Pierre-Edouard Portier
|
96e55f2c72
|
experiment with associating mca clusters with nakr errors
|
2023-03-27 12:04:40 +02:00 |
Pierre-Edouard Portier
|
5689bf6fef
|
Nystroem approximation is now correct (and simplified).
|
2023-03-19 18:48:47 +01:00 |
Pierre-Edouard Portier
|
3c83c256b5
|
progress on nystroem approximation implementation ; there is still a problem: for a small sample size the result of krr should be the same as the one of nystroem approx krr where the number of landmarks is equal to the number of training points. For now, it isn't.
|
2023-03-13 13:01:40 +01:00 |
Pierre-Edouard Portier
|
38d8f0a401
|
Experiments. Where is the error in nakr?
|
2023-01-29 17:52:53 +01:00 |
Pierre-Edouard Portier
|
472b35e38c
|
minor
|
2023-01-23 22:04:52 +01:00 |
Pierre-Edouard Portier
|
d3639fe2ef
|
one hot encoding function / correspondence analysis based landmarks for nystroem approx / experiment on housing dataset
|
2023-01-22 21:27:38 +01:00 |