CN108873401A - Liquid crystal display response time prediction technique based on big data - Google Patents
Liquid crystal display response time prediction technique based on big data Download PDFInfo
- Publication number
- CN108873401A CN108873401A CN201810650117.2A CN201810650117A CN108873401A CN 108873401 A CN108873401 A CN 108873401A CN 201810650117 A CN201810650117 A CN 201810650117A CN 108873401 A CN108873401 A CN 108873401A
- Authority
- CN
- China
- Prior art keywords
- characteristic
- liquid crystal
- data
- column
- crystal display
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G02—OPTICS
- G02F—OPTICAL DEVICES OR ARRANGEMENTS FOR THE CONTROL OF LIGHT BY MODIFICATION OF THE OPTICAL PROPERTIES OF THE MEDIA OF THE ELEMENTS INVOLVED THEREIN; NON-LINEAR OPTICS; FREQUENCY-CHANGING OF LIGHT; OPTICAL LOGIC ELEMENTS; OPTICAL ANALOGUE/DIGITAL CONVERTERS
- G02F1/00—Devices or arrangements for the control of the intensity, colour, phase, polarisation or direction of light arriving from an independent light source, e.g. switching, gating or modulating; Non-linear optics
- G02F1/01—Devices or arrangements for the control of the intensity, colour, phase, polarisation or direction of light arriving from an independent light source, e.g. switching, gating or modulating; Non-linear optics for the control of the intensity, phase, polarisation or colour
- G02F1/13—Devices or arrangements for the control of the intensity, colour, phase, polarisation or direction of light arriving from an independent light source, e.g. switching, gating or modulating; Non-linear optics for the control of the intensity, phase, polarisation or colour based on liquid crystals, e.g. single liquid crystal display cells
- G02F1/1306—Details
- G02F1/1309—Repairing; Testing
Landscapes
- Physics & Mathematics (AREA)
- Nonlinear Science (AREA)
- Chemical & Material Sciences (AREA)
- Crystallography & Structural Chemistry (AREA)
- General Physics & Mathematics (AREA)
- Optics & Photonics (AREA)
- Liquid Crystal (AREA)
Abstract
The liquid crystal display response time prediction technique based on big data that the invention discloses a kind of solves the lower technical problem of precision of prediction in the prior art, realizes that step is：Obtain production process of liquid crystal displays data set I1；Judge data set I1In with the presence or absence of missing characteristic；To I1The characteristic of middle missing is filled；Extract the key feature data that complete characteristic is concentrated；Construct liquid crystal display response time prediction model；The liquid crystal display response time is predicted.The discreteness that the present invention is arranged according to the supplemental characteristic of the used equipment of the every procedure of liquid crystal display in production process and its corresponding characteristic missing, characteristic missing column are filled, interference of the data to prediction model after avoiding Missing Data Filling, the precision of prediction of liquid crystal display prediction model is further improved, can be used for the prediction of liquid crystal display response time.
Description
Technical field
The invention belongs to field of engineering technology, are related to the prediction technique of liquid crystal display response time a kind of, and in particular to
A kind of prediction technique of the liquid crystal display response time based on big data, can be used for the prediction of liquid crystal display response time.
Background technique
Semiconductor industry is an industrialization and the higher industry of the level of informatization.The industrialization of height and information-based give count
Possibility is created according to analysis.However, the production process of liquid crystal display is complex, the liquid crystal display that producing line produces daily
It is ten hundreds of, it include procedures up to a hundred, every procedure is likely to that the quality of product can be had an impact.Although different field pair
The requirement of liquid crystal display device performance is not quite similar, but when wide viewing angle, high contrast, wide colour gamut, low-power consumption and quick response
Between be technological challenge that always liquid crystal display is faced, wherein the response time of liquid crystal display is especially focus of attention.
The liquid crystal display response time refers to each pixel of liquid crystal display to the reaction speed of input signal, i.e. pixel
By blackout it is bright or by bright turn it is dark required for the time.Wherein pixel is known as upstream time, pixel by the time of the bright needs of blackout
The time that point is secretly needed by bright turn is known as downgoing time.When the response time of liquid crystal display is long, the object that is moved in picture
Cognition leaves " hangover of image ".Performance in a computer, uses mouse sometimes, does not see the position of cursor.
