问题
I've followed the advice laid out here for calculating the average of circular data:
https://en.wikipedia.org/wiki/Mean_of_circular_quantities
But I'd also like to calculate standard deviation as well.
#A vector of directional data (separated by 20 degrees each)
Dir2<-c(350,20,40)
#Degrees to Radians
D2R<-0.0174532925
#Radians to Degrees
Rad2<-Dir2 * D2R
Sin2<-sin(Rad2)
SinAvg<-mean(Sin2)
Cos2<-cos(Rad2)
CosAvg<-mean(Cos2)
RADAVG<-atan2(SinAvg, CosAvg)
DirAvg<-RADAVG * R2D
The above gives me the average, but I don't know how to calculate the SD
I tried to just take the mean of the standard deviation for both the sine and cos, but I get varying answers.
SinSD<-sd(Sin2)
CosSD<-sd(Cos2)
mean(CosSD, SinSD)
回答1:
You may use the circular
package for that:
x <- circular(Rad2)
mean(x)
# Circular Data:
# Type = angles
# Units = radians
# Template = none
# Modulo = asis
# Zero = 0
# Rotation = counter
# [1] 0.2928188 # The same as yours
sd(x)
# [1] 0.3615802
Manually,
sqrt(-2 * log(sqrt(sum(Sin2)^2 + sum(Cos2)^2) / length(Rad2)))
# [1] 0.3615802
which can be seen from the source code of sd.circular
.
See also here and here.
来源:https://stackoverflow.com/questions/55616697/how-to-calculate-standard-deviation-of-circular-data