Probability Question

[anthonyporter]

Here is my problem:
I have data about 1000 people in database A
The data on each person can be categorized into 5 groups: 1) demographic; 2) health; 3) dental; 4) financial; 5) loan
I am converting all data in database A to database B
After each conversion test run I assess each category of data for how accurately it was converted to database b:
Results of 1st conversion test run:
80% of demographic data in database A matches the demographic data database B
75% of health data in database A matches the demographic data database B
65% of dental data in database A matches the demographic data database B
90% of financial data in database A matches the demographic data database B
85% of loan data in database A matches the demographic data database B

From the results of the 1st test run, how can I determine the probability that any given person has all of their data converted correctly (i.e., all categories of data for a person in database A matches their data in database b)?

 

[tkhunny]

If you assume the independence of the five processes (which i have to believe is wonderfully dubious) and that the amount of data in teh five separate proces is aboutthe same it is easy enough, but you will not be happy about it.

0.80 * 0.75 * 0.65 * 0.90 * 0.85

If you want it a piece at a time:

Demographic Only

80% Good 20% Bad

Add the health

75% of 80% = 60% Good Left (40% already ruled bad, either now or previously)

Add the Dental

65% of 60% = 39% Good left (61% already ruled bad, either now or previously)

Add the Fianncial

90% of 39% = 35.1% Good left (64.9% already ruled bad, either now or previously)

Add the Loan

85% of 35.1% = 29.835% Good left (70.165% already ruled bad, either now or previously)