A Node Selection Paradigm for Crowdsourcing Service Based on Region Feature in Crowd Sensing

Algorithm 1

Density-based spatial clustering of applications with noise (DBSCAN).

Input: Trace dataset ,

Radius , Density Threshold

Output: Density-based trace dataset

1:mark all the nodes in unvisited

2:while not all the nodes are visited do

3:select an unvisited node from

4:mark node visited

5:define the -domain of as the area with radius centered on

6:if  -domain of has at least nodes then

7:create a new cluster and add into

8:make as a set of the -domain of

9:for each   do

10:if   is unvisited then

11:mark as visited

12:if  -domain of has at least nodes then

13:put the nodes into

14:end if

15:if   is not the member of any clusters then

16:put into

17:end if

18:end if

19:end for

20:output

21:else

22:mark as noise

23:end if

24:end while