# US 2018/0322399 A1 · Endpoint detection in manufacturing process by near infrared spectroscopy and machine learning techniques > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2018/0322399 A1 - Document ID: US20180322399A1 - Status: Published application - Kind: A1 - Application number: 15/586,678 - Filed: 2017-05-04 - Published: 2018-11-08 - Earliest claimed: 2017-05-04 - Assignees: Viavi Solutions Inc. - Applicants: Viavi Solutions Inc. - Inventors: Changmeng HSIUNG; Peng Zou; Lan Sun - Web page: https://patentgrep.com/patent/US20180322399A1 ## CPC - G06N 5/04: Inference or reasoning models - G06N 20/00: Machine learning - G06N 20/10: using kernel methods, e.g. support vector machines [SVM] - G06N 20/20: Ensemble learning - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract A device may receive training spectral data associated with a manufacturing process that transitions from an unsteady state to a steady state. The device may generate, based on the training spectral data, a plurality of iterations of a support vector machine (SVM) classification model. The device may determine, based on the plurality of iterations of the SVM classification model, a plurality of predicted transition times associated with the manufacturing process. A predicted transition time, of the plurality of predicted transition times, may identify a time, during the manufacturing process, that a corresponding iteration of the SVM classification model predicts that the manufacturing process transitioned from the unsteady state to the steady state. The device may generate, based on the plurality of predicted transition times, a final SVM classification model associated with determining whether the manufacturing process has reached the steady state. ## Claims ### Claim 1 A device, comprising: one or more processors to: receive training spectral data associated with a manufacturing process that transitions from an unsteady state to a steady state; generate, based on the training spectral data, a plurality of iterations of a support vector machine (SVM) classification model; determine, based on the plurality of iterations of the SVM classification model, a plurality of predicted transition times associated with the manufacturing process, a predicted transition time, of the plurality of predicted transition times, identifying a time, during the manufacturing process, that a corresponding iteration of the SVM classification model predicts that the manufacturing process transitioned from the unsteady state to the steady state; and generate, based on the plurality of predicted transition times, a final SVM classification model associated with determining whether the manufacturing process has reached the steady state. ### Claim 2 (depends on 1) The device of claim 1, where the one or more processors are further to: receive additional spectral data associated with the manufacturing process; and determine, based on the final SVM classification model and the additional spectral data, that the manufacturing process has not reached the steady state. ### Claim 3 (depends on 1) The device of claim 1, where the one or more processors are further to: receive additional spectral data associated with the manufacturing process; and determine, based on the final SVM classification model and the additional spectral data, that the manufacturing process has reached the steady state. ### Claim 4 (depends on 3) The device of claim 3, where the one or more processors are further to: provide an indication that the manufacturing process has reached the steady state. ### Claim 5 (depends on 3) The device of claim 3, where the one or more processors are further to: determine, based on determining that the manufacturing process has reached the steady state, a quantitative metric associated with the steady state; and provide information associated with the quantitative metric. ### Claim 6 (depends on 1) The device of claim 1, where the one or more processors, when generating the plurality of iterations of the SVM classification model, are to: create, based on the training spectral data, a set of unsteady state data and a set of steady state data; and generate an iteration, of the plurality of iterations of the SVM classification model, based on the set of unsteady state data and the set of steady state data. ### Claim 7 (depends on 6) The device of claim 6, where the one or more processors, when determining the plurality of predicted transition times, are to: determine, based on the training spectral data and the iteration of the SVM classification model, the predicted transition time, associated with the iteration of the SVM classification model, the predicted transition time being one of the plurality of predicted transition times. ### Claim 8 (depends on 1) The device of claim 1, where the one or more processors are further to: perform a dimension reduction, based on the training spectral data, before generating the plurality of iterations of the SVM classification model. ### Claim 9 A non-transitory computer-readable medium storing instructions, the instructions comprising: one or more instructions that, when executed by one or more processors, cause the one or more processors to: receive training spectral data associated with a first performance of a manufacturing process that transitions from an unsteady state to a steady state; iteratively generate, based on the training spectral data, a support vector machine (SVM) classification model associated with determining whether another performance of the manufacturing process has transitioned from the unsteady state to the steady state; receive additional spectral data associated with a second performance of the manufacturing