Equipment indicator excellence grade prediction model training method, monitoring system and method

Abstract

A method for training a model for predicting the quality level of equipment indicators, comprising: obtaining historical parameter values ​​of multiple equipment indicators (S110); performing dimensionality reduction processing on the historical parameter values ​​according to the correlation between the multiple equipment indicators to obtain sample data about the reduced-dimensional variables (S120); determining a positive sample set and a negative sample set according to the quality level of the sample data of the reduced-dimensional variables (S130); training a machine learning model with the positive sample set and the negative sample set to obtain a prediction model for predicting the quality level of the target equipment indicator (S140).

Description

Background Art

[0002] Obviously, during the production process, display panels need to be processed by multiple process equipment in sequence. At the same time, the final display panel products inevitably have a certain probability of various defects, and the quality level of the product is essentially related to the process equipment, specifically caused by the parameter values ​​of the equipment indicators.

[0003] Therefore, determining the correlation between the equipment indicators of the process equipment and the quality level of the product is of great significance for locating defects and adjusting the production process. Summary of the invention

[0004] The embodiments of the present disclosure provide a training method for an equipment indicator excellence grade prediction model, an equipment indicator monitoring system and method, and a computer-readable medium.

[0005] In one embodiment of the present disclosure, a method for training a device indicator excellence grade prediction model is provided, comprising:

[0006] Get historical parameter values ​​of multiple device indicators;

[0007] According to the correlation between the above-mentioned multiple equipment indicators, the above-mentioned historical parameter values ​​are subjected to dimensionality reduction processing to obtain sample data on the dimensionality reduction variables;

[0008] According to the excellence level of the sample data of the above-mentioned dimension reduction variables, a positive sample set and a negative sample set are determined;

[0009] The machine learning model is trained by the positive sample set and the negative sample set to obtain a prediction model for predicting the quality level of the target device indicator.

[0010] In an embodiment of the present disclosure, the dimensionality reduction processing of the historical parameter values ​​according to the association relationship between the multiple device indicators includes:

[0011] Analyze the correlation between the above multiple equipment indicators through principal component analysis algorithm;

[0012] Based on the above association relationship, the related equipment indicators are combined to obtain a reduced dimension variable;

[0013] According to the historical parameter values ​​of the equipment indicators involved in the above-mentioned combination operation, the sample data of the above-mentioned dimension reduction variables are determined.

[0014] In one embodiment of the present disclosure, the training of the machine learning model by the positive sample set and the negative sample set includes:

[0015] Sampling the dimension-reduced variables in the positive sample set and the negative sample set respectively to obtain sampled variables;

[0016] Calculating the eigenvector of the Fourier basis of each of the above sampled variables to determine the correlation between the above sampled variables;

[0017] Processing the eigenvectors of the Fourier basis based on a linear regression algorithm to obtain screening variables;

[0018] The machine learning model is trained based on the above screening variables and the parameter values ​​corresponding to the above screening variables.

[0019] In one embodiment of the present disclosure, the sample data includes training samples and test samples;

[0020] The above-mentioned determination of the positive sample set and the negative sample set according to the excellence level of the sample data of the above-mentioned dimension reduction variable includes:

[0021] According to the excellence level of the training samples in the sample data of the above-mentioned dimension reduction variables, a positive sample set and a negative sample set are determined;

[0022] The method further includes: evaluating the prediction model through the test sample.

[0023] In one embodiment of the present disclosure, the prediction model is evaluated by one or more of the following evaluation indicators, including: accuracy, recall, and area under the receiver operating characteristic curve AUC.

[0024] In an embodiment of the present disclosure, the obtaining of the historical parameter value of each of the above-mentioned device indicators includes:

[0025] Loading source data about equipment indicators into the data warehouse for data mining analysis;

[0026] The data after the mining analysis is stored in a non-relational database for determining the historical parameter values.

[0027] In one embodiment of the present disclosure, the method further includes:

[0028] Get the device index to be cleaned up whose parameter value is null in the data object;

[0029] Obtain the mode value or average value of the attribute value of the device indicator to be cleaned in other data objects;

[0030] Determine the above-mentioned mode value or the above-mentioned average value as the attribute value of the device indicator to be cleaned in the above-mentioned data object.

[0031] In one embodiment of the present disclosure, the above-mentioned machine learning model is a Bayesian algorithm model, a decision tree algorithm model or a neural network model.

[0032] In one embodiment of the present disclosure, a monitoring system for device indicators is provided, comprising: a distributed storage device, and one or more processors, wherein:

[0033] The distributed storage device is configured to respectively obtain first parameter values ​​to be measured for M target device indicators, where M is an integer greater than 1;

[0034] The one or more processors are configured to monitor the excellence of the M target device indicators based on the prediction model.

[0035] In one embodiment of the present disclosure, the one or more processors are further specifically configured to:

[0036] According to the correlation relationship between the M target device indicators, the first parameter value to be measured is subjected to dimensionality reduction processing to obtain the second parameter value to be measured about N target dimensionality reduction variables, wherein N is a positive integer less than M;

[0037] The second parameter value to be measured is input into the prediction model to monitor the quality of the M target device indicators through the output value of the prediction model.

[0038] In one embodiment of the present disclosure, the one or more processors are specifically configured to:

[0039] Analyze the correlation between the above M target device indicators through the principal component analysis algorithm;

[0040] Based on the above association relationship, the associated equipment indicators are combined to obtain N target dimension reduction variables;

[0041] According to the first parameter to be measured of the target device indicator involved in the above-mentioned combination operation, the second parameter to be measured of the above-mentioned target dimension reduction variable is determined.

[0042] In one embodiment of the present disclosure, the one or more processors are specifically configured to:

[0043] Input the second parameter values ​​to be measured for the N target dimension reduction variables into the prediction model, and calculate the eigenvectors of the Fourier basis corresponding to the N target dimension reduction variables respectively;

[0044] Processing the eigenvectors of the N Fourier bases based on a linear regression algorithm to obtain N' target screening variables for the above-mentioned excellence level, wherein N' is less than N;

[0045] Based on the above prediction model, the predicted probability that the parameter value of the above target screening variable belongs to each excellence level is obtained;

[0046] The quality level corresponding to the maximum value of the predicted probability is determined as the output value of the above prediction model.

[0047] In one embodiment of the present disclosure, the one or more processors are further configured to:

[0048] If the output value of the above prediction model is that the above-mentioned excellence level is lower than the preset level value, then the above-mentioned target dimensionality reduction variable is processed by dimensionality increase to obtain a set of faulty equipment indicators.

[0049] In one embodiment of the present disclosure, the one or more processors are further configured to:

[0050] Determine an excellent parameter value interval of the target device indicator to be monitored based on the prediction model, wherein the target device indicator to be monitored is any one of the M target device indicators;

[0051] The current parameter value of the above-mentioned target device indicator to be monitored is monitored in real time according to the above-mentioned excellent parameter value range.

