CN109255186A - A kind of industrial process flexible measurement method based on output constraint AP-XGBOOST model - Google Patents
A kind of industrial process flexible measurement method based on output constraint AP-XGBOOST model Download PDFInfo
- Publication number
- CN109255186A CN109255186A CN201811060031.0A CN201811060031A CN109255186A CN 109255186 A CN109255186 A CN 109255186A CN 201811060031 A CN201811060031 A CN 201811060031A CN 109255186 A CN109255186 A CN 109255186A
- Authority
- CN
- China
- Prior art keywords
- matrix
- sample
- xgboost
- output constraint
- measurement method
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Geometry (AREA)
- Computer Hardware Design (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention discloses a kind of industrial process flexible measurement method based on output constraint AP-XGBOOST model, this method passes through partial least squares algorithm first, original variable space projection to latent variables space is subjected to measuring similarity, then classified using affine propagation clustering algorithm to multiple operating modes process, soft sensor modeling of the different limit grad enhancement tree-models for multi-state complex process is finally established on obtained data set.Compared to the soft-sensing model for combining traditional affine propagation clustering algorithm, affine propagation clustering based on output constraint-limit grad enhancement tree AP-XGBOOST model can effectively improve the accuracy to multiple operating modes process cluster, and it being capable of variable relation in accurate mapping complex production process, it is not limited by data distribution hypothesis, the size of data set, is provided simultaneously with good interpretation and robustness.
Description
Technical field
The invention belongs to industrial process control fields, and in particular to a kind of work based on output constraint AP-XGBOOST model
Industry process flexible measurement method.
Background technique
With the increase of industrial process complexity, the appearance of the features such as multi-modal, multi-state would generally make hard measurement mould
Type predictive ability is degenerated.For this problem, multi-model Modeling Method is suggested: passing through clustering algorithm, gauss hybrid models etc.
Method classifies to multi-state data sample, then using machine learning such as statistical method, support vector regression, neural networks
The sample set that method is directed under different operating conditions establishes hard measurement submodel respectively.However there are several places' deficiencies for such method: 1, gathering
Class result is difficult to accurately reflect the multi-state classification of process.2, traditional clustering method such as K mean cluster method needs artificial in advance
Determine class number, it is difficult to accurately divide data set according to different operating conditions;Though affine propagation clustering has certain advantage,
Since influence degree of the different auxiliary variables for target variable is different, the selection of sample still has certain limitation.3,
Hard measurement submodel is difficult to accurate characterization complex process, and the propositions such as LI Xiu-liang are using support vector regression as submodule
Type, Jie Yu etc. propose Gauss-Markov dynamic fusion submodel approach, traditional statistics, probabilistic model and machine in normal service
The contradiction that model is difficult to balance model accuracy and complexity during multi-model modeling is practised, the generalization ability of model is difficult to
To guarantee.
Summary of the invention
It is an object of the invention to solve to be generally in different operation operating conditions due to complex process so as to cause soft
The problems such as measurement model is degenerated, using a kind of limit grad enhancement tree (AP- based on output constraint affine propagation clustering algorithm
XGBOOST) the industrial process flexible measurement method of model, compared to traditional clustering algorithm, the radiation based on output constraint is propagated
Algorithm can more accurately distinguish the different operating stages of process, and effectively be promoted by establishing XGBOOST submodel
To the prediction effect of complex nonlinear nongausian process.
