Outlier detection for spectroscopic classification
In some implementations, a device may determine that an unknown sample is an outlier sample by using an aggregated classification model. The device may determine that one or more spectroscopic measurements are not performed accurately based on determining that the unknown sample is the outlier sample. The device may cause one or more actions based on determining the one or more spectroscopic measurements are not performed accurately.
1 . A method, comprising:
receiving, by a device that includes a processor, information identifying results of a set of spectroscopic measurements of a training set of known samples and a validation set of known samples;
causing, by the device, one or more spectrometers to perform one or more spectroscopic measurements of an unknown sample;
receiving, by the device and based on causing the one or more spectrometers to perform the one or more spectroscopic measurements of the unknown sample, information identifying a result of the one or more spectroscopic measurements of the unknown sample; and
causing, by the device, one or more actions based on a local classification model indicating that the unknown sample is an outlier sample,
the local classification model being based on an aggregated classification model and the result of the one or more spectroscopic measurements of the unknown sample, and
the aggregated classification model being based on the information identifying the results of the set of spectroscopic measurements of the training set and the validation set.
2 . The method of claim 1 , wherein the aggregated classification model includes at least one class relating to a material of interest for a spectroscopic determination.
3 . The method of claim 1 , wherein the aggregated classification model includes a no-match class relating to at least one of at least one material that is not of interest or a baseline spectroscopic measurement.
4 . The method of claim 1 , further comprising:
selecting a set of top classes from the aggregated classification model; and
generating the local classification model by generating a set of local classes of the local classification model based on the set of top classes from the aggregated classification model,
wherein a particular local class, of the set of local classes, includes the unknown sample.
5 . The method of claim 4 , further comprising:
determining that the local classification model indicates that the unknown sample is the outlier sample based on an evaluation of the unknown sample relative to the particular local class or the set of local classes.
6 . The method of claim 1 , wherein causing the one or more actions comprises:
providing output indicating that the outlier sample is included in a no-match class.
7 . A system, comprising:
one or more memories; and
one or more processors, coupled to the one or more memories, configured to:
receive information identifying results of a set of spectroscopic measurements;
cause one or more spectrometers to perform one or more spectroscopic measurements of an unknown sample;
receive, based on the one or more spectrometers being caused to perform the one or more spectroscopic measurements of the unknown sample, information identifying a result of the one or more spectroscopic measurements of the unknown sample; and
cause one or more actions based on a local classification model indicating that the unknown sample is an outlier sample,
the local classification model being based on an aggregated classification model and the result of the one or more spectroscopic measurements of the unknown sample, and
the aggregated classification model being based on the information identifying the results of the set of spectroscopic measurements.
8 . The system of claim 7 , wherein the aggregated classification model includes at least one class relating to a material of interest for a spectroscopic determination.
9 . The system of claim 7 , wherein the aggregated classification model includes a no-match class relating to at least one of at least one material that is not of interest or a baseline spectroscopic measurement.
10 . The system of claim 7 , wherein the one or more processors are further configured to:
select a set of top classes from the aggregated classification model; and
generate the local classification model by generating a set of local classes of the local classification model based on the set of top classes from the aggregated classification model,
wherein a particular local class, of the set of local classes, includes the unknown sample.
11 . The system of claim 10 , wherein the one or more processors are further configured to:
determine that the local classification model indicates that the unknown sample is the outlier sample based on an evaluation of the unknown sample relative to the particular local class.
12 . The system of claim 7 , wherein the one or more processors are further configured to:
determine that the local classification model indicates that the unknown sample is the outlier sample based on an evaluation of the unknown sample relative to the set of local classes.
13 . The system of claim 7 , wherein the one or more processors, to cause the one or more actions, are configured to:
provide output indicating that the outlier sample is included in a no-match class.
14 . A non-transitory computer-readable medium storing a set of instructions, the set of instructions comprising:
one or more instructions that, when executed by one or more processors, cause the one or more processors to:
receive information identifying results of a set of spectroscopic measurements of a training set of known samples and a validation set of known samples;
cause one or more spectrometers to perform one or more spectroscopic measurements of an unknown sample;
receive, based on the one or more spectrometers being caused to perform the one or more spectroscopic measurements of the unknown sample, information identifying a result of the one or more spectroscopic measurements of the unknown sample; and
cause one or more actions based on a local classification model indicating information regarding the unknown sample,
the local classification model being based on an aggregated classification model and the result of the one or more spectroscopic measurements of the unknown sample, and
the aggregated classification model being based on the information identifying the results of the set of spectroscopic measurements of the training set and the validation set.
15 . The non-transitory computer-readable medium of claim 14 , wherein the aggregated classification model includes at least one class relating to a material of interest for a spectroscopic determination.
16 . The non-transitory computer-readable medium of claim 14 , wherein the aggregated classification model includes a no-match class relating to at least one of at least one material that is not of interest or a baseline spectroscopic measurement.
17 . The non-transitory computer-readable medium of claim 14 , wherein the one or more instructions further cause the one or more processors to:
generate the local classification model by generating a set of local classes of the local classification model; and
determine that the local classification model indicates that the unknown sample is an outlier sample based on an evaluation of the unknown sample relative to the set of local classes.
18 . The method of claim 1 , further comprising:
causing the one or more spectrometers to perform the set of spectroscopic measurements of the training set and the validation set,
wherein the information identifying the results of the set of spectroscopic measurements of the training set and the validation set are received based on causing the one or more spectrometers to perform the set of spectroscopic measurements of the training set and the validation set.
19 . The system of claim 7 , wherein the one or more processors are further configured to:
cause the one or more spectrometers to perform the set of spectroscopic measurements on a training set of known samples and a validation set of known samples; and
wherein the one or more processors, to receive the information identifying the results of the set of spectroscopic measurements, are configured to:
receive the information identifying the results of the set of spectroscopic measurements based on the set of spectroscopic measurements being performed on the training set and the validation set.
20 . The non-transitory computer-readable medium of claim 14 , wherein the one or more instructions further cause the one or more processors to:
cause the one or more spectrometers to perform the set of spectroscopic measurements of the training set and the validation set.