[0052] In one embodiment of the present disclosure, the system further includes: a display device configured to:

[0053] The good parameter value range of the target device indicator to be monitored is displayed, and the current parameter value of the target device indicator to be monitored obtained in real time is displayed.

[0054] In one embodiment of the present disclosure, the system further includes: an alarm device configured to:

[0055] When the difference between the current parameter value of the target device indicator to be monitored and the boundary value of the excellent parameter value interval is less than a preset value, an alarm is issued.

[0056] In one embodiment of the present disclosure, a method for monitoring device indicators is provided, including:

[0057] Obtaining first parameter values ​​to be measured for M target device indicators respectively, where M is an integer greater than 1;

[0058] The prediction model provided by the training method embodiment of the above-mentioned equipment excellence grade prediction model monitors the excellence of the above-mentioned M target equipment indicators.

[0059] In one embodiment of the present disclosure, the monitoring of the excellence of the M target device indicators based on the prediction model includes:

[0060] According to the correlation relationship between the M target device indicators, the first parameter value to be measured is subjected to dimensionality reduction processing to obtain the second parameter value to be measured about N target dimensionality reduction variables, wherein N is a positive integer less than M;

[0061] The second parameter value to be measured is input into the prediction model to monitor the quality of the M target device indicators through the output value of the prediction model.

[0062] In an embodiment of the present disclosure, the dimensionality reduction processing of the first parameter value to be measured according to the association relationship between the M target device indicators includes:

[0063] Analyze the correlation between the above M target device indicators through the principal component analysis algorithm;

[0064] Based on the above association relationship, the associated equipment indicators are combined to obtain N target dimension reduction variables;

[0065] According to the first parameter to be measured of the target device indicator involved in the above-mentioned combination operation, the second parameter to be measured of the above-mentioned target dimension reduction variable is determined.

[0066] In one embodiment of the present disclosure, a computer-readable medium is provided, on which a computer program is stored, wherein when the program is executed by a processor, the training method of the equipment indicator excellence grade prediction model as described above is implemented, and the monitoring method of the equipment indicator as described above is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] The accompanying drawings are used to provide a further understanding of the embodiments of the present disclosure and constitute a part of the specification. Together with the embodiments of the present disclosure, they are used to explain the present disclosure and do not constitute a limitation of the present disclosure. The above and other features and advantages will become more apparent to those skilled in the art by describing detailed example embodiments with reference to the accompanying drawings, in which:

[0068] Figure 1 A flowchart of a method for training a device indicator excellence level prediction model provided in an embodiment of the present disclosure;

[0069] Figure 2 A flowchart of a method for obtaining historical parameter values ​​of a device indicator provided by an embodiment of the present disclosure;

[0070] Figure 3 A schematic diagram of a process flow of a dimensionality reduction processing method provided in an embodiment of the present disclosure;

[0071] Figure 4 A flowchart of a model training method provided in an embodiment of the present disclosure;

[0072] Figure 5 A schematic diagram of the structure of a monitoring system for equipment indicators provided in an embodiment of the present disclosure;

[0073] Figure 6A schematic diagram of the structure of an analysis device in a monitoring system for device indicators provided by an embodiment of the present disclosure;

[0074] Figure 7 A schematic diagram of the structure of another device indicator monitoring system provided by an embodiment of the present disclosure;

[0075] Figure 8 A schematic diagram of a monitoring method for device indicators provided in an embodiment of the present disclosure;

[0076] Fig. 9 A schematic diagram of a computer-readable medium provided for an embodiment of the present disclosure;

[0077] Fig.10 A schematic diagram of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0078] In order to enable those skilled in the art to better understand the technical solutions of the embodiments of the present disclosure, the system and method for analyzing the causes of product defects and the computer-readable medium provided by the embodiments of the present disclosure are described in detail below in conjunction with the accompanying drawings.

[0079] The embodiments of the present disclosure will be described more fully below with reference to the accompanying drawings, but the embodiments shown may be embodied in different forms and should not be construed as being limited to the embodiments set forth in the present disclosure. On the contrary, the purpose of providing these embodiments is to make the present disclosure thorough and complete and to enable those skilled in the art to fully understand the scope of the present disclosure.

[0080] The embodiments of the present disclosure may be described with reference to plan views and / or cross-sectional views by way of ideal schematic diagrams of the present disclosure. Therefore, the exemplary illustrations may be modified according to manufacturing techniques and / or tolerances.

[0081] In the absence of conflict, the various embodiments of the present disclosure and the various features therein may be combined with each other.

[0082] The terms used in the present disclosure are only used to describe specific embodiments and are not intended to limit the present disclosure. The term "and / or" as used in the present disclosure includes any and all combinations of one or more related enumerated items. The singular forms "one" and "the" as used in the present disclosure are also intended to include plural forms, unless the context clearly indicates otherwise. The terms "including", "made of..." as used in the present disclosure specify the presence of the features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or groups thereof.

[0083] Unless otherwise defined, all terms (including technical and scientific terms) used in this disclosure have the same meaning as commonly understood by those of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant art and this disclosure, and will not be interpreted as having an idealized or overly formal meaning unless explicitly defined in this disclosure.

[0084] The embodiments of the present disclosure are not limited to the embodiments shown in the drawings, but include modifications of the configurations formed based on the manufacturing process. Therefore, the regions illustrated in the drawings have schematic properties, and the shapes of the regions shown in the drawings illustrate the specific shapes of the regions of the elements, but are not intended to be limiting.

[0085] Many products (such as display panels) are produced through production lines. Each production line includes multiple process stations, each of which is used to perform certain processing (such as cleaning, deposition, exposure, etching, box alignment, testing, etc.) on products (including semi-finished products). At the same time, each process station usually has multiple process equipment for the same processing; of course, although the processing is the same in theory, the actual processing effects of different process equipment are not exactly the same due to differences in model, status, etc.

[0086] The production process of each product needs to go through multiple process stations, and different products may go through different process stations during the production process; and products passing through the same process station may also be processed by different process equipment. Therefore, in a production line, each process equipment will participate in the production process of some products, but not all products, that is, each process equipment participates in and only participates in the production process of some products.

[0087] In a first aspect, an embodiment of the present disclosure provides a method for training a device indicator excellence grade prediction model.

[0088] The equipment indicator quality grade prediction model trained by the training method provided by the embodiment of the present disclosure can predict the parameter values ​​of the equipment indicators, and monitor the predicted parameter values ​​in real time through historical parameter trends, and then predict the possibility of defect occurrence; the correlation between each process equipment in the production line and various defects of the product, that is, determine the cause of the product defect, so as to locate the defect, adjust the production process, etc.

[0089] Furthermore, the equipment indicator excellence grade prediction model obtained by training through the training method provided in the embodiment of the present disclosure can also obtain the optimal trend diagram of the parameters of each equipment indicator, so that problems can be discovered as early as possible during the production of similar products, ensuring that the equipment status is the optimal production status during the production of similar products, thereby improving the yield of similar products in batches.

[0090] Among them, "defective" refers to quality defects in products, which may lead to reduced product quality or even scrapping, and may also cause the product to need rework or repair.