The purpose of the present invention is achieved through the following technical solutions:
A kind of industrial process flexible measurement method based on output constraint AP-XGBOOST model, which is characterized in that including with
Lower step:
(1) the input auxiliary variable matrix that given historical sample integrates is X ∈ RL×N, target variable matrix Y ∈ RL;Wherein L and
N respectively represents sample number and input variable dimension, by Partial Least Squares to auxiliary variable matrix X and target variable matrix Y
Supervision dimensionality reduction: X=TP is carried outT+ E, Y=TRT+F;Wherein, matrix T is that historical sample matrix is projected by Partial Least Squares
Obtained new sample matrix, P are the loading matrix of historical sample matrix, and R is the loading matrix of target variable matrix Y, E and F
Corresponding residual matrix when respectively with Partial Least Squares fitting X and Y;
(2) by sample matrix T calculate historical sample concentrate data point between similarity matrix S:s (i, j)=- | |
ti-tj||2, wherein s (i, j) is the element in matrix S, the element t in representative sample matrix TiAnd tjBetween similarity it is big
It is small;Carry out iteration according to similarity matrix S and updates Certainty Factor matrix and evidence availability matrix
Wherein, r (i, k), r (k, k), r (j, k) are the element of Certainty Factor matrix, and a (i, k), a (j, i) are respectively
The element of evidence availability matrix;
(3) when Certainty Factor matrix and evidence availability matrix converges on fixed value or the number of iterations reaches maximum value
When, calculate each cluster centre point I=1 ..., θ, and determine cluster centre number θ
With corresponding Sub Data Set(m=1... θ);
(4) XGBOOST modeling algorithm is called to construct multiple submodel H (m) on Sub Data Set, (m=1... θ);When new
Sample xqWhen arrival, according to formula tq=xqP calculates sample xqVariable t after being projected by Partial Least Squaresq;
(5) t is calculatedqFrom different cluster centre pointsThe distance between, and determine therefrom that the cluster nearest apart from the sample
CenterAnd by tqThe XGBOOST submodel H (m) constructed in the step (4) corresponding with the cluster centre is inputted, is calculated final defeated
It is worth out, as model is directed to original sample xqPredicted value.
Further, the step (1) specifically: the latent variables space dimension of the Partial Least Squares is K, is led to
Least square method is crossed to be expressed as the input and output of history data set
Further, in the step (2) when calculating Certainty Factor matrix and availability matrix, Certainty Factor square
The value of element a (i, k) initially selected in element r (i, k) and availability matrix initially selected is set as 0 in battle array.
Further, when updating element a (i, k) of availability matrix in the step (2), if i=k, a (i, k)
Calculation method are as follows:
Further, t is measured in the step (5)qFrom different cluster centre pointsThe distance between be tqIn cluster
The size of Euclidean distance between heart point.
Beneficial effects of the present invention are as follows:
Industrial process flexible measurement method of the invention is using the AP-XGBOOST algorithm based on output constraint in complicated multiplexing
There is better predictive ability and lower prediction error in condition problem;Affine propagation clustering (AP) algorithm based on output constraint
Sample Similarity is measured according to hidden variable, output information is effectively utilized, avoids influence of the information redundancy to clustering precision, it is more quasi-
Really by the different producing condition classifications of process, more accurate submodel is established on this basis;Meanwhile based on integrated study
XGBOOST model then improves model for the mapping ability of non-linear non-gaussian complex process.
Detailed description of the invention
Fig. 1 is the situation of change of CO residual concentration in process gas in historical sample under different operating conditions;Scheme (1a)-(1c)
Respectively indicate operating condition 1,2,3;
Fig. 2 is that prediction of the AP-SVR method in test data set shows curve graph, and figure (2a)-(2c) respectively indicates operating condition
1,2,3;
Fig. 3 is that prediction of the AP-XGBOOST method in test data set shows curve graph;Figure (3a)-(3c) respectively indicates
Operating condition 1,2,3;
Fig. 4 is that prediction of the AP-XGBOOST method of output constraint in test data set shows curve graph, is schemed (4a)-
(4c) respectively indicates operating condition 1,2,3.
Specific embodiment
Below according to attached drawing and preferred embodiment the present invention is described in detail, the objects and effects of the present invention will become brighter
White, below in conjunction with drawings and examples, the present invention will be described in further detail.It should be appreciated that described herein specific
Embodiment is only used to explain the present invention, is not intended to limit the present invention.
