Assignment 6 Part 1e and 1f

Assignment 6 Part 1e and 1f

by Ujas Patel -
Number of replies: 8

Hello, 

I would like to clarify a few points in relation to the sub-parts of section 1 of the assignment. They are as follows:

1) For 1e, avg.many.n.birthdays, is there a specific format for the output of the function? Currently, my function returns a vector of average probabilities and has a length of n -- is this okay or should it be different?

2) For 1f, run.chisq.test, should Feb. 29 be excluded from the data file before a sample of n birthdays are generated? I.e., we are to only include 365 dates -- is that correct?

3) For 1f, once again, my function first generates a sample of n birthdays and then I convert it to a frequency table to feed into chisq.test(). I see a chi-square output, however, I am not sure if the frequency table is the correct input for this test. Am I on the right track?

I greatly appreciate all clarifications, thank you!

In reply to Ujas Patel

Re: Assignment 6 Part 1e and 1f

by Erik Spence -
1) Yes, it should return a vector of averages.

2) Yes.

3) You need some sort of frequency calculation. However, the calculation must include those values that have zero counts.
In reply to Erik Spence

Re: Assignment 6 Part 1e and 1f

by Iciar Iturmendi Sabater -
As a follow up of q 3)(1f) - I am struggling to figure out how to calculate frequencies including the values that have zero counts. I have used the factor and summary functions to come up with frequency counts for both the randomly sampled and bdata sampled birthday vectors, but these don't include zero counts. I have thought of creating a 3rd vector (days<-1:365) to count frequencies of birthdays based on days, but can't figure out how to incorporate it to the factor and summary functions. I was wondering if you could advice whether I am on a right path to solve this problem, or if you could give some alternative advice on how to count frequencies accounting for values from 1:365 with zero counts.

Thank you for your help
In reply to Iciar Iturmendi Sabater

Re: Assignment 6 Part 1e and 1f

by Erik Spence -
I just used a vector and a loop to build the input data. Probably using a table first would be more efficient.
In reply to Erik Spence

Re: Assignment 6 Part 1e and 1f

by Erik Spence -
Confirmed, as other students have noted, it's more efficient to use table without using a loop.
In reply to Erik Spence

Re: Assignment 6 Part 1e and 1f

by Ujas Patel -
Hi Erik,

Agreed, using a table without using a loop would be more efficient as chisq.test can directly take this as an input. However, I am having trouble creating a table which includes both zero and non-zero counts for birthdays. I am able to create a data frame but not a table. Can you hint at how I can achieve this?

Thank you!
In reply to Ujas Patel

Re: Assignment 6 Part 1e and 1f

by Erik Spence -
No, I didn't originally figure it out myself, so I'll leave this question unanswered. Any technique you can figure out to get the correct input, including loops, will be accepted.
In reply to Erik Spence

Re: Assignment 6 Part 1e and 1f

by Edgardo Torres -
Just to clarify, when we say zero count, we mean that if the length of the frequency table is <365, we should add "0"s to the frequency table, up-to 365, correct?
In reply to Edgardo Torres

Re: Assignment 6 Part 1e and 1f

by Erik Spence -
Yes, sort of. But you need to add them to the correct birthdays that have zero counts.