[0091] Among them, defects can be divided into different categories according to needs. For example, they can be classified according to the direct impact of defects on product performance, such as bright line defects, dark line defects, firefly defects (hot spot), etc.; or, they can also be classified according to the specific causes of defects, such as gate line and data line short circuit defects, alignment defects, etc.; or, they can also be classified according to the general causes of defects, such as array process defects, color film process defects, etc.; or, they can also be classified according to the severity of defects, such as defects that lead to scrapping, defects that lead to reduced quality, etc.; or, they can also not distinguish the types of defects, that is, as long as the product has any "defect", it is considered to have "defects", otherwise it is considered to be "good".

[0092] Of course, the above correlation is for "one" defect, that is, the same process equipment has different correlations with different types of defects. Among them, "the correlation between process equipment and a certain defect" refers to the degree of influence of the participation of the process equipment on the probability of whether the product will have this defect.

[0093] In some embodiments, the product is a display panel production line. The disclosed embodiments can be used to determine the correlation between each process equipment and defects in the display panel production line during the production process of the display panel (such as a liquid crystal display panel, an organic light emitting diode display panel, etc.). Of course, the disclosed embodiments can also be used for other products.

[0094] refer to Figure 1 , the embodiment of the present disclosure provides a training method for a device indicator excellence level prediction model, steps S110 to S140.

[0095] In step S110, historical parameter values ​​of multiple device indicators are obtained.

[0096] The above equipment indicators are related to the quality of the product (such as OLED), including: temperature, pressure, humidity, heating time, cooling time, etc. The above equipment indicators can be indicators of multiple equipment or indicators of a certain equipment. The historical parameter values ​​of each of the above equipment indicators include: temperature value within a preset historical time length, pressure value within a preset historical time length, pressure value within a preset historical time length, etc.

[0097] In an exemplary embodiment, in the related art, a monitoring model is established for each equipment indicator to realize parameter monitoring of the equipment indicator, and the parameter monitoring of each equipment indicator needs to be monitored manually. However, more equipment indicators will lead to more data in the monitoring model. In order to save manpower and material resources, only some important parameters are managed, resulting in a smaller monitoring range and the monitoring accuracy needs to be improved.

[0098] In order to overcome the above problems, this technical solution uses the factory data collection (Factory Data Collection, referred to as: FDC) system data as the big data basis for training the above-mentioned equipment indicator excellence grade prediction model. Exemplarily, the FDC system is a system for real-time monitoring and analysis of production equipment in the product production process data (PROCESS data), which can reduce the scrap rate and production time and improve product quality and output through real-time monitoring and analysis of data. Thus, various indicators and parameters of the equipment are collected. This solution determines the samples for training the above-mentioned classification model based on multiple equipment indicators and their historical parameter values, which can effectively expand the monitoring scope and improve the monitoring accuracy.

[0099] In an exemplary embodiment, Figure 2 The illustrated embodiment provides a specific implementation method of obtaining the historical parameter value of each device indicator in the above step S110, including:

[0100] Step S210, loading source data about equipment indicators into a data warehouse (eg, Hive) for data mining analysis; and, step S220, storing the data after the mining analysis into a non-relational database (eg, Hbase) for determining the historical parameter values.

[0101] Specifically, the source data of the above-mentioned equipment indicators are derived from the state variable (SV) data in the FDC system. Among them, the SV data is used to represent the real-time state parameters of the equipment indicators (such as the current temperature is 80 degrees Celsius). In this embodiment, the BLOB (BinaryLargeObject) field in the SV data can be parsed to generate visual data. Among them, visual data refers to conveying data to users visually by means of graphical means, so as to convey and communicate information more clearly and effectively.

[0102] In this embodiment, the scattered, scattered, and non-standard SV data (which may include the above-mentioned visualization data) can also be integrated together through ETL (Extraction-Transformation-Loading, which is the process of data extraction (Extract), transformation (Transform), and loading (Load), and imported into the data warehouse (such as HIVE) to provide an analysis basis for further data decision-making. Therefore, the massive structured full data can be stored, calculated, mined and analyzed in the data warehouse.

[0103] Furthermore, the data in the data warehouse HIVE is stored in a non-relational database (such as Hbase) through ETL, so as to store massive semi-structured data through Hbase and realize data retrieval and real-time analysis. Thus, the historical parameter values ​​of each of the above equipment indicators can be determined respectively.

[0104] In an exemplary embodiment, in order to improve the prediction accuracy of the trained model, data cleaning is also performed when determining the sample set. Specifically:

[0105] Obtain a device index to be cleaned whose parameter value is a null value in a data object; obtain a mode value or an average value of an attribute value of the device index to be cleaned in other data objects; and determine the mode value or the average value as the attribute value of the device index to be cleaned in the data object.

[0106] Exemplarily, the data stored in the non-relational database Hbase is cleaned. The parameter value is null as the device indicator to be cleaned. Specifically, if the null value is numeric, the missing attribute value is filled according to the average value of the attribute in all other objects; if the null value is numeric, the missing value is filled with the value with the largest number of values ​​of the attribute in all other objects according to the principle of majority in statistics, so as to supplement the missing parameter value with the value with the highest probability, which is conducive to improving the prediction accuracy of the trained model.

[0107] Continue to refer Figure 1 In step S120, according to the correlation relationship between the multiple equipment indicators, the above historical parameter values ​​are subjected to dimensionality reduction processing to obtain sample data about the dimensionality reduction variables.

[0108] In an exemplary embodiment, in the related art, a monitoring model is established for each device indicator to realize parameter monitoring of the device indicator, and thus the parameter monitoring of each device indicator needs to be monitored separately. Since a monitoring model is established for each device indicator to realize parameter monitoring of the device indicator, there is a lack of correlation between the various device indicators. When a problem occurs in the target monitoring device, it is difficult to obtain other device parameter information related to the target device parameter, resulting in low monitoring efficiency.

[0109] In order to overcome the above problems, this technical solution performs dimensionality reduction processing on relevant historical parameter values ​​according to the correlation between equipment indicators to obtain sample data on the dimensionality reduction variables.

[0110] Specifically, Figure 3 The flowchart of a dimensionality reduction processing method provided in the embodiment of the present disclosure can be used as a specific implementation of step S120. Figure 3 , the method shown in this embodiment includes:

[0111] Step S310, analyzing the correlation between the multiple equipment indicators through a principal component analysis algorithm; Step S320, combining the related equipment indicators based on the correlation to obtain a reduced-dimensionality variable; and, Step S330, determining the sample data of the reduced-dimensionality variable based on the historical parameter values ​​of the equipment indicators involved in the combination operation.