The prediction of liquid crystal display response time is mainly analyzed according to the data in production process of liquid crystal displays, benefit
Model is established with various assessment algorithms, a kind of technology predicted the liquid crystal display response time.It is correct to carry out liquid crystal
The prediction for showing the device response time can not only obtain the liquid crystal display response time of prediction, and according to liquid crystal display
The key feature data that response time is affected, can be with the corresponding process of key monitoring key feature data, to improve liquid
The product quality of crystal display.
The main thought for the prediction of liquid crystal display response time is by obtaining production process of liquid crystal displays at present
The data that middle detection generates change liquid crystal display composition material and according to the physical characteristic of liquid crystal to liquid crystal display
Response time is predicted.The flashing backlight based on wavelet transformation in Southeast China University's journal volume 41 in 2011 is put down as the summer shakes
A kind of LCD response time evaluation method based on wavelet transform filtering is disclosed herein in the opinion that the LCD response time is estimated, passes through reality
Border measures the liquid crystal response under continuous background light between Pyatyi gray scale and flashes the background photoresponse of bias light, constructs flashing
LCD luminosity response under bias light, according to the data that liquid crystal display detection generates, using wavelet transform filtering method,
By wavelet decomposition, threshold process, inverse transformation reconstruct, the liquid crystal display response time is carried out by reduction liquid crystal response signal
Prediction.This method reduce the errors between the predicted value of liquid crystal display response time and true value, but have a defect that
In the filling to missing data, the brightness data of liquid crystal display detection generation is only utilized, does not make full use of
The characteristic of liquid crystal display in production process, leads to the mistake between the predicted value and true value of liquid crystal display response time
Difference is larger.
Summary of the invention
It is an object of the invention to overcome the problems of the above-mentioned prior art, a kind of liquid crystal based on big data is proposed
Show device response time prediction technique, for solving the lower technical problem of precision of prediction existing in the prior art.
To achieve the above object, the technical solution that the present invention takes includes the following steps：
(1) production process of liquid crystal displays data set I is obtained1：
The supplemental characteristic of characteristic and the used equipment of every procedure to liquid crystal display in the production process of acquisition
Duplicate removal is carried out, production process of liquid crystal displays data set I is obtained1；
(2) judge data set I1In with the presence or absence of missing characteristic：
Count I1Middle characteristic quantity n and non-empty characteristic quantity m, if n is not equal to m, I1The middle spy that there is missing
Levy data；
(3) to I1The characteristic of middle missing is filled：
I is arranged in (3a)1The discreteness threshold value of middle characteristic is β；
(3b) is according to I1In the used equipment of every procedure supplemental characteristic and its corresponding characteristic, calculate feature
There are the length of the column of missing data in data, while there are the classifications of characteristic in missing data column for statistics, and by depositing
The classification of characteristic in the length of missing data column and column, calculates the discrete property coefficient of characteristic missing column；
(3c) lacks the mode pair of column data with characteristic when the discrete property coefficient of characteristic missing column is greater than β
I1The characteristic of middle missing is filled, and when the discrete property coefficient of characteristic missing column is less than or equal to β, uses characteristic
According to the mean value of missing column data to I1The characteristic of middle missing is filled, and obtains complete characteristic data set；
(4) the key feature data that complete characteristic is concentrated are extracted：
Complete characteristic data set is normalized in (4a), the characteristic data set I after being normalized2；
(4b) is to I2In each characteristic carry out one-way analysis of variance：
Calculate I2In each characteristic statistical check amount, and will be greater than pre-set threshold value FaStatistical check amount
Corresponding characteristic merges, and obtains characteristic data set I3；
(4c) is to I3In characteristic two-by-two carry out distance correlation analysis：
Calculate I3In the distance between characteristic related coefficient two-by-two, and the distance that pre-set threshold value a will be less than
The corresponding characteristic of related coefficient merges, and obtains key feature data set I4；
(5) liquid crystal display response time prediction model is constructed：
(5a) is by key feature data set I4In 70% data as training set data train, remaining data conduct
Test set data test；
(5b) is trained training set data train using XGboost algorithm, and it is pre- to obtain the liquid crystal display response time
Survey model；
(6) the liquid crystal display response time is predicted：
Test set data test is inputted into liquid crystal display response time prediction model, when obtaining the response of liquid crystal display
Between.
