a rectangular block that is N x L of pixel values of continuum-normalized spectral intensities (training set)
a parallel block of inverse variance values
a block of N x K labels
a rectangular block that is M x L of pixel values of continuum-normalized spectral intensities (test set)
a parallel block of inverse variance values
a value for the "polynomial order" of the model
Output:
a block of M x K labels
a block of Q x L parameters (coefficients and scatters and so on)
Continuum Wavelengths Code:
Input:
the first three inputs to the Main Code
Output:
a length-L mask of continuum True/False values
Continuum Normalization Code:
Input:
a N x L or M x L block of pixel values
a parallel block of inverse variance values
the length-L continuum mask (from above)
Output
a N x L or M x L block of continuum-normalized intensities
a parallel block of rescaled inverse variance values
APOGEE Munging Code:
Input:
a URL, or maybe nothing
Output:
a length-L rest-frame wavelength vector
a block N x L of training-set pixel intensities
a parallel block of inverse variances
a block N x K of labels
a block M x L of test-set pixel intensities
a parallel block of inverse variances
Comments:
This code should get the data, shift to the rest frame, read the bad-pixel masks, and build the inverse variance vectors and package them all into rectangular blocks.
Main Code:
Input:
Output:
Continuum Wavelengths Code:
Input:
Output:
Continuum Normalization Code:
Input:
Output
APOGEE Munging Code:
Input:
Output:
Comments:
This code should get the data, shift to the rest frame, read the bad-pixel masks, and build the inverse variance vectors and package them all into rectangular blocks.
Workflow: