Product performance prediction modeling to predict final product performance in case of device exception

1. A product performance prediction modeling method, comprising:

acquiring first sample data, the first sample data including first device outlier data generated in a process of manufacturing a product by a device;

acquiring product information and a production line configuration simulation parameter of a production line relating to location of the device, the product information being product information of the product manufactured by the production line;

selecting a product performance simulation model to perform a simulation test on a performance of the product to obtain product performance simulation data, wherein the product performance simulation model is selected according to the first device outlier data, the production line configuration simulation parameter and the product information; and

inputting the first device outlier data, the production line configuration simulation parameter, the product information and the product performance simulation data into a machine learning model to perform machine learning training to obtain a product performance prediction model.

2. The method of claim 1 , wherein the acquiring of the production line configuration simulation parameter comprises:

acquiring an application type of the device and a category of the product manufactured by the device, the application type of the device including an application field of the device and a production line in which the device is applied; and

acquiring the production line configuration simulation parameter by matching a category of the product manufactured by the device, the application field of the device and the production line in which the device is applied.

3. The method of claim 1 , wherein inputting of the first device outlier data, the production line configuration simulation parameter, the product information and the product performance simulation data, comprises:

partitioning the first device outlier data to obtain training set data and test set data;

inputting the training set data, the production line configuration simulation parameter, the product information and the product performance simulation data into the machine learning model to perform training to obtain an initial product performance prediction model; and

inputting the test set data, the production line configuration simulation parameter, the product information and the product performance simulation data into the initial product performance prediction model, and verifying the initial product performance prediction model, to obtain the product performance prediction model.

4. The method of claim 3 , wherein the inputting of the test set data, the production line configuration simulation parameter, the product information and the product performance simulation data, comprises:

inputting the test set data, the production line configuration simulation parameter, the product information and the product performance simulation data into the initial product performance prediction model to test the initial product performance prediction model to obtain a product performance prediction accuracy of the initial product performance prediction model; and

comparing the product performance prediction accuracy with a threshold, and determining a mapping relationship between the first device outlier data and a product performance prediction result upon the comparing indicating that the product performance prediction accuracy is higher than the threshold, to obtain the product performance prediction model.

5. The method of claim 1 , further comprising:

acquiring second sample data, the second sample data including device outlier data generated in the process of manufacturing the product by the device, the second sample data not overlapping with the first sample data; and

updating the product performance prediction model according to the second sample data.

6. The method of claim 5 , wherein the updating of the product performance prediction model, comprises:

inputting the second sample data, the production line configuration simulation parameter, the product information and the product performance simulation data into the product performance prediction model to perform prediction to obtain a product performance prediction result;

judging whether a deviation between the product performance prediction result and an actual value is greater than a prediction standard; and

updating the product performance prediction model upon the judging indicating that the deviation is greater than the prediction standard.

7. The method of claim 6 , further comprising:

adjusting, upon the judging indicating that the deviation is than the prediction standard, the production line configuration simulation parameter until the deviation between the product performance prediction result and the actual value is smaller than the prediction standard.

8. The method of claim 1 , further comprising:

acquiring third sample data, the third sample data including third device outlier data generated in the process of manufacturing the product by the device;

acquiring product performance test data corresponding to the third sample data;

inputting the third sample data, the production line configuration simulation parameter, the product information and the product performance test data into the product performance prediction model to perform verification;

determining a mapping relationship between the third device outlier data in the third sample data and the product performance test data; and

updating the product performance prediction model.

9. The method of claim 1 , further comprising:

acquiring original data;

judging whether the original data pertains to outlier data;

further judging whether the original data is associated with a product performance upon the original data being judged to not pertain to the outlier data;

inputting, upon the further judging indicating that the original data is associated with the product performance, the original data into the product performance prediction model for prediction to obtain a product performance prediction result; and

updating the product performance prediction model according to a comparison result between the product performance prediction result and a product performance actual value.

10. A product performance prediction method, comprising:

acquiring device outlier data;

acquiring product information and a production line configuration simulation parameter of a production line relating to location of a device generating the device outlier data, the product information being product information of a product manufactured by the production line; and

inputting the device outlier data, the production line configuration simulation parameter and the product information into the product performance prediction model established via the product performance prediction modeling method of claim 1 , so as to obtain a product performance prediction result corresponding to the device outlier data.

11. The method of claim 10 , further comprising:

acquiring an initial product performance simulation model according to the device outlier data, the production line configuration simulation parameter and the product information; and

correcting a model parameter of the initial product performance simulation model according to the product performance prediction result, so that a deviation between product performance simulation data of the initial product performance simulation model and the product performance prediction result satisfies a standard, and so that the product performance simulation model is obtained.

12. The method of claim 2 , further comprising:

acquiring second sample data, the second sample data including second device outlier data generated in the process of manufacturing the product by the device, the second sample data not overlapping with the first sample data; and

updating the product performance prediction model according to the second sample data.

13. The method of claim 3 , further comprising:

acquiring second sample data, the second sample data including second device outlier data generated in the process of manufacturing the product by the device, the second sample data not overlapping with the first sample data; and

updating the product performance prediction model according to the second sample data.

14. The method of claim 2 , further comprising:

acquiring third sample data, the third sample data including third device outlier data generated in the process of manufacturing the product by the device;

acquiring product performance test data corresponding to the third sample data;

inputting the third sample data, the production line configuration simulation parameter, the product information and the product performance test data into the product performance prediction model to perform verification;

determining a mapping relationship between the third device outlier data in the third sample data and the product performance test data; and

updating the product performance prediction model.

15. The method of claim 3 , further comprising:

acquiring third sample data, the third sample data including third device outlier data generated in the process of manufacturing the product by the device;

acquiring product performance test data corresponding to the third sample data;

inputting the third sample data, the production line configuration simulation parameter, the product information and the product performance test data into the product performance prediction model to perform verification;

determining a mapping relationship between the third device outlier data in the third sample data and the product performance test data; and

updating the product performance prediction model.

16. A product performance prediction modeling apparatus, comprising:

a sample acquisition module configured to acquire first sample data, the first sample data including device outlier data generated in a process of manufacturing a product by a device;

a parameter acquisition module configured to acquire product information and a production line configuration simulation parameter of a production line relating to location of the device, the product information being product information of the product manufactured by the production line;

a performance simulation module configured to select a product performance simulation model to perform a simulation test on a performance of the to obtain product performance simulation data, wherein the product performance simulation model is selected according to the device outlier data, the production line configuration simulation parameter and the product information; and

a modeling module configured to input the device outlier data, the production line configuration simulation parameter, the product information and the product performance simulation data into a machine learning model to perform machine learning training to obtain a product performance prediction model.

17. A computer device, comprising:

a memory storing a computer program; and

a processor configured to, when the processor executes the computer program, cause the computer device to at least

acquire first sample data, the first sample data including device outlier data generated in a process of manufacturing a product by a device,

acquire product information and a production line configuration simulation parameter of a production line relating to location of the device, the product information being product information of the product manufactured by the production line,

select a product performance simulation model to perform a simulation test on a performance of the product to obtain product performance simulation data, wherein the product performance simulation model is selected according to the device outlier data, the production line configuration simulation parameter and the product information, and

input the device outlier data, the production line configuration simulation parameter, the product information and the product performance simulation data into a machine learning model to perform machine learning training to obtain a product performance prediction model.