Using prior knowledge in SVD-based NMR spectroscopy - the ATP example
Yngve Selén, Peter Stoica, Niclas Sandgren, Sabine Van Huffel
Abstract
Yngve Selén, Peter Stoica, Niclas Sandgren, Sabine Van Huffel
Abstract
We introduce the KNOB-SVD (knowledge based singular value decomposition) method for exploiting prior knowledge in NMR spectroscopy based on the singular value decomposition (SVD) of the data matrix. The ATP (adenosine triphosphate) complex, often modeled as a sum of seven exponentially damped sinusoids, is used as a vehicle for describing our SVD-based method throughout the paper. By means of a simple numerical example we show that our method provides more accurate parameter estimates than a commonly used general-purpose SVD-based method and a previously suggested prior knowledge-based SVD method using the same type of prior knowledge as considered herein.
OpenAlex reports 4 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
We introduce the KNOB-SVD (knowledge based singular value decomposition) method for exploiting prior knowledge in NMR spectroscopy based on the singular value decomposition (SVD) of the data matrix. The ATP (adenosine triphosphate) complex, often modeled as a sum of seven exponentially damped sinusoids, is used as a vehicle for describing our SVD-based method throughout the paper. By means of a simple numerical example we show that our method provides more accurate parameter estimates than a commonly used general-purpose SVD-based method and a previously suggested prior knowledge-based SVD method using the same type of prior knowledge as considered herein.
Key concepts: Singular value decomposition, Singular value, Matrix (chemical analysis), Decomposition, Computer science, Value (mathematics), Nuclear magnetic resonance spectroscopy, Matrix decomposition