How to get the original matrix from derived matrix?!

2 次查看（过去 30 天）

Jussara Roque 2015-5-13

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/216603-how-to-get-the-original-matrix-from-derived-matrix

评论： Walter Roberson 2015-5-14

Hello!!

I am deriving the matrix of original spectra to make some corrections in all spectra. But after these corrections I need the original matrix. So, how do I get the original matrix back from the derived matrix?

Please, i need help!!

Many thanks for your help!

Jussara

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

回答（1 个）

Star Strider 2015-5-13

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/216603-how-to-get-the-original-matrix-from-derived-matrix#answer_179024

I would simply copy the original matrix to a separate variable. Then there is no need to recover it.

7 个评论
显示 5更早的评论隐藏 5更早的评论

Jussara Roque 2015-5-14

I'm using the savgol for calculate the derived matrix the samples and the spectra the solvent. And to correct the solvent signal, I'm subtracting the spectra of each sample and the spectra the solvent. So, the spectra resulting this operation is obtained derived and I want this new matrix with the original data (without derived but with correction the solvent applied in the derived matrix).

Walter Roberson 2015-5-14

在 MATLAB Online 中打开

Star Strider: the "baseline" has been subtracted out. Something akin to:

derived_matrix = f(A);   %for some formula f
[slope, intercept] = do_fitting(derived_matrix);
corrected_derivation = (derived_matrix - intercept)/slope;   %normalize using estimated parameters
corrected_A = f_inverse(corrected_derivation);

That is, representation is extracted from the original matrix, used to calculate a correction factor, the correction is applied to the representation information, and now the corrected representation should be transformed back to the original structure.

At the moment we can't say anything about how to do that back-calculation because we do not know how the representation is calculated.

请先登录，再进行评论。

请先登录，再回答此问题。