process; and determine, based on the SVM classification model and the additional spectral data, whether the second performance of the manufacturing process has transitioned from the unsteady state to the steady state. ### Claim 10 (depends on 9) The non-transitory computer-readable medium of claim 9, where the one or more instructions that cause the one or more processors to determine whether the second performance of the manufacturing process has transitioned from the unsteady state to the steady state, cause the one or more processors to: determine that the second performance of the manufacturing process has transitioned from the unsteady state to the steady state; and provide an indication that the second performance of the manufacturing process has transitioned from the unsteady state to the steady state. ### Claim 11 (depends on 10) The non-transitory computer-readable medium of claim 10, where the one or more instructions, when executed by the one or more processors, further cause the one or more processors to: determine, based on determining that the second performance of the manufacturing process has transitioned from the unsteady state to the steady state, a quantitative metric associated with the second performance of the manufacturing process; and provide information associated with the quantitative metric. ### Claim 12 (depends on 11) The non-transitory computer-readable medium of claim 11, where the one or more instructions, when executed by the one or more processors, further cause the one or more processors to: identify a regression model associated with the quantitative metric; and where the one or more instructions, that cause the one or more processors to determine the quantitative metric, cause the one or more processors to: determine the quantitative metric based on the regression model. ### Claim 13 (depends on 9) The non-transitory computer-readable medium of claim 9, where the one or more instructions, that cause the one or more processors to iteratively generate the SVM classification model, cause the one or more processors to: create, based on the training spectral data, a first set of unsteady state data and a first set of steady state data; generate a first iteration of the SVM classification model based on the first set of unsteady state data and the first set of steady state data; create, based on the training spectral data, a second set of unsteady state data and a second set of steady state data; and generate a second iteration of the SVM classification model based on the second set of unsteady state data and the second set of steady state data. ### Claim 14 (depends on 13) The non-transitory computer-readable medium of claim 13, where the one or more instructions, when executed by the one or more processors, further cause the one or more processors to: determine, based on the first iteration of the SVM classification model, a predicted transition time associated with the first iteration of the SVM classification model; and where the one or more instructions that cause the one or more processors to create the second set of unsteady state data, cause the one or more processors to: create the second set of unsteady state data based on the predicted transition time associated with the first iteration of the SVM classification model. ### Claim 15 A method, comprising: receiving, by a device, first spectral data associated with a first performance of a manufacturing process that transitions from an unsteady state to a steady state; generating, by the device and based on the first spectral data, a plurality of iterations of a support vector machine (SVM) classification model; determining, by the device and based on the plurality of iterations of the SVM classification model, a plurality of predicted transition times associated with the first performance of the manufacturing process; generating, by the device and based on the plurality of predicted transition times, a final SVM classification model associated with determining whether another performance of the manufacturing process has reached the steady state; receiving, by the device, second spectral data associated with a second performance of the manufacturing process; and determining, by the device, whether the second performance of the manufacturing process has reached the steady state based on the final SVM classification model and the second spectral data. ### Claim 16 (depends on 15) The method of claim 15, where determining whether the second performance of the manufacturing process has reached the steady state comprises: determining that the second performance of the manufacturing process has not reached the steady state. ### Claim 17 (depends on 15) The method of claim 15, where determining whether the second performance of the manufacturing process has reached the steady state comprises: determining that the second performance of the manufacturing process has reached the steady state. ### Claim 18 (depends on 17) The method of claim 17, further comprising: providing an indication that the manufacturing process has reached the steady state, where providing the indication causes an action, associated with the manufacturing process, to be automatically performed. ### Claim 19 (depends on 17) The method of claim 17, further comprising: determining, based on a regression model associated with the manufacturing process, a quantitative metric associated with the steady state; and providing information associated with the quantitative metric. ### Claim 20 (depends on 15) The method of claim 15, further comprising: performing a dimension reduction, associated with the first spectral data, before generating the plurality of iterations of the SVM classification model.