The present invention is generally in different operation operating conditions for complex process so as to cause soft-sensing model degeneration etc.
Problem measures Sample Similarity using affine propagation clustering (AP) algorithm based on output constraint, first by original variable space
It projects to latent variables space and carries out measuring similarity, then classified using affine propagation clustering algorithm to multiple operating modes process, finally
Soft survey of different limit grad enhancement tree (XGBOOST) models for multi-state complex process is established on obtained data set
Amount modeling.
A kind of industrial process flexible measurement method based on output constraint AP-XGBOOST model, which is characterized in that including with
Lower step:
(1) the input auxiliary variable matrix that given historical sample integrates is X ∈ RL×N, target variable matrix Y ∈ RL;Wherein L and
N respectively represents sample number and input variable dimension, by Partial Least Squares to auxiliary variable matrix X and target variable matrix Y
Supervision dimensionality reduction: X=TP is carried outT+ E, Y=TRT+F;Wherein, matrix T is that historical sample matrix is projected by Partial Least Squares
Obtained new sample matrix, P are the loading matrix of historical sample matrix, and R is the loading matrix of target variable matrix Y, E and F
Corresponding residual matrix when respectively with Partial Least Squares fitting X and Y;
(2) by sample matrix T calculate historical sample concentrate data point between similarity matrix S:s (i, j)=- | |
ti-tj||2, wherein s (i, j) is the element in matrix S, the element t in representative sample matrix TiAnd tjBetween similarity it is big
It is small;Carry out iteration according to similarity matrix S and updates Certainty Factor matrix and evidence availability matrix
Wherein, r (i, k), r (k, k), r (j, k) are the element of Certainty Factor matrix, and a (i, k), a (j, i) are respectively
The element of evidence availability matrix;
(3) when Certainty Factor matrix and evidence availability matrix converges on fixed value or the number of iterations reaches maximum value
When, calculate each cluster centre point I=1 ..., θ, and determine cluster centre number θ
With corresponding Sub Data Set(m=1... θ);
(4) XGBOOST modeling algorithm is called to construct multiple submodel H (m) on Sub Data Set, (m=1... θ);When new
Sample xqWhen arrival, according to formula tq=xqP calculates sample xqVariable t after being projected by Partial Least Squaresq;
(5) t is calculatedqFrom different cluster centre pointsThe distance between, and determine therefrom that the cluster nearest apart from the sample
CenterAnd by tqThe XGBOOST submodel H (m) constructed in the step (4) corresponding with the cluster centre is inputted, is calculated final defeated
It is worth out, as model is directed to original sample xqPredicted value.
Preferably, the step (1) specifically: the latent variables space dimension of the Partial Least Squares is K, is passed through
The input and output of history data set are expressed as by least square method
Preferably, in the step (2) when calculating Certainty Factor matrix and availability matrix, Certainty Factor matrix
In the value of element a (i, k) initially initially selected in selected element r (i, k) and availability matrix be set as 0.
Preferably, when updating element a (i, k) of availability matrix in the step (2), if the meter of i=k, a (i, k)
Calculation method are as follows:
Preferably, t is measured in the step (5)qFrom different cluster centre pointsThe distance between be tqWith cluster centre
The size of Euclidean distance between point.
Verified below by way of unit is just become for the conversion process of CO in synthesis ammonia platform proposed based on output
Constrain the validity that AP-XGBOOST model is applied to multi-state complex process.It is the key that in synthetic ammonia process that height, which becomes unit,
Production unit.During ammonia synthesis, a kind of important process materials are hydrogen, it is generated by methane decarbonization device.So
And carbon is still with CO and CO2Form be present in process gas.The major function of height change unit is will be intractable
Carbon monoxide (CO) is converted into carbon dioxide (CO2), in subsequent CO2Absorbing carbon dioxide in absorption tower.When process gas stream
When crossing height converter unit, the overall chemical reaction of generation is described as follows:
Different reaction temperatures can generate different types of as a result, therefore respectively using difference in high temperature tower and cryogenic columns
Catalyst realize, to meet a variety of production requirements.The final target of device is to reduce in process gas to the maximum extent
CO content, and the technique is steadily run in a manner of energy-efficient.Therefore, it realizes and reduces this target of CO content in process gas
First is that the content of CO is remained in measurement unit outer tube with most important step.Real process passes through offline laboratory point
Analysis, with the content of extremely low sample rate measurement remnants CO.In order to predict the concentration of remnants CO in process gas, 26 are had chosen herein
A performance variable and state variable establish soft-sensing model as auxiliary variable.
