Looping a Rayleigh test to find highest value

Question

I am a beginner with R, so hopefully this will be an easy fix.

I am trying to use a for loop on a dataset for neuron firing direction in order to:

Incrementally add the next value from the dataset to a vector
Run a Rayleigh test on that vector and save it to a variable
Test if the Rayleigh test I just ran has a larger statistic than the the Rayleigh test in the last loop just before it, as well as having a p-value of less than .05
If the value is larger, save the statistic value, so that the next loop can compare to it
If the value is larger, save the vector

So far I have this for the code, and after going through it for a long time I'm at a loss for why it's not working. Every time I run it, the for loop goes all the way to the end and just reports the rayleigh value and vector for the whole dataset, which I know for sure isn't correct. (I'm using the circular package for the rayleigh test function)

# This first line is just to create an initial rayleigh statistic to compare to in the loop that is low

best_rayleigh <- rayleigh.test(1:10)
data_vector <- c()

for (i in firing_directions) {
  data_vector <- append(data_vector, i) 
  ray_lee_test <- rayleigh.test(data_vector) 
  if ((ray_lee_test$statistic>best_rayleigh$statistic)&(ray_lee_test$p.value<=.05)) {
    best_rayleigh <- ray_lee_test 
    best_rayleigh_vector <- data_vector
  } else {
    NULL
  }
}

Any help is appreciated. Thank you!

Update: I tried using && instead of single & in the if statement, however it returned the same result

Try double `&&` in the `if` condition. See this [SO post](https://stackoverflow.com/questions/6558921/boolean-operators-and). — Rui Barradas, Jun 30 '20 at 21:08
Please try to give a minimal example, see [mcve], so we can try your code — denis, Jun 30 '20 at 21:45

StupidWolf · Answer 1 · 2020-06-30T23:01:58.727

I think the NULL is throwing up some issues. Not sure what will have if your throw a NULL. You only update the vector if it passes your criteria:

library(circular)
firing_directions= rvonmises(n=25, mu=circular(pi), kappa=2)

best_rayleigh <- rayleigh.test(1:10)
data_vector <- c()    

for (i in firing_directions){
  data_vector <- c(data_vector, i) 
  ray_lee_test <- rayleigh.test(data_vector) 
     if ((ray_lee_test$statistic>best_rayleigh$statistic)&(ray_lee_test$p.value<=.05)) {
        best_rayleigh <- ray_lee_test 
        best_rayleigh_vector <- data_vector
     } 
  }

score 0 · Accepted Answer · answered Jul 01 '20 at 06:43

The following code doesn't give warnings and selects the vector with highest test statistic and "significant" p-value.

library(circular)

set.seed(2020)
firing_directions <- rvonmises(n = 25, mu = circular(pi), kappa = 2)
plot(firing_directions)

best_rayleigh <- rayleigh.test(circular(1:10))
for(i in seq_along(firing_directions)){
  dv <- firing_directions[seq_len(i)]
  rltest <- rayleigh.test(dv)
  if((rltest$statistic > best_rayleigh$statistic) && (rltest$p.value <= 0.05)){
    best_rayleigh <- rltest 
    best_rayleigh_vector <- dv
  }
}

best_rayleigh
#
#       Rayleigh Test of Uniformity 
#       General Unimodal Alternative 
#
#Test Statistic:  0.8048 
#P-value:  0.0298 

best_rayleigh_vector
#Circular Data: 
#Type = angles 
#Units = radians 
#Template = none 
#Modulo = asis 
#Zero = 0 
#Rotation = counter 
#[1] 4.172219 2.510826 2.997495 4.095335 3.655613

Looping a Rayleigh test to find highest value

2 Answers2