CODEWITHSUNDEEP
rstatisticsprobability
penguin
penguin

Reputation: 1377

How to create a function from a numeric vector which takes a value as input and outputs probability of the value being present in the vector?

I would like to create a function or object from a numeric vector that takes a numeric value as input and gives probability of presence of that value in the numeric vector. I would like to create an object like my_object as in code below

numeric_vector = iris$Sepal.Length
# Create a function or object from numeric vector
my_object = generate_function(iris$Sepal.Length)
prob_of_2 = my_object(2)
print(prob_of_2) # outputs probability of presence of 2 in numeric vector

I have tried using the stats::density function to create my_object kind of object. But the problem is that I cannot use a single value to output probability.

numeric_vector = iris$Sepal.Length
my_function = data.frame(value = density(numeric_vector)$x, probability = density(numeric_vector)$y)
prob_of_2 = my_function[my_function$value == 2, ]$probability 
print(prob_of_2) # outputs numeric(0)

I have to use a range, which I want to avoid

epsilon = 1.5
prob_of_2 = sum(my_function[my_function$value >= 2 - epsilon & my_function$value <= 2 + epsilon, ]$probability)
print(prob_of_2) # outputs 0.0007896299

I hope the question is clear and makes sense. Any help is really appreciated.

Upvotes: 0

Views: 119

Answers (1)

user12728748
user12728748

Reputation: 8506

You can use approxfun for that

approx <- function(x, vec=iris$Sepal.Length, ...){
    approxfun(density(vec, ...))(x)
}
approx(4.5)
#>[1] 0.1707666

approx(2, width=4)
#>[1] 0.002516293

see also: https://stats.stackexchange.com/questions/78711/how-to-find-estimate-probability-density-function-from-density-function-in-r

Upvotes: 1

Related Questions