This week we dicussed the advantages and potential use cases of interpolation between bases. The key points included:
It can be useful when we care about the orientation of the projection. For example, it can be used when we are interested in detecting non-linear association in the projections.
In case of bases interpolation, we do not need to do the preprojection. It is suggested in Computational Methods for High-Dimensional Rotations in Data Visualization (Buja et. al 2004) because of the computational simplicity.
The summary of next steps forward is listed below.
[x] Borrow a copy of Matrix computations (Golub & Van)
[x] Create R package structure in the repository using "usethis" package
[x] Code to make 2D rotation between arbitrary vectors x_1 (12) and x_2 (12)
(i) Calculate the angle between the vectors
(ii) Break the angle into k pieces
(iii) Rotate x_1 k times to get to x_2
(iv) Make diagnostic plots to show the process
(v) List possible edge cases that need to be detected
[x] Further reading on data applications in detecting non-linear structures
This week we dicussed the advantages and potential use cases of interpolation between bases. The key points included:
The summary of next steps forward is listed below.
[x] Borrow a copy of Matrix computations (Golub & Van)
[x] Create R package structure in the repository using "usethis" package
[x] Code to make 2D rotation between arbitrary vectors x_1 (12) and x_2 (12)
(i) Calculate the angle between the vectors
(ii) Break the angle into k pieces
(iii) Rotate x_1 k times to get to x_2
(iv) Make diagnostic plots to show the process
(v) List possible edge cases that need to be detected
[x] Further reading on data applications in detecting non-linear structures
Cross-currency rates
Physics
Simulated data