Sept 2020 Paper 2

Discussion in 'CP2' started by bernvall, Apr 6, 2022.

  1. bernvall

    bernvall Member

    This refers to the chi-square test given in the excel sheet for the exponential distribution, in sheet "Data"

    For the Observed, since we do not know the number of claims under the excess, the colleague assumed to make it equal to the expected (532) which is in red font.

    Now, part of the formula to get to the expected value of 532 is the total number of observed (which includes the ones below the excess of $50) where this total stands at 1351. However, this 1351 is only able to be fully calculated if you already have the full frequency, and hence the ones below the excess.
    Therefore it seems there is a sort of circular error in trying to get to the answer.
     
  2. Harashima Senju

    Harashima Senju Ton up Member

    The student that carried out the original model noted the following in his/her audit trail.

    Note the entries for the first two rows are different. For the 0 to 50 range, the frequency has been assumed to be in line with the empirical frequency. For the 50 to 100 range, this is set equal to the cumulative frequency for this range. For all other ranges, the frequency is calculated as the cumulative frequency for this range less the cumulative frequency for the previous range.

    I wouldn't worry about why they did it this way
     

Share This Page