangFull Text:PDF
GTID:2417330578953146Subject:Applied Statistics
Abstract/Summary:PDF Full Text Request
The 21st Chengdu International Automobile Exhibition was held from August 31 to September 9,2018.Chengdu International Automobile Exhibition ranks among the top four A-class auto shows in China.Based on the 21st Chengdu International Automobile Exhibition,data mining using machine learning technology and computer technology is particularly necessary and meaningful.This paper first builds an automatic identification model for automobile brands and cars with three kinds of machine learning methods which are random forest,multi-layer neural network and support vector machine.The automatic identification model provides the standard brand name and the standard car name according to the brand name and car name of the eight car websites,that are Pacific Auto,Sina Auto,Sohu Auto,Cheshi,Emao,Ifeng,Netease Auto and Tencent Auto and updates the training set in a step-by-step manner to get a standard database.This database contains 17972 pieces of information on seven indicators that are brand_name,brand_id,car_name,car_id,standard_brand,standard_car,and web.Furthermore,the paper builds the LDA model to infer the topic of the news report about Chengdu Auto Show from August 28 to September 12 from three angles,then gets six topics which are the detecting test of vehicle,performance of products,brand strategy,reputation and marketing,listing new car and product planning and the number of news articles that talk about each topic every day.This article has the following innovations:Firstly,the data used in the research is up-to-date,especially the news data used for topic inference is updated daily and time-sensitive.Secondly,a database of standard brand names and standard car names was obtained,which provided support for calculating the user's preferences of brand and car,as well as facilitating future research on other car shows.Thirdly,the results of the topic inference from the brand dimension found that there is a transfer tendency on the attention of the six topics in pre-show,exhibition and post-show,which will provide business opportunities for these brands to a certain extent.
Keywords/Search Tags:The 21st Chengdu International Automobile Exhibition, Machine Learning, brand, car, News report
PDF Full Text Request
Related items