Back to results
US 2020/0051669 A1Published application
Method for Characterizing a Sample by Data Analysis
Assigneenumares AG
InventorsFritz Huber, Renate Kirchhöfer, Volker Pfahlert
Life of this patent
Estimating the term…
Abstract
A method for characterizing a sample is disclosed, having the following steps: providing at least one analysis result having a plurality of values, wherein the analysis result was generated by the analysis of a sample by at least one analysis method; determining the value of at least one mathematic relation between at least two values of the plurality of values; generating a characterizing signature of the sample on the basis of the value of the at least one mathematic relation. Furthermore, a method for characterizing a system is disclosed in which method the preceding method is used.
Claims13
A method for characterizing a system, comprising the following steps:
- providing a first analysis result having a plurality of values, wherein the first analysis result was generated by an analysis of a first sample taken from a system by an analysis method chosen from a group consisting of NMR spectroscopy, mass spectrometry, electron spin resonance, vibrational spectroscopy, UV/VIS spectroscopy, and fluorescence spectroscopy;
- determining a first set of values of a quotient between pairs of two values of a first plurality of values;
- generating a characterizing signature of a first sample on a basis of the first set of values of the quotient by weighting the first set of values of the quotient with respect to their significance by carrying out the following steps:
- a) determining a first deviation between the first set of values of the quotient and a first average correlation matrix that has been obtained by summing up a plurality of a second set of values of a quotient, wherein the second set of values of the quotient has been obtained by providing a second analysis result having a second plurality of values, wherein the second analysis result method is the same analysis used for the first analysis result, and determining the second set of values of a quotient between the first set of values;
- b) determining a second deviation between the first set of values of the quotient and a second average correlation matrix that has been obtained by summing up a plurality of a third set of values of a quotient, wherein the third set of values of the quotient has been obtained by providing a third analysis result having a third plurality of values, wherein the third analysis result was generated by an analysis of a sample belonging to a second group by the same analysis method used for the first analysis result, and determining the third set of values of a quotient between a pair of two values of the third plurality of values;
- c) providing a weighting set of values having a value of 1.0 each;
- d) raising one single value of the weighting set by a factor;
- e) multiplying each value of the first set of values of quotients with an according value of the weighting set;
- f) determining a first deviation between the first set of values of the quotient and the first average correlation matrix, and determining a second deviation between the first set of values of the quotient and the second average correlation matrix;
- g) if the determined first deviation is smaller than in the preceding step and/or if the determined second deviation is bigger than in the preceding step raising the single value of the weighting set again by a factor;
- if the determined first deviation is bigger than in the preceding step and/or if the determined second deviation is smaller than in the preceding step, lowering the single value of the weighting set by a factor;
- h) repeating steps e) to g) until the single value reaches asymptotically a value being best for separation, or asymptotically zero;
- i) setting the single value to zero if it is lower than a pre-determined threshold value;
- and
- j) repeating steps d) to i) for all other values of the weighting set in order to obtain an optimized weighting set serving as characterizing signature;
- multiplying the values of the quotient with the values of the characterizing signature to obtain a weighted set of values being characteristic for the first sample;
- and
- using the weighted set of values for characterizing the first sample of the system.
15 more elements
A computer program product for characterizing a system, comprising a non-transitory computer-readable medium having program instructions that, when executed, cause a processor to perform the steps of:
- receiving a first analysis result having a plurality of values, wherein the first analysis result was generated by an analysis of a first sample taken from a system by an analysis method chosen from the group consisting of NMR spectroscopy, mass spectrometry, electron spin resonance, vibrational spectroscopy, UV/VIS spectroscopy, and fluorescence spectroscopy;
- determining a first set of values of a quotient between pairs of two values of a first plurality of values;
- generating a characterizing signature of a first sample on a basis of the first set of values of the quotient by weighting the first set of values of the quotient with respect to their significance by carrying out the following steps:
- a) determining a first deviation between the first set of values of the quotient and a first average correlation matrix that has been obtained by summing up a plurality of a second set of values of a quotient, wherein the second set of values of the quotient has been obtained by providing a second analysis result having a second plurality of values, wherein the second analysis result method is the same analysis used for the first analysis result, and determining the second set of values of a quotient between the first set of values;
- b) determining a second deviation between the first set of values of the quotient and a second average correlation matrix that has been obtained by summing up a plurality of a third set of values of a quotient, wherein the third set of values of the quotient has been obtained by providing a third analysis result having a third plurality of values, wherein the third analysis result was generated by an analysis of a sample belonging to a second group by the same analysis method used for the first analysis result, and determining the third set of values of a quotient between a pair of two values of the third plurality of values;
- c) providing a weighting set of values having a value of 1.0 each;
- d) raising one single value of the weighting set by a factor;
- e) multiplying each value of the first set of values of quotients with an according value of the weighting set;
- f) determining a first deviation between the first set of values of the quotient and the first average correlation matrix, and determining a second deviation between the first set of values of the quotient and the second average correlation matrix;
- g) if the determined first deviation is smaller than in the preceding step and/or if the determined second deviation is bigger than in the preceding step raising the single value of the weighting set again by a factor;
- if the determined first deviation is bigger than in the preceding step and/or if the determined second deviation is smaller than in the preceding step, lowering the single value of the weighting set by a factor;
- h) repeating steps e) to g) until the single value reaches asymptotically a value being best for separation, or asymptotically zero;
- i) setting the single value to zero if it is lower than a pre-determined threshold value;
- and
- j) repeating steps d) to i) for all other values of the weighting set in order to obtain an optimized weighting set serving as characterizing signature;
- multiplying the values of the quotient with the values of the characterizing signature to obtain a weighted set of values being characteristic for the first sample;
- and
- using the weighted set of values for characterizing the first sample of the system.
15 more elements
Dependencies are inferred from the claim text.