Compared with prior art, the present invention having the following advantages that：
The present invention is based on the supplemental characteristic of the characteristic and the used equipment of every procedure that acquire, according to producing
The discreteness of the supplemental characteristic of the used equipment of the every procedure of liquid crystal display and its corresponding characteristic missing column, right in journey
Characteristic missing column are filled, and using obtained complete characterization data set building liquid crystal display prediction model, and existing
The brightness data for having technology to generate using liquid crystal display detection is filled the liquid crystal display of building to missing data
Prediction model is compared, and it is pre- to effectively reduce liquid crystal display for interference of the data after reducing Missing Data Filling to prediction model
The error between the response time and true response time that model prediction obtains is surveyed, the pre- of liquid crystal display prediction model is improved
Survey precision.
Detailed description of the invention
Fig. 1 is implementation flow chart of the invention.
Specific implementation method
Below in conjunction with the drawings and specific embodiments, present invention is further described in detail.
Referring to Fig.1, a kind of liquid crystal display response time prediction technique based on big data, includes the following steps：
Step 1) obtains production process of liquid crystal displays data set I1：
The supplemental characteristic of characteristic and the used equipment of every procedure to liquid crystal display in the production process of acquisition
Duplicate removal is carried out, production process of liquid crystal displays data set I is obtained1；
The production process of liquid crystal display includes many process procedures, and includes many works in each process procedure
Sequence, each process can have an impact the response time of liquid crystal display, acquire the spy of liquid crystal display in production process
Levy the supplemental characteristic of data and the used equipment of every procedure, it can be ensured that when acquisition may be comprising influencing liquid crystal display response
Between total data, provide a preferable data basis for the filling of subsequent missing data.It avoids only walking using single process
Suddenly the data generated predict the liquid crystal display response time by constructing prediction model, generate because losing other processes
Influence the liquid crystal display response time data, cause to have an impact the precision of prediction model.
The characteristic of liquid crystal display in production process, including display sizes, the liquid crystal coefficient of viscosity, liquid crystal dielectric system
Number, voltage, electric current, liquid crystal inclination angle, thickness of metal film, coating thickness, box data processed, substrate distortion value, liquid crystal infusion volume, liquid crystal
Box thickness and etching inclination angle.
The supplemental characteristic of every used equipment of procedure, voltage, power and consumption including the used equipment of every procedure
Time.
Production process of liquid crystal displays includes procedures up to a hundred, is produced as unit of process, liquid crystal display
Device parameter data used in characteristic and every procedure in production process are by device sensor and artificial note
What the mode of record was acquired.
When acquiring the supplemental characteristic of the characteristic and the used equipment of every procedure of liquid crystal display in production process, it is
The supplemental characteristic of liquid crystal display characteristic and equipment is acquired after current process, but not to current process
The characteristic and device parameter data of change acquire again, thus the spy of liquid crystal display in collected production process
Duplicate characteristic can be generated in sign data.If directly utilizing the characteristic of liquid crystal display in collected production process
According to, will increase prediction the liquid crystal display response time required for the time.
In the present invention, according to the mice function in R language to the characteristic of liquid crystal display in collected production process
Duplicate removal is carried out according to the supplemental characteristic with the used equipment of every procedure, obtains production process of liquid crystal displays data set I1。
Step 2) judges data set I1In with the presence or absence of missing characteristic：
Count I1Middle characteristic quantity n and non-empty characteristic quantity m, if n is not equal to m, I1The middle spy that there is missing
Levy data；
In acquiring production process when the supplemental characteristic of the characteristic and the used equipment of every procedure of liquid crystal display,
Due to the uncertainty using manual record mode, error of omission may occur when recording data and generate missing data, because
This is needed to data set I1Judged with the presence or absence of the characteristic of missing.
Step 3) is to I1The characteristic of middle missing is filled：
Step 3a) setting I1The discreteness threshold value of middle characteristic is β；
Step 3b) according to I1In the used equipment of every procedure supplemental characteristic and its corresponding characteristic, calculate
There are the length of the column of missing data in characteristic, while there are the classifications of characteristic in missing data column for statistics, and lead to
Cross the discrete property coefficient that characteristic missing column are calculated there are the classification of characteristic in the length of missing data column and column；
Step 3c) when the discrete property coefficient of characteristic missing column is greater than β, with the mode of characteristic missing column data
To I1The characteristic of middle missing is filled, and when the discrete property coefficient of characteristic missing column is less than or equal to β, uses feature
The mean value of shortage of data column data is to I1The characteristic of middle missing is filled, and obtains complete characteristic data set；
Step 4) extracts the key feature data that complete characteristic is concentrated：
Step 4a) complete characteristic data set is normalized, the characteristic data set I after being normalized2；
Complete characteristic data set is normalized, data between different characteristic data in elimination complete characterization data set
The inconsistent bring dimension impact of unit, so that all characteristics are under same referential.
The present invention is standardized using min-max, and also referred to as deviation standardizes, and is to carry out linear change to former data, is made to tie
Fruit is mapped between [0-1], and calculation formula is：
x*=(x-xmin)/(xmax-xmin)
Wherein xmaxFor sample maximum, xminFor sample minimum, x is former data, x*For the data after normalization.
Step 4b) to I2In each characteristic carry out one-way analysis of variance：
Calculate I2In each characteristic statistical check amount, and will be greater than pre-set threshold value FaStatistical check amount
Corresponding characteristic merges, and obtains characteristic data set I3；
One-way analysis of variance：Test data is analyzed, whether the equal multiple Normal Means of variance test
It is equal, and then judge whether influence of each factor to single test index be significant.
In the present invention, according to the aov function in R language to I2In each characteristic carry out one-way analysis of variance, obtain
To characteristic data set I3。
Step 4c) to I3In characteristic two-by-two carry out distance correlation analysis：
Calculate I3In the distance between characteristic related coefficient two-by-two, and the distance that pre-set threshold value a will be less than
The corresponding characteristic of related coefficient merges, and obtains key feature data set I4；
Apart from correlation analysis：Study the system of the correlativity between the stochastic variable that two or more are in par
Count analysis method.
Calculate I3In the distance between characteristic related coefficient two-by-two, realize that step is：
Step 4c1) calculate I3The distance between characteristic matrix a in middle X columnj,kAnd in Y column between characteristic away from
Distance matrix b is obtained from matrixj,k, calculation formula is respectively：
aj,k=| | Xj-Xk| |, j, k=1,2 ..., n
bj,k=| | Yj-Yk| |, j, k=1,2 ..., n
Wherein j indicates I3Jth row, k indicate I3Row k, Xj、XkIt respectively indicates in X in the characteristic and X of jth row
The characteristic of row k, Yj、YkRespectively indicate the characteristic of kth column in the characteristic of jth row and Y in Y；
Step 4c2) calculate I3The centre distance matrix A of characteristic in middle X columnj,kWith Y column in characteristic center away from
From matrix Bj,k, calculation formula is respectively：
WhereinIt is aj,kThe average value of middle jth row,It is aj,kThe average value of middle kth column,It is matrix aj,kIt is total
Value,It is bj,kThe average value of middle jth row,It is bj,kThe average value of middle kth column,It is matrix bj,kGrand mean；
Step 4c3) calculate I3The distance between characteristic and characteristic in Y column covariance in middle X columnCalculation formula is：
Step 4c4) calculate I3The distance between characteristic and characteristic in Y column variance in middle X columnMeter
Calculating formula is：
Step 4c5) calculate I3The distance between characteristic and characteristic in Y column related coefficient in middle X columnIt calculates
Formula is：
Step 5) constructs liquid crystal display response time prediction model：
Step 5a) by key feature data set I4In 70% data as training set data train, remaining data is made
For test set data test；
Step 5b) training set data train is trained using XGboost algorithm, when obtaining liquid crystal display response
Between prediction model；
Realize that step is：
Step 5b1) iterative parameter of XGboost algorithm is set；
Step 5b2) training set data train is imported among XGboost algorithm, it realizes to training set data train's
Fitting, obtains liquid crystal display response time prediction model；
The advantages of integrated study is relative to a body Model is integrated study by mode appropriate, integrates many " individuals
Model ", obtained final mask performance are more excellent than the performance of a body Model.XGboost is boosting in integrated study
One kind of race's algorithm is the different Weak Classifier boosting algorithm of training set, has many advantages, such as that regression fit precision is high, speed is fast.
Step 6) predicts the liquid crystal display response time：
Test set data test is inputted into liquid crystal display response time prediction model, when obtaining the response of liquid crystal display
Between.
Effect of the invention can be further illustrated by following practice：
1) this example is under 7 (× 64) system of Intel (R) Core (TM) i5-4210U@2.10GHz, Windows,
On RStudio operation platform, modeling practice is completed；
2) by using the characteristic of liquid crystal display in the production process of certain enterprise's production of liquid crystal displays collection in worksite
Modeling practice is carried out according to the supplemental characteristic with the used equipment of every procedure；
3) in data prediction, using min-max standardized method to the characteristic of liquid crystal display in production process
According to every procedure be normalized using the supplemental characteristic of equipment, the data set after being normalized, according to characteristic
The discreteness of missing column carries out the filling of missing characteristic, obtains complete characteristic data set；
4) one-way analysis of variance, setting statistical check amount threshold value are 0.05, obtain the statistical check amount pair greater than 0.05
The data set that the characteristic answered is constituted；
5) distance correlation analyze, setting apart from correlation coefficient threshold be 0.1, obtain less than 0.1 apart from related coefficient
The key feature data set that corresponding characteristic is constituted；
6) training set divided according to key feature data set, passes through the XGboost regression model structure based on integrated study
Liquid crystal display response time prediction model is built, the number of iterations n_estimators=10000 is set；
7) test set for dividing key feature data set inputs liquid crystal display response time prediction model, obtains liquid
The predicted value of crystal display response time.
Claims (3)
1. a kind of liquid crystal display response time prediction technique based on big data, which is characterized in that include the following steps：
(1) production process of liquid crystal displays data set I is obtained1：
The supplemental characteristic of characteristic and the used equipment of every procedure to liquid crystal display in the production process of acquisition carries out
Duplicate removal obtains production process of liquid crystal displays data set I1；
(2) judge data set I1In with the presence or absence of missing characteristic：
Count I1Middle characteristic quantity n and non-empty characteristic quantity m, if n is not equal to m, I1The middle characteristic that there is missing
According to；
(3) to I1The characteristic of middle missing is filled：
I is arranged in (3a)1The discreteness threshold value of middle characteristic is β；
(3b) is according to I1In the used equipment of every procedure supplemental characteristic and its corresponding characteristic, calculate characteristic
Middle there are the length of the column of missing data, while there are the classifications of characteristic in missing data column for statistics, and are lacked by existing
The classification for losing characteristic in the length and column of data column calculates the discrete property coefficient of characteristic missing column；
(3c) lacks the mode of column data to I with characteristic when the discrete property coefficient of characteristic missing column is greater than β1In lack
The characteristic of mistake is filled, and when the discrete property coefficient of characteristic missing column is less than or equal to β, is lacked with characteristic
The mean value of column data is to I1The characteristic of middle missing is filled, and obtains complete characteristic data set；
(4) the key feature data that complete characteristic is concentrated are extracted：
Complete characteristic data set is normalized in (4a), the characteristic data set I after being normalized2；
(4b) is to I2In each characteristic carry out one-way analysis of variance：
Calculate I2In each characteristic statistical check amount, and will be greater than pre-set threshold value FaStatistical check