[0112] In an exemplary embodiment, since there is a certain correlation between the equipment indicators, the equipment indicators with the correlation can be taken as a set (i.e., an "equipment indicator set"). A set of equipment indicator sets can also be called a "Recipe". Specifically refers to a set of equipment indicator sets related to completing a certain function of a product in the product manufacturing process; for example, in the OLED manufacturing process, a set of equipment parameters related to the packaging stage (such as temperature, pressure, and electrical properties). At the same time, with the improvement of technology, the parameter values ​​of a set of Recipes are also constantly changing. Then, taking the above-mentioned Recipe related to the packaging stage as an example, before a certain packaging technology improvement, it can be called the first version of the Recipe, and after the packaging technology improvement, it can be called the second version of the Recipe. In this embodiment, by comparing the differences in parameter values ​​in Recipes between different versions, a reference basis is provided for optimizing the Recipe.

[0113] Furthermore, quantitative statistical results are provided for each equipment index and different process steps in a set of recipes, so that the closely related equipment indexes (variables) are reduced in dimension to comprehensive indexes (i.e., new variables with fewer numbers), and the obtained comprehensive indexes are unrelated to each other. Therefore, fewer comprehensive indexes can be used to represent the information contained in a larger number of original equipment indexes.

[0114] Exemplarily, the PCA (Principal Component Analysis algorithm is a method for preprocessing high-dimensional feature data. Dimensionality reduction is to retain the most important features of high-dimensional data and remove unimportant features, so as to achieve the purpose of improving data processing speed. Dimensionality reduction can save us a lot of time and cost within a certain range of information loss) method is used to preprocess the data, map the high-dimensional data to a low-dimensional space, and maximize the variance of the data in the dimension, so as to use fewer data dimensions, retain more characteristics of the original data points, and maximize the retention of the intrinsic information of the data.

[0115] Furthermore, in step S130, a positive sample set and a negative sample set are determined according to the excellence level of the sample data of the above-mentioned dimension reduction variable.

[0116] In an exemplary embodiment, the quality level of the sample data is determined according to the quality level of the product. For example, the quality of the product is divided into A, B, C and D. Accordingly, the sample data related to the manufacture of A-level products can be divided into A-level, the sample data related to the manufacture of B-level products can be divided into B-level, the sample data related to the manufacture of C-level products can be divided into C-level, and the sample data related to the manufacture of D-level products can be divided into D-level. Further, the quality level of the training samples belonging to A-level and B-level in the sample data of the reduced dimension variable can be determined as a positive sample set, and the quality level of the training samples belonging to C-level and D-level in the sample data of the reduced dimension variable can be determined as a negative sample set.

[0117] Exemplarily, the final inspection data of the OLED product is used to determine the positive sample set and the negative sample set. Specifically, if the final inspection data of the OLED product is good, the equipment index parameters for producing the OLED product are used as positive samples to determine the positive sample set; if the final inspection data of the OLED product is defective, the equipment index parameters for producing the OLED product are used as negative samples to determine the negative sample set. It should be noted that in order to ensure the accuracy of the prediction, the positive and negative sample sets are determined for a single model of product. In addition, in order to avoid inaccurate predictions caused by uneven sample distribution, for a single model of product, the positive sample set contains at least 10,000 positive sample information, and the negative sample set contains at least 10,000 negative sample information.

[0118] Furthermore, in step S140, a machine learning model is trained by using the positive sample set and the negative sample set to obtain a classification model for predicting the quality level of the target device indicator.

[0119] In an exemplary embodiment, Figure 4The flowchart of a model training method provided in the embodiment of the present disclosure can be used as a specific implementation of step S140. Figure 4 , the method shown in this embodiment includes:

[0120] Step S410, sampling the reduced-dimensional variables in the positive sample set and the negative sample set respectively to obtain sampling variables; step S420, calculating the eigenvector of the Fourier basis of each of the sampling variables to determine the correlation between the sampling variables; step S430, processing the eigenvector of the Fourier basis based on a linear regression algorithm to obtain screening variables; and, step S440, training a machine learning model based on the screening variables and the parameter values ​​corresponding to the screening variables.

[0121] Exemplarily, the positive sample set is randomly divided into training positive samples and testing positive samples, and the negative sample set is randomly divided into training negative samples and testing negative samples. The embodiment shown in the figure trains a machine learning model by training positive samples and training negative samples. The machine learning model is a Bayesian algorithm model, a decision tree algorithm model or a neural network classification model.

[0122] Exemplarily, in the parameter space corresponding to the training samples, the dimension reduction variables corresponding to the equipment indicators are randomly sampled to obtain sampling variables. Then, the eigenvector of the low-degree Fourier basis is calculated for the parameter value corresponding to each sampling variable to determine the correlation between the sampling variables. Further, the eigenvector of the above Fourier basis is processed based on a linear regression algorithm (such as the lasso algorithm) to obtain a screening variable. After completing a round of screening, the screening variable is a feature that has a greater impact on the product quality level. Based on at least one round of screening, the screening features and their corresponding parameters are used to train the machine learning model. Through Figure 4 The technical solution shown can remove noise variables that do not have a large enough impact on the product quality level, which is beneficial to improving the model training speed.

[0123] In an exemplary embodiment, the technical solution also includes testing a goodness level prediction model (referred to as "model to be tested") by testing positive samples and testing negative samples, and using at least one test indicator to verify the test results of the model to be tested, and predicting the goodness level of the model that meets the test indicator.

[0124] Specifically, the model performance is evaluated by one or more of the following evaluation indicators, including: accuracy, recall, KS (Kolmogorov-Smirnov) value and receiver operating characteristic curve (Receiver Operating Characteristic curve, abbreviated as: ROC) area under AUC (a model evaluation indicator, specifically used to evaluate the predictive value of the model; it is the abbreviation of Area Under ROC Curve) to evaluate the above iteratively optimized prediction model.

[0125] In an exemplary embodiment, the method for testing the model to be tested specifically includes:

[0126] First, the descriptive features of the test samples are input into the model to be tested, and the output data of the model are as follows: true positive TP, true negative TN, false negative FN and false positive FP. Among them, TP is the number of samples that are still in the positive class after being judged as the positive class in the test sample set by the model to be tested, TN is the number of samples that are still in the negative class after being judged as the negative class in the test sample set by the model to be tested, FN is the number of samples that are still in the negative class after being judged as the positive class in the test sample set by the model to be tested, and FP is the number of samples that are still in the positive class after being judged as the negative class in the test sample set by the model to be tested. Positive and negative classes refer to the two categories manually annotated for the samples in the first part, that is, if a sample is manually annotated to belong to a specific class, then the sample belongs to the positive class, and samples that do not belong to the specific class belong to the negative class.

[0127] Secondly, the test results of the model to be tested are calculated according to the true positive TP, true negative TN, false negative FN and false positive FP.

[0128] In an exemplary embodiment, the test index may be an AUC or a KS value. Specifically:

[0129] In an exemplary embodiment, the false positive rate FPR and the true positive rate TPR are determined using Formula 1 and Formula 2:

[0130] FPR=FP / (FP+TN) Formula 1

[0131] TPR=TP / (TP+FN) Formula 2

[0132] Furthermore, the ROC curve is plotted with FPR as the horizontal axis and TPR as the vertical axis. The ROC curve is the characteristic curve of each indicator obtained, which is used to show the relationship between each indicator, and further calculates the area under the ROC curve AUC. The ROC curve is the characteristic curve of each indicator obtained, which is used to show the relationship between each indicator. AUC is the area under the ROC curve. The larger the AUC, the higher the predictive value of the model, and the AUC can be used to test the model to be tested. And when the evaluation result is that the AUC value meets the preset threshold, the obtained model can be used to predict the quality level.

