Extract values within interquartile range in array

Hi,

I want to extract values within an interquartile range in an array. Let say that the array looks as follows: v= [1 2 3 4 5 6 ......95 96 97 98 99]. The interquartile range is 25 and 75. How can I extract the array values from 25 to 75 to a new array? If possible, can I select other ranges than bottom and top 25%?

Thanks!

Answer 1

[the cyclist]

Use the quantile command to get whichever quantiles you want.

Then extract using logical indexing (as described on this documentation page).

v = 1:99;
q = quantile(v,[0.25 0.75])
v2 = v(v>q(1) & v<q(2))

[Vlatko Milic]

Thank you, exactly what I was looking for! I just have one question about the follown: Let's say I have 10 000 digits in the cell and select 10% quantile. But my quantile value at 10% and 90 % will be between two integers. How will the function operate then? What values will be selected?

[the cyclist]

This logical value

v>q(1)

will be true when the value in your vector is strictly larger than the quantile (regardless of whether it is an integer or not). So if all your values in v are integers, it will only select the integers that are larger than the "in-between" quantile value.

[Vlatko Milic]

okej, good explanation. Thank you again! :)

[Vlatko Milic]

The cyclist, I tried with : v= 0:10; q = quantile(v,[0.25 0.75])

This gives med 2.25 and 7.75. As it is the quartiles I thought I would get 2.5 and 7.5. What do I do wrong?

[the cyclist]

You did not do anything wrong, but your expectation was wrong. :-)

Technically, any value between 2 and 3 will qualify as a 0.25 quantile of that vector. The definition is not strict enough to determine a unique value. There are many different procedures for selecting a single value. (See, e.g., this Wikipedia page or this MathWorld page.)

The procedure that MATLAB uses is described on this documentation page.

[Vlatko Milic]

I see, that sounds reasonable.Thank you for the pedagogic explanation :)

[Sesilia Iileka]

Hi i want kind of the same thing but from a loop

like this:

for i = length(x)

xy = find(u==i | u>i & u4<(i+1)); #xy should find indices of u within that range i.e u(i)>=i<u(i+1)

N(i) = length(xy) #giving me counts of i

end

but then, the last iteration (last value of i) will be wrong by this expression.

how do i correct this please?

Thanks a lot

Answer 2

[madhan ravi]

v=1:100;

new_array = v(v >=25 & v<=75) %values between 25 to 75

new_array1 = v(v<25 | v>75) % values less than 25 and greater than 75

[madhan ravi]

The extraction of ranges can be done by logical indexing.

[Vlatko Milic]

Yes, but let's say that I don't know what the interquartile range is before the statement. So I don't know the values. What do I need to do then?

[Vlatko Milic]

Thank you!

[madhan ravi]

Anytime :) couldn’t help you though because have no experience in statistics field :)

[Vlatko Milic]

No worries, thank you friend :)

[Sesilia Iileka]

Hi i want kind of the same thing but from a loop

like this:

for i = length(x)

xy = find(u==i | u>i & u4<(i+1)); #xy should find indices of u within that range i.e u(i)>=i<u(i+1)

N(i) = length(xy) #giving me counts of i

end

but then, the last iteration (last value of i) will be wrong by this expression.

how do i correct this please?

Thanks a lot