The Student Room Group

meaning of binomial?

An email spam filter correctly classifies mails (i.e. a spam mail as spam, a not-spam as not-spam) 2/3 of the times. Classification decisions are independent of each other. You test the classifier with a set of 3 emails. The classifier will stop when it performs one wrong classification (e.g. when it says that a spam email is not spam, or vice-versa), or when it correctly classifies all 3 emails. Answer the following:
a) Is this a binomial experiment? Explain your answer.
b) What is the probability that the classifier will make at most 1 correct classification?
c) What is the standard deviation in the number of correct classifications?
What have u tried?
Meanwhile, the answer to question 1 is already started in the question !
Original post by Faith&hardwork
What have u tried?
Meanwhile, the answer to question 1 is already started in the question !

according to the first sentence it would be binomial however im not sure there is a fixed number of trials since it mentions that if one email is wrongly classified then the rest are ignored and not carried on with. for question b, i would work out probability of 0 times and 1 time and just add them up and question c, use the standard deviation formular sqr of np(1-p) where n is trials and p is possibility of success
(edited 2 years ago)
Original post by Carrying a torch
according to the first sentence it would be binomial however im not sure there is a fixed number of trials since it mentions that if one email is wrongly classified then the rest are ignored and not carried on with. for question b, i would work out probability of 0 times and 1 time and just add them up and question c, use the standard deviation formular sqr of np(1-p) where n is trials and p is the possibility of success

I think you run the test with 3 sets of email
For a Binomial distribution, there must be a fixed probability of success, once the classifier is faulty it stops so as not to alter the probability
Original post by Anonymous
An email spam filter correctly classifies mails (i.e. a spam mail as spam, a not-spam as not-spam) 2/3 of the times. Classification decisions are independent of each other. You test the classifier with a set of 3 emails. The classifier will stop when it performs one wrong classification (e.g. when it says that a spam email is not spam, or vice-versa), or when it correctly classifies all 3 emails. Answer the following:
a) Is this a binomial experiment? Explain your answer.
b) What is the probability that the classifier will make at most 1 correct classification?
c) What is the standard deviation in the number of correct classifications?


You've got some reponses - not all necessarily correct. What are your thoughts? And what have you been able to do so far?

Quick Reply

Latest