[0133] The KS (each time we select a different threshold, we can get a set of FPR and TPR, that is, a point on the ROC curve) curve is the difference between the two curves at each threshold. KS = max (TPR-FPR), which is the maximum value of the difference between TPR and FPR; the KS value can reflect the optimal distinction effect of the model. At this time, the threshold taken is generally used as the optimal threshold for defining good and bad users. The value range of the KS value is [0, 1], and the larger the KS value, the higher the prediction accuracy of the model. For example, KS>0.2 can be considered that the model has relatively good prediction accuracy.

[0134] In an exemplary embodiment, the model test index may also be accuracy and recall. Specifically, the accuracy p and the recall r are calculated according to Formula 3 and Formula 4 respectively;

[0135] p=TP / (TP+FP) Formula 3

[0136] r=TP / (TP+FN) Formula 4

[0137] If the setting conditions corresponding to the test indicators are: if the accuracy test result is greater than p' (preset value), the accuracy setting conditions are met, otherwise the accuracy setting conditions are not met, and if the recall test result is greater than r' (preset value), the recall setting conditions are met, otherwise the recall setting conditions are not met.

[0138] In an exemplary embodiment, when the test results meet the set conditions corresponding to the test indicators, the model to be tested can be used as a prediction model for predicting the parameter values ​​of the equipment indicators; when the test results do not meet the set conditions, the above-mentioned model to be tested continues to be iteratively optimized until the test results of the model to be tested meet the set conditions.

[0139] In a second aspect, an embodiment of the present disclosure provides a system for monitoring equipment indicators.

[0140] Since the technical solution involves multiple factory equipment in multiple factories, the amount of raw data about equipment indicators and parameter values ​​is very large. For example, the raw data generated by all factory equipment every day may be hundreds of GB, and the data generated every hour may be tens of GB. There are two main solutions for storing and computing massive structured data: the grid computing solution of RDBMS (Relational Database Management System, RDBMS); the big data solution of Distributed File System (Distributed File System, DFS).

[0141] Among them, the grid computing of RDBMS is to divide the problem that requires a very huge computing power into many small parts, and then distribute these parts to many computers for processing respectively, and finally integrate these calculation results. For example, as a specific example, Oracle RAC (Real Application Cluster) is the core technology of grid computing supported by Oracle database, in which all servers can directly access all data in the database. However, the application system of RDBMS grid computing cannot meet the user's requirements when the amount of data is large. For example, due to the limited expansion space of the hardware, when the data increases to a large enough order of magnitude, the efficiency of processing data will be very low due to the bottleneck of hard disk input / output.

[0142] Big data technology based on distributed file management allows the use of multiple cheap hardware devices to build large clusters to process massive data. For example, the Hive tool is a data warehouse tool based on Hadoop, which can be used for data extraction, transformation and loading (ETL). The Hive tool defines a simple SQL-like query language, and also allows the use of custom MapReduce mappers and reducers to perform complex analysis tasks that the default tool cannot complete. The Hive tool does not have a special data storage format, nor does it create indexes for data. Users can freely organize tables and process data in non-relational databases. It can be seen that the parallel processing of distributed file management can meet the storage and processing requirements of massive data. Users can process simple data through SQL queries, and custom functions can be used for complex processing. Therefore, when analyzing massive data in a factory, it is necessary to extract data from the factory database into a distributed file system (such as the distributed storage device 501 / 701 described below), specifically:

[0143] Through ETL, the scattered, messy, and non-standard SV data are integrated together and imported into the data warehouse HIVE to provide an analytical basis for further data decision-making. Therefore, the massive structured full data can be stored, calculated, mined and analyzed in the data warehouse HIVE. Furthermore, through ETL, the data in the data warehouse HIVE is stored in the non-relational database Hbase, so that Hbase can be used to store massive semi-structured data, and realize data retrieval and real-time data analysis.

[0144] Therefore, on the one hand, the original data will not be damaged, and on the other hand, the efficiency of data analysis is improved.

[0145] Reference Figure 5 The device indicator monitoring system 500 provided in the embodiment of the present disclosure includes a distributed storage device 501 and one or more processors 502 .

[0146] The distributed storage device 501 is configured to obtain first parameter values ​​to be measured for M target device indicators, respectively, where M is an integer greater than 1; and store production data generated by factory equipment. The one or more processors 502 are configured to monitor the excellence of the M target device indicators based on the classification model provided in the above embodiment. Perform an operation of determining correlation.

[0147] The distributed storage device 501 stores production data from factory equipment. Factory equipment refers to any equipment in each factory, which may include process equipment in each process site, and management equipment used to manage production lines in the factory, etc. Production data refers to any information related to production, including which products each process equipment in each production line participates in the production process, and whether each product is defective and what kind of defect it is.

[0148] Reference Figure 6 , the one or more processors (such as CPU) 502 may be in the analysis device 600. Exemplarily, the analysis device 600 further includes a memory (such as a hard disk) 601 that may store the required program, and the processor 502 is connected to the memory 601 through an I / O interface 602, so that information interaction can be achieved, so that the processor 502 can perform the required operation according to the program stored in the memory 601 to achieve the operation of determining the correlation.

[0149] Among them, the distributed storage device 501 stores relatively complete data (such as a non-relational database), and the distributed storage device 501 includes multiple hardware memories, and different hardware memories are distributed in different physical locations (such as in different factories, or in different production lines), and information is transmitted between each other through the network, so that its data is distributed relationally, but logically constitutes a database based on big data technology.

[0150] Reference Figure 7 The monitoring system 700 shown includes: a distributed storage device 701 , a source system 702 , an analysis device 703 including one or more processors, and a display device 704 .

[0151] Among them, the source system 702 contains a large amount of raw data of different factory equipment (such as parameter values ​​about temperature, parameter values ​​about pressure, parameter values ​​about electrical properties, etc.). Exemplarily, the above raw numbers can be processed by PLC (Programmable Logic Controller). Further, the processed data is stored in CIM PC (a machine storing equipment related information), and then EIS (Equipment international System) classifies and transmits the required data from CIM PC to further store in multiple production and manufacturing systems, such as YMS (Yield Management System) to store yield related data, FDC system, MES (Manufacturing Execution System), MDW (Manufactory Data Warehouse, Manufacturing Data Warehouse) and other systems in relational databases (such as Oracle, Mysql, etc.). Further, the above data is stored in the HADOOP big data platform by ETL to perform relevant calculations in HADOOP. For example, the data stored in the HADOOP big data platform can be displayed in corresponding reports using the BO (BUSINESS OBJECT) tool.