This method is using the 100000 variable sample datas collected in the platform history production process.When according to acquisition
Between sequence, choose the first eight ten thousand data point for modeling, rear 20,000 data points are for test model for future time instance CO
The estimated performance of concentration avoids modeling space from being overlapped with space is predicted with this.Fig. 1 shows under operating conditions different in historical sample
The situation of change of CO residual concentration in process gas.
When establishing the AP-XGBOOST model based on output constraint, maximal tree depth and the number of iterations are respectively set to 3 Hes
200, XGBOOST submodule shape parameter is then obtained by cross validation.Fig. 2-Fig. 4 is AP-SVR model, AP-XGBOOST respectively
Model, the AP-XGBOOST based on output constraint the prediction performance in test data set respectively.
AP algorithm it can be seen from Fig. 2-Fig. 4 based on output constraint measures Sample Similarity, effectively benefit according to hidden variable
With output information, influence of the information redundancy to clustering precision is avoided, more accurately by the different producing condition classifications of process;Meanwhile
XGBOOST algorithm improves model for the mapping ability of non-linear non-gaussian complex process.By to synthetic ammonia process height
The practical study for becoming unit illustrates that method proposed by the present invention can efficiently solve the multi-state in complex process, non-
Linearly, the characteristics such as non-gaussian have preferable accuracy when solving the problems, such as complex process soft sensor modeling.
It will appreciated by the skilled person that being not used to limit the foregoing is merely the preferred embodiment of invention
System invention, although invention is described in detail referring to previous examples, for those skilled in the art, still
It can modify to the technical solution of aforementioned each case history or equivalent replacement of some of the technical features.It is all
Within the spirit and principle of invention, modification, equivalent replacement for being made etc. be should be included within the protection scope of invention.
Claims (5)
1. a kind of industrial process flexible measurement method based on output constraint AP-XGBOOST model, which is characterized in that including following
Step:
(1) the input auxiliary variable matrix that given historical sample integrates is X ∈ RL×N, target variable matrix Y ∈ RL;Wherein L and N points
Other representative sample number and input variable dimension carry out auxiliary variable matrix X and target variable matrix Y by Partial Least Squares
There is supervision dimensionality reduction: X=TPT+ E, Y=TRT+F;Wherein, matrix T is that historical sample matrix projects to obtain by Partial Least Squares
New sample matrix, P be historical sample matrix loading matrix, R be target variable matrix Y loading matrix, E and F difference
Corresponding residual matrix when for Partial Least Squares fitting X and Y.
(2) by sample matrix T calculate historical sample concentrate data point between similarity matrix S:s (i, j)=- | | ti-tj
||2, wherein s (i, j) is the element in matrix S, the element t in representative sample matrix TiAnd tjBetween similarity size;According to
Carry out iteration according to similarity matrix S and updates Certainty Factor matrix and evidence availability matrix
Wherein, r (i, k), r (k, k), r (j, k) are the element of Certainty Factor matrix, and a (i, k), a (j, i) are respectively evidence
The element of availability matrix;
(3) when Certainty Factor matrix and evidence availability matrix converge on fixed value or the number of iterations reaches maximum value,
Calculate each cluster centre point And determine cluster centre number θ and right
The Sub Data Set answered
(4) XGBOOST modeling algorithm is called to construct multiple submodel H (m) on Sub Data Set, (m=1... θ);As new samples xq
When arrival, according to formula tq=xqP calculates sample xqVariable t after being projected by Partial Least Squaresq;
(5) t is calculatedqFrom different cluster centre pointsThe distance between, and determine therefrom that the cluster centre nearest apart from the sampleAnd by tqThe XGBOOST submodel H (m) constructed in the step (4) corresponding with the cluster centre is inputted, final output is calculated
Value, as model are directed to original sample xqPredicted value.
2. the industrial process flexible measurement method according to claim 1 based on output constraint AP-XGBOOST model, special
Sign is, the step (1) specifically: the latent variables space dimension of the Partial Least Squares is K, passes through least square
The input and output of history data set are expressed as by method
3. the industrial process flexible measurement method according to claim 1 or 2 based on output constraint AP-XGBOOST model,
It is characterized in that, it is initial in Certainty Factor matrix in the step (2) when calculating Certainty Factor matrix and availability matrix
The value of element a (i, k) initially selected is set as 0 in selected element r (i, k) and availability matrix.
4. the industrial process flexible measurement method according to claim 3 based on output constraint AP-XGBOOST model, special
Sign is, when updating element a (i, k) of availability matrix in the step (2), if the calculation method of i=k, a (i, k) are as follows:
5. the industrial process flexible measurement method according to claim 4 based on output constraint AP-XGBOOST model, special
Sign is, t is measured in the step (5)qFrom different cluster centre pointsThe distance between be tqThe Europe between cluster centre point
The size of family name's distance.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811060031.0A CN109255186A (en) | 2018-09-12 | 2018-09-12 | A kind of industrial process flexible measurement method based on output constraint AP-XGBOOST model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811060031.0A CN109255186A (en) | 2018-09-12 | 2018-09-12 | A kind of industrial process flexible measurement method based on output constraint AP-XGBOOST model |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109255186A true CN109255186A (en) | 2019-01-22 |
Family
ID=65047405
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811060031.0A Pending CN109255186A (en) | 2018-09-12 | 2018-09-12 | A kind of industrial process flexible measurement method based on output constraint AP-XGBOOST model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109255186A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111291020A (en) * | 2019-11-11 | 2020-06-16 | 中国计量大学 | Dynamic process soft measurement modeling method based on local weighted linear dynamic system |
CN111582308A (en) * | 2020-04-01 | 2020-08-25 | 上海华虹宏力半导体制造有限公司 | Method, device and storage medium for calculating reliability of OPC modeling data |
CN112199829A (en) * | 2020-09-27 | 2021-01-08 | 浙江大学 | Multi-working-condition multi-stage batch process monitoring method based on density peak clustering and instant learning |
CN115455668A (en) * | 2022-08-26 | 2022-12-09 | 哈尔滨工业大学(深圳) | Simulation data generation method and device for table data and electronic equipment |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20160339177A1 (en) * | 2015-05-22 | 2016-11-24 | Iowa State University Research Foundation,Inc. | Automatic insulin delivery system |
-
2018
- 2018-09-12 CN CN201811060031.0A patent/CN109255186A/en active Pending
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20160339177A1 (en) * | 2015-05-22 | 2016-11-24 | Iowa State University Research Foundation,Inc. | Automatic insulin delivery system |
Non-Patent Citations (6)
Title |
---|
TIANQI CHEN等: ""XGBoost: A Scalable Tree Boosting System"", 《PROCEEDINGS OF THE 22ND ACM SIGKDD INTERNATIONAL COFERENCE ON KNOWLEDGE DISCOVERY AND DATA MINING》 * |
ZHENGGUANG XU等: ""Complex production process prediction model based on EMD-XGBOOST-RLSE"", 《THE 9TH INTERNATIONAL CONFERENCE ON MODELLING, IDENTIFICATION AND CONTROL》 * |
俞金寿: "《工业过程先进控制》", 28 February 2002, 中国石化出版社 * |
袁小锋: ""基于即时学习的复杂非线性过程软测量建模及应用"", 《中国博士学位论文全文数据库 信息科技辑》 * |
车笑卿等: ""基于仿射传播聚类的局部TDGPR 的自适应软测量建模"", 《计算机与应用化学》 * |
郭崇慧: "《大数据与中国古村落保护》", 30 June 2017, 华南理工大学出版社 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111291020A (en) * | 2019-11-11 | 2020-06-16 | 中国计量大学 | Dynamic process soft measurement modeling method based on local weighted linear dynamic system |
CN111582308A (en) * | 2020-04-01 | 2020-08-25 | 上海华虹宏力半导体制造有限公司 | Method, device and storage medium for calculating reliability of OPC modeling data |
CN111582308B (en) * | 2020-04-01 | 2023-08-25 | 上海华虹宏力半导体制造有限公司 | Method, device and storage medium for calculating reliability of OPC modeling data |
CN112199829A (en) * | 2020-09-27 | 2021-01-08 | 浙江大学 | Multi-working-condition multi-stage batch process monitoring method based on density peak clustering and instant learning |
CN115455668A (en) * | 2022-08-26 | 2022-12-09 | 哈尔滨工业大学(深圳) | Simulation data generation method and device for table data and electronic equipment |
CN115455668B (en) * | 2022-08-26 | 2023-05-30 | 哈尔滨工业大学(深圳) | Simulation data generation method and device for table data and electronic equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109255186A (en) | A kind of industrial process flexible measurement method based on output constraint AP-XGBOOST model | |
Yuan et al. | Deep learning-based feature representation and its application for soft sensor modeling with variable-wise weighted SAE | |
CN104778298B (en) | Gaussian process based on EGMM returns soft-measuring modeling method | |
Hothorn | CRAN task view: Machine learning & statistical learning | |
CN106897774B (en) | Multiple soft measurement algorithm cluster modeling methods based on Monte Carlo cross validation | |
CN109543009A (en) | Text similarity assessment system and text similarity appraisal procedure | |
CN104915522B (en) | The hybrid modeling method and system of cohesive process priori and data-driven model | |
CN106897775A (en) | Soft-measuring modeling method based on Bayes's integrated study | |
Jin et al. | Transformer-based map-matching model with limited labeled data using transfer-learning approach | |
Yao et al. | Nonlinear Gaussian mixture regression for multimode quality prediction with partially labeled data | |
CN110795522B (en) | Method and device for predicting track position of mobile user | |
CN113673242A (en) | Text classification method based on K-neighborhood node algorithm and comparative learning | |
CN116643989A (en) | Defect prediction method for carrying out deep semantic understanding by adopting graph structure | |
Song et al. | Elastic structural analysis based on graph neural network without labeled data | |
CN106448660B (en) | It is a kind of introduce big data analysis natural language smeared out boundary determine method | |
Li et al. | Development of a Novel Soft Sensor with Long Short‐Term Memory Network and Normalized Mutual Information Feature Selection | |
CN105844303A (en) | Sampling type clustering integration method based on local and global information | |
CN114239397A (en) | Soft measurement modeling method based on dynamic feature extraction and local weighted deep learning | |
Jiang et al. | RAGAN: Regression attention generative adversarial networks | |
CN110515354A (en) | A kind of complex-curved Digitized Closed Loop precision machining method based on intellectualized technology | |
CN115775231A (en) | Cascade R-CNN-based hardware defect detection method and system | |
CN113076421B (en) | Social noise text entity relationship extraction optimization method and system | |
CN115424000A (en) | Pointer instrument identification method, system, equipment and storage medium | |
Zhang et al. | An unsupervised leak detection method with aggregating prediction and reconstruction along projection pathway for natural gas gathering pipelines | |
Wang et al. | Shapelet classification algorithm based on efficient subsequence matching |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190122 |