Timeline for Why feature normalization can increase the biometric recognition accuracy?
Current License: CC BY-SA 4.0
7 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 2, 2022 at 4:16 | comment | added | Noha | @Oxbowerce For two fingerprints from the same person, I think that their wavelet feature vectors will be in a similar scale, whereas the large difference in scale will be between two feature vectors from two fingerprints for different persons. And we need minimum distance between samples from the same person, and maximum distance between different persons. | |
| Jan 2, 2022 at 1:25 | answer | added | Brian Spiering | timeline score: 1 | |
| Jan 1, 2022 at 21:58 | comment | added | Oxbowerce | Let's take the extreme case where you have two features with values in the range 0-1000000 and 0-100 respectively. In that case the first feature will have a relatively larger effect when assigning new points when using Euclidean distance. See also this answer from stats.stackexchange which also contains a visual representation. | |
| Jan 1, 2022 at 21:25 | comment | added | Noha | @Oxbowerce Could you please give a numeric example? | |
| Jan 1, 2022 at 21:10 | comment | added | Oxbowerce | Because when not normalizing your features before using a KNN model the model gives a higher relative weight to features with a larger range of values than to features with a smaller range of values. | |
| S Jan 1, 2022 at 20:35 | review | First questions | |||
| Jan 2, 2022 at 1:55 | |||||
| S Jan 1, 2022 at 20:35 | history | asked | Noha | CC BY-SA 4.0 |