Font Size: a A A

Analysis Of User Characteristics Based On A Run Application

Posted on:2019-11-12Degree:MasterType:Thesis
Country:ChinaCandidate:R D YanFull Text:PDF
GTID:2417330566475736Subject:Applied Statistics
Abstract/Summary:PDF Full Text Request
With the rising of the Internet,Internet of things and artificial intelligence,running applications and various smart devices spring up,the market is getting mature and users are gradually increasing.More and more people like to use running applications while running.And after running,they like to dry up their own trajectory in a circle of friends.Therefore,These behaviors have accumulated a lot of data,so it is a very significant thing to study the characteristics and behaviors of these users who use running applications.This paper selects an application from the market,according to its database data,some data indicators will be selected,its user characteristics and user behavior will also be analyzed.Firstly,the users are analyzed basing on the indicators,and the running users are classified by clustering to obtain different levels of users.Then,according to the characteristics of different levels of users,which factors have significant impact on different levels of users will be analyzed by using ordinal regression.According to research,geographical distribution shows that the largest number of users is in Guangdong Province,followed by Jiangsu;from an age point of view,users are mainly middle-aged people over 40 years old,but young users are gradually increasing.After classifying users,it was found that most of the users belonged to general users,and all of them were occasionally using a running app.The users with commercial value were ashes,loyal users,and important development users.Its main distribution area is East China Area.The number of days of registration for ashes users,loyal users,and important development users is mostly over 150 days,but there are still a small number of users whose registration days are within 100 days.Further study of the relationship between indicators and different categories of users,found that registration days,weight and gender have a significant impact on different categories of users.
Keywords/Search Tags:Running applications, Clustering, Ordinal regression
PDF Full Text Request
Related items