|
Tuning a search engine to attain two different scenarios of retrieval. |
||||
| Scenario 1. Query with specificity of 99.99% is insufficient for a database of 16 million records. |
||||
|
|
||||
| The truth |
||||
|
|
||||
| relevant records |
irrelevant records |
|||
|
|
||||
| search engine |
records returned to user |
495 |
1,600 |
2,095 |
| records eliminated |
5 |
15,997,900 |
||
|
|
||||
| 500 |
15,999,500 |
16,000,000 |
||
|
|
||||
| odds ratio |
1,000,000.00 |
|||
| Specificity |
99.99% |
|||
| sensitivity (recall) |
99.01% |
|||
| Precision |
23.63% |
|||
|
|
||||
| Scenario 2. The price for a very high specificity: Missing a large number of relevant records. |
||||
|
|
||||
| The truth |
||||
|
|
||||
| relevant records |
irrelevant records |
|||
|
|
||||
| search engine |
records returned to user |
250 |
16 |
266 |
| records eliminated |
250 |
15,999,484 |
||
|
|
||||
| 500 |
15,999,500 |
16,000,000 |
||
|
|
||||
| odds ratio |
1,000,000.00 |
|||
| Specificity |
99.9999% |
|||
| sensitivity (recall) |
50.00% |
|||
| Precision |
93.99% |
|||
Siadaty et al. BMC Medical Informatics and Decision Making 2007 7:1 doi:10.1186/1472-6947-7-1 |
||||