Hi,
I don't like asking for help .. can someone check my work please ?
In a biometric system under testing there are 500 users. 200 were legitimate users (authorised), and 300 were not ( unauthorised)
At one threshold the system:-
Shut out 10 individual users it should have admitted, and admitted 2 individual it should have shut out.
What is the false rate match (FMR) at this point ?
is it 2/300 = 0.67% ???
Next the false non match rate
is it 10 / 200 = 5 % ???
If the threshold of the system is now decreased significantly which is the most likely outcome ..
The % of false matches will increase and the % of false non matches will decrease ???
Thank you .. for your kind help and time - I do appreciate it
Sure, I can help you check your work. Let's go through each question and calculate the answers step by step.
First, let's determine the False Match Rate (FMR) at the given threshold. The FMR represents the percentage of unauthorised users who are incorrectly matched or identified as legitimate.
To calculate the FMR, we divide the number of false matches (admitting unauthorized users) by the total number of unauthorized users:
FMR = Number of False Matches / Number of Unauthorized Users
In this case, you mentioned that 2 individual users were admitted when they should have been shut out. Therefore, the number of false matches is 2. The total number of unauthorized users is 300.
FMR = 2 / 300 ≈ 0.67%
So, your calculation for the FMR is correct.
Moving on to the next question about the False Non-Match Rate (FNMR). The FNMR represents the percentage of legitimate users who are incorrectly not matched or identified as unauthorized.
To calculate the FNMR, we divide the number of false non-matches (shutting out legitimate users) by the total number of authorized users:
FNMR = Number of False Non-Matches / Number of Authorized Users
According to your information, 10 individual users were shut out when they should have been admitted. The total number of authorized users is 200.
FNMR = 10 / 200 = 0.05 = 5%
So your calculation for the FNMR is correct as well.
Now, let's address the last part of your question. If the threshold of the system is decreased significantly, the most likely outcome is that the percentage of false matches will increase, and the percentage of false non-matches will decrease. This means that the likelihood of admitting unauthorized users (false matches) will be higher, while the likelihood of rejecting legitimate users (false non-matches) will be lower.
I hope this helps you verify your work. If you have any further questions, feel free to ask!