Jump to content


Photo

FLightR outlier detection

FLightR geolocation

  • Please log in to reply
1 reply to this topic

#1 L Halpin

L Halpin
  • Society Members
  • 10 posts
  • Vancouver, Canada,
  • Canada

Posted 14 July 2017 - 07:46 PM

Hi Folks,

 

 

Can anyone describe how outlier detection works in FLightR?  How does it determine which points are outliers and what are the criteria for outlier detection?

 

check.outliers=TRUE is part of the run.particle.filter function.

 

 

Thanks,

Luke



#2 Eldar

Eldar
  • General Members
  • 47 posts
  • Netherlands,
  • Netherlands

Posted 17 July 2017 - 02:06 AM

  Hi Luke,

 I do not think it is well documented, so I will try to explain it here.

Imagine we are at time point (twilight) A, particle filter simulates new positions for point B and from those for point C. After it checks 

1. If AB average movement was more than 50 km

AND (

2. AB average distance was > AC.distance*1.3

OR

3. AB.distance>AC.distance AND angle ABC<100 degrees

 

If this condition is satisfied it exclude point B.

   I know it is not very smart filtering, so I am not a big fan of it. I will be happy if someone programmes something better, e.g. Douglas filter







Also tagged with one or more of these keywords: FLightR, geolocation

0 user(s) are reading this topic

0 members(s), 0 guests(s) and 0 anonymous member(s)