[0152] Furthermore, in order to facilitate the subsequent data reading of the analysis device 703 and reduce the load on the factory equipment and the manufacturing system, OGG (Oracle Golden Gate, a real-time data transmission tool of ORACLE) can be used to transmit the data of the source system to the data lake 71 of the distributed storage device (such as Hadoop Distributed File System, HDFS) 701 in real time.

[0153] The data in the distributed storage device 701 can be stored in a data warehouse (such as Hive) or a non-relational database (such as Hbase) format. Exemplarily, the data from the source system 702 is first stored in the data lake; afterwards, data cleaning, data conversion and other pre-processing can be continued in Hive according to the application theme, scenario, etc. of the data. Exemplarily, in the data warehouse 72, data with different themes (such as production history theme, detection data theme, point measurement theme, equipment data theme) are obtained through ETL processing and other methods. Further, data with different scenarios (such as sudden failure analysis scenario, correlation analysis scenario) are stored in the data mart 73. Exemplarily, the data in the data mart 73 can be specifically stored in the non-relational database Hbase. The above data mart 73 can be connected to the analysis device 703 and the like through an API interface to realize data interaction with these devices.

[0154] For example, the data mart 73 transmits the parameter data corresponding to the collected device indicators to one or more processors in the analysis device 703 through the API interface for data processing (for specific implementation, see Figure 2 and Figure 3 The sample set is obtained in the embodiment shown in the figure; further, in one or more processors in the analysis device 703, the model is trained by the training positive samples and the training negative samples in the sample set (for a specific implementation, see Figure 4 In addition, in one or more processors in the analysis device 703, the model evaluation is performed using the test positive samples and the test negative samples in the sample set (for specific implementation, please refer to the above-mentioned model test embodiment).

[0155] In an exemplary embodiment, given that the training samples of the above-mentioned prediction model are trained based on reduced-dimensional variables obtained by performing dimensionality reduction processing on multiple device indicators, when predicting the quality level of the device indicators through the prediction model, the target device indicators (i.e., the indicators of the device to be tested) obtained in the distributed storage device 701 need to be transmitted to the analysis device 703, so that data processing (i.e., dimensionality reduction processing) can be performed on the target device indicators through one or more processors.

[0156] Specifically, in this embodiment, the parameter values ​​corresponding to the M target device indicators are recorded as "first parameter values ​​to be measured", and the parameter values ​​corresponding to the N target reduced dimension variables obtained after the M target device indicators are subjected to dimensionality reduction processing are recorded as "second parameter values ​​to be measured". Then one or more processors in the analysis device 703 are further configured to: perform dimensionality reduction processing on the first parameter values ​​to be measured according to the association relationship between the M target device indicators to obtain second parameter values ​​to be measured for the N target reduced dimension variables, where N is a positive integer less than M; and input the second parameter values ​​to be measured into the prediction model to monitor the excellence of the M target device indicators through the output values ​​of the prediction model.

[0157] It can be seen that this technical solution performs dimensionality reduction processing according to the correlation between the target equipment indicators, and uses a smaller number of comprehensive indicators to represent the information contained in a larger number of original equipment indicators.

[0158] The process of one or more processors in the analysis device 703 performing dimensionality reduction processing on the above-mentioned M target device indicators includes: analyzing the correlation relationship between the M target device indicators through a principal component analysis algorithm; combining the associated device indicators based on the correlation relationship to obtain N target dimensionality reduction variables; and determining the second parameter to be measured of the target dimensionality reduction variable based on the first parameter to be measured of the target device indicator involved in the combination operation.

[0159] It should be noted that, since one or more processors in the analysis device 703 perform the above-mentioned dimensionality reduction process and Figure 3 The embodiments shown are the same, so they will not be described again here.

[0160] refer to Figure 7 , the display device 704 can be used to display an "interactive interface", which may include a sub-interface for displaying analysis results (such as the quality level), a sub-interface for controlling the system for product defect cause analysis to perform required work (such as task setting), and a sub-interface for controlling each process equipment (such as setting its process parameters), etc. In other words, through the "interactive interface" of the display device, the user can fully interact with the system for product defect cause analysis (control and receive results). Exemplarily, the user can provide relevant parameters for algorithm design in the analysis device 703 through the interactive interface, and the user can also select a model for predicting the quality level of the equipment index (i.e., model selection) through the interactive interface, etc.

[0161] In an exemplary embodiment, the process of predicting the quality level of the above-mentioned target dimensionality reduction variables and their corresponding second parameter values ​​to be measured by the trained prediction model is as follows, that is, one or more processors in the analysis device 703 are specifically configured to: input the second parameter values ​​to be measured regarding the N target dimensionality reduction variables into the prediction model, and respectively calculate the eigenvectors of the Fourier basis corresponding to the N target dimensionality reduction variables; process the eigenvectors of the N Fourier basis based on a linear regression algorithm to obtain N' target screening variables regarding the quality level, wherein N' is less than N; obtain the predicted probability that the parameter values ​​of the target screening variables belong to each quality level based on the prediction model; and determine that the quality level corresponding to the maximum value of the predicted probability is the output value of the prediction model.

[0162] In an exemplary embodiment, a specific implementation method for determining a fault indicator set is as follows: If the output value of the prediction model is that the excellence level is lower than a preset level value, then one or more processors in the analysis device 703 are further configured to: By performing a dimensionality increase processing corresponding to the dimensionality reduction processing on the target screening variable, a device indicator set related to the target screening variable (such as RecipeX) can be obtained. Further, a group of device indicators included in the device indicator set RecipeX is determined as a faulty device indicator set. Furthermore, a correlation analysis is performed on the group of device indicators included in RecipeX (i.e., Figure 7 The correlation analysis subject shown) is used to determine the cause of the defect. Exemplarily, if the quality grades of the product are divided into Grade A, Grade B, Grade C and Grade D, Grade B can be taken as the above-mentioned preset grade value. Thus, if the output value of the prediction model is Grade C and Grade D, the equipment indicator set M related to the above-mentioned target screening variable is obtained to determine that the equipment indicator set M is a faulty equipment indicator set. Of course, the above-mentioned preset grade value can also be Grade C, then the faulty equipment indicator set is determined when the output value of the prediction model is Grade D. In other words, the above-mentioned faulty equipment indicator set is related to the pre-divided product quality grades and the predicted values ​​of the model.

[0163] In an exemplary embodiment, after determining the faulty device indicator set, the good parameter ranges corresponding to the respective device indicators in the indicator set are further determined. Specifically, since the current parameter value of at least one device indicator in the faulty device indicator set is outside the corresponding good parameter range, the current parameter value of the faulty device indicator can be adjusted according to the good parameter range, so that the indicator value of the faulty device is restored to the corresponding good parameter range in time, thereby reducing the defective rate of the product.

[0164] It should be noted that since one or more processors in the analysis device 703 perform the above data processing process and Figure 4The embodiments shown are the same, so they will not be described again here.

[0165] In an exemplary embodiment, one or more processors in the analysis device 703 are further configured to: determine a good parameter value interval of the target device indicator to be monitored based on the prediction model. Among them, any one of the M target device indicators is used as the target device indicator to be monitored in this embodiment. Specifically, the good parameter value interval of the target device indicator to be monitored is determined according to the trained prediction model. Figure 8 , the horizontal axis represents M target device indicators, and the vertical axis is used to represent the good parameter value interval of each target device indicator. Therefore, the current parameter value of the above-mentioned target device indicator to be monitored can be monitored in real time through the good parameter value interval.

[0166] refer to Figure 8 In an exemplary embodiment, for the target device indicator Q1 to be monitored, the parameter value of Q1 is obtained in real time. When the parameter value of the target device indicator Q1 to be monitored is within the corresponding excellent parameter value range, it means that the monitoring result of the above Q1 is good, and it also means that the current parameter value of Q1 will not lead to the production of defective products. In this embodiment, the current parameter value of Q1 is exemplarily displayed as a dot. In addition, the situation of another target device indicator Q2 is the same as the above target device indicator Q1 to be monitored, and the target device indicator Q2 will not be described in detail. In another exemplary embodiment, for the target device indicator Q1 to be monitored M , get Q in real time M When the parameter value of the target device indicator QM to be monitored is outside the corresponding excellent parameter value range, it indicates that the above Q M The monitoring result is bad, which also means that Q M The current parameter value of Q may result in defective products. M The current parameter value is displayed as a triangle to remind staff to take improvement measures in time.

[0167] It can be seen that this technical solution can provide a history of abnormal parameters throughout the life cycle through big data prediction, which can specifically confirm the impact of equipment parameters on products. In addition, real-time monitoring is achieved through the trend line of variable parameters, which is conducive to timely determining parameter value bad warnings and reducing product defect rates.

[0168] refer to Figure 7 The display device 704 has a display function for displaying the analysis results (such as correlation) calculated by the analysis device.

[0169] In some embodiments, the display device also has a display function, which is used to display the analysis results (such as the excellence level, etc.) calculated by the analysis device. The display device may include one or more displays, including one or more terminals with a display function, so that the analysis device can send the correlation obtained by its analysis to the display device, and the display device then displays it.

[0170] In this embodiment, the display interface of the display device 704 is as described above. Figure 8 The display shown is as follows: specifically, the good parameter value interval of the above-mentioned target device indicator to be monitored is displayed, wherein the good parameter value interval refers to the parameter value range of the target device indicator to be monitored that will not lead to the production of defective products; and the display interface also displays the current parameter value of the target device indicator to be monitored acquired in real time, wherein when the current parameter value is within the corresponding good parameter value interval, it is displayed as a dot (which can be changed according to actual needs, for example, it can be displayed as a red dot), and when the current parameter value is outside the corresponding good parameter value interval, it is displayed as a triangular dot (which can be changed according to actual needs, for example, it can be displayed as a green dot). Thus, the good parameter value interval and current parameter value of any target device indicator are vividly displayed to the staff, so that the staff can monitor the above-mentioned first-dimensional variables and take corresponding measures in time for bad warnings.

[0171] In an exemplary embodiment, the monitoring system 700 further includes an alarm device. Specifically, the alarm device is configured to sound an alarm (such as a buzzer or a voice alarm light) when the difference between the current parameter value of the target device indicator to be monitored and the boundary value of the excellent parameter value interval is less than a preset value, so as to remind the staff of the occurrence of a bad warning, thereby facilitating the staff to take corresponding measures for the bad warning in a timely manner.

[0172] In the embodiments provided by the present disclosure, firstly, the FDC system data is used as the big data basis for the prediction and analysis of the product quality level. Further, the distributed storage device can efficiently realize the collection and preliminary processing of the original data of multiple factory equipment through big data. One or more processors (which can be arranged in the analysis device) can obtain the required data from the distributed storage device. Based on the trained prediction model, the excellent parameter value intervals of the M target equipment indicators can be determined to monitor each target equipment indicator in real time, predict the possibility of defective occurrence and alarm function, and the defective prediction can discover some important parameters that have not been paid attention to before by managing all parameters to the FDC system, improve the control range of the FDC system, and then more accurately locate the defective, adjust the production process, etc. In addition, the defective prediction can automatically generate a model and provide it to the FDC system to help the FDC system reasonably confirm the range of relevant equipment indicators, thereby saving manual time and being able to discover problems as early as possible. When ensuring the production of similar products, the equipment status is the optimal production status, and the yield of similar products is improved in batches.

[0173] refer to Fig. 9 As shown, the technical solution also provides a computer readable medium 900, which can be a portable compact disk read-only memory (CD-ROM) and includes program code, and can be run on a terminal device, such as a personal computer. However, the program product of the present disclosure is not limited thereto. In this document, a readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, apparatus, or device.

[0174] The above-mentioned program product can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RandomAccess Memory, RAM), a read-only memory (Read-Only Memory, ROM), an erasable programmable read-only memory (erasable programmable read-only memory, EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0175] Computer readable signal media may include data signals propagated in baseband or as part of a carrier wave, in which readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Readable signal media may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0176] The program code embodied on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the foregoing.

[0177] Program code for performing the operations of the present disclosure may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computing device (e.g., via the Internet using an Internet Service Provider).

[0178] In addition, in an exemplary embodiment of the present disclosure, an electronic device capable of implementing the above method is also provided.

[0179] Those skilled in the art will appreciate that various aspects of the present disclosure may be implemented as systems, methods or program products. Therefore, various aspects of the present disclosure may be specifically implemented in the following forms, namely: complete hardware implementation, complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software, which may be collectively referred to herein as "circuits", "modules" or "systems".

[0180] Refer to the following Fig.10 The electronic device 1000 according to this embodiment of the present disclosure is described. Fig.10 The electronic device 1000 shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0181] like Fig.10As shown, the electronic device 1000 is in the form of a general computing device. The components of the electronic device 1000 may include but are not limited to: the at least one processing unit 1010, the at least one storage unit 1020, and a bus 1030 connecting different system components (including the storage unit 1020 and the processing unit 1010).

[0182] The storage unit stores program codes, and the program codes can be executed by the processing unit 1010, so that the processing unit 1010 executes the method steps described in the above embodiments of this specification.

[0183] The storage unit 1020 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 10201 and / or a cache memory unit 10202 , and may further include a read-only memory unit (ROM) 10203 .

[0184] The storage unit 1020 may also include a program / utility 10204 having a set (at least one) of program modules 10205, such program modules 10205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0185] Bus 1030 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0186] The electronic device 1000 may also communicate with one or more external devices 1100 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 1000, and / or communicate with any device that enables the electronic device 1000 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 1050. Further, the I / O interface 1050 is connected to the display unit 1040 to transmit the content to be displayed to the display unit 1040 via the I / O interface 1050 for the user to view.

[0187] Furthermore, the electronic device 1000 can also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN) and / or a public network, such as the Internet) through a network adapter 1060. As shown in the figure, the network adapter 1060 communicates with other modules of the electronic device 1000 through the bus 1030. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with the electronic device 1000, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0188] Through the description of the above implementation, it is easy for those skilled in the art to understand that the example implementation described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the implementation of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the implementation of the present disclosure.

[0189] In addition, the above-mentioned figures are only schematic illustrations of the processes included in the method according to the exemplary embodiments of the present disclosure, and are not intended to be limiting. It is easy to understand that the processes shown in the above-mentioned figures do not indicate or limit the time sequence of these processes. In addition, it is also easy to understand that these processes can be performed synchronously or asynchronously, for example, in multiple modules.

[0190] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing what is disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary technical means in the art that are not disclosed in the present disclosure. The specification and embodiments are to be considered merely as exemplary, and the true scope and spirit of the present disclosure are indicated by the claims.

Claims

1. A training method for a device indicator quality level prediction model, wherein: The method comprises: Obtaining historical parameter values ​​of multiple equipment indicators; the equipment indicators include one or more of temperature, pressure, humidity, heating time, and cooling time related to product quality; Analyze the correlation between the multiple equipment indicators by using a principal component analysis algorithm; combine the associated equipment indicators based on the correlation to obtain a reduced-dimensionality variable; and determine sample data of the reduced-dimensionality variable based on historical parameter values ​​of the equipment indicators involved in the combination operation; Determining a positive sample set and a negative sample set according to the excellence level of the sample data of the dimension reduction variable; The reduced-dimensional variables in the positive sample set and the negative sample set are sampled respectively to obtain sampling variables; the eigenvector of the Fourier basis of each of the sampling variables is calculated to determine the correlation between the sampling variables; the eigenvector of the Fourier basis is processed based on a linear regression algorithm to obtain screening variables; a machine learning model is trained based on the screening variables and the parameter values ​​corresponding to the screening variables to obtain a prediction model for predicting the quality level of the target device indicator.

2. The method according to claim 1, wherein: The sample data includes training samples and test samples; The step of determining a positive sample set and a negative sample set according to the excellence level of the sample data of the dimension reduction variable comprises: Determining a positive sample set and a negative sample set according to the excellence level of the training samples in the sample data of the dimension reduction variable; The method further includes: evaluating the prediction model using the test sample.

3. The method according to claim 1, wherein: The prediction model is evaluated by one or more of the following evaluation indicators, including: accuracy, recall, and area under the receiver operating characteristic curve (AUC).

4. The method according to claim 1, wherein: The obtaining of the historical parameter value of each device indicator includes: Loading source data about equipment indicators into the data warehouse for data mining analysis; The data after the mining analysis is stored in a non-relational database for determining the historical parameter value.

5. The method according to claim 4, wherein: The method further comprises: Get the device index to be cleaned up whose parameter value is null in the data object; Obtain the mode value or average value of the attribute value of the device indicator to be cleaned in other data objects; The mode value or the average value is determined as the attribute value of the device indicator to be cleaned in the data object.

6. The method according to claim 1, wherein: The machine learning model is a Bayesian algorithm model, a decision tree algorithm model or a neural network model.

7. A monitoring system for equipment indicators, wherein: The system includes: a distributed storage device, and one or more processors, wherein: The distributed storage device is configured to respectively obtain first parameter values ​​to be measured for M target device indicators, where M is an integer greater than 1; The one or more processors are configured to monitor the quality of the M target device indicators based on the prediction model described in any one of claims 1 to 6.

8. The system according to claim 7, wherein: The one or more processors are further specifically configured to: According to the association relationship between the M target device indicators, the first parameter value to be measured is subjected to dimensionality reduction processing to obtain the second parameter value to be measured related to N target dimensionality reduction variables, wherein N is a positive integer less than M; The second parameter value to be measured is input into the prediction model to monitor the quality of the M target device indicators through the output value of the prediction model.

9. The system according to claim 8, wherein: The one or more processors are specifically configured to: Analyze the correlation between the M target device indicators by using a principal component analysis algorithm; Based on the association relationship, the associated equipment indicators are combined to obtain N target dimension reduction variables; According to the first parameter to be measured of the target device indicator involved in the combination operation, a second parameter to be measured of the target dimension reduction variable is determined.

10. The system according to claim 9, wherein: The one or more processors are specifically configured to: Inputting the second parameter values ​​to be measured about the N target dimension reduction variables into the prediction model, and respectively calculating the eigenvectors of the Fourier basis corresponding to the N target dimension reduction variables; Processing the eigenvectors of the N Fourier bases based on a linear regression algorithm to obtain N' target screening variables related to the excellence level, wherein N' is less than N; Based on the prediction model, the predicted probability that the parameter value of the target screening variable belongs to each excellence level is obtained; The goodness level corresponding to the maximum value of the predicted probability is determined as the output value of the prediction model.

11. The system according to claim 10, wherein: The one or more processors are further configured to: If the output value of the prediction model is that the excellence level is lower than the preset level value, the target dimension reduction variable is processed by dimension increase to obtain a set of faulty equipment indicators.

12. The system according to claim 7, wherein: The one or more processors are further configured to: Determining an excellent parameter value interval of a target device indicator to be monitored based on the prediction model, wherein the target device indicator to be monitored is any one of the M target device indicators; The current parameter value of the target device indicator to be monitored is monitored in real time according to the excellent parameter value interval.

13. The system according to claim 12, wherein: The system further includes: a display device configured to: The good parameter value range of the target device indicator to be monitored is displayed, and the current parameter value of the target device indicator to be monitored obtained in real time is displayed.

14. The system according to claim 12, wherein: The system further comprises: an alarm device configured to: When the difference between the current parameter value of the target device indicator to be monitored and the boundary value of the good parameter value interval is less than a preset value, an alarm is issued.

15. A method for monitoring equipment indicators, wherein: The method comprises: Obtaining first parameter values ​​to be measured for M target device indicators respectively, where M is an integer greater than 1; The quality of the M target device indicators is monitored based on the prediction model described in any one of claims 1 to 6.

16. The method according to claim 15, wherein: The monitoring of the excellence of the M target device indicators based on the prediction model includes: According to the association relationship between the M target device indicators, the first parameter value to be measured is subjected to dimensionality reduction processing to obtain the second parameter value to be measured related to N target dimensionality reduction variables, wherein N is a positive integer less than M; The second parameter value to be measured is input into the prediction model to monitor the quality of the M target device indicators through the output value of the prediction model.

17. The method according to claim 16, wherein: The performing dimensionality reduction processing on the first parameter value to be measured according to the association relationship between the M target device indicators includes: Analyze the correlation between the M target device indicators by using a principal component analysis algorithm; Based on the association relationship, the associated equipment indicators are combined to obtain N target dimension reduction variables; According to the first parameter to be measured of the target device indicator involved in the combination operation, a second parameter to be measured of the target dimension reduction variable is determined.

18. A computer readable medium having a computer program stored thereon, wherein: When the program is executed by the processor, it implements the training method of the equipment indicator excellence grade prediction model as described in any one of claims 1 to 6, or implements the equipment indicator monitoring method as described in any one of claims 15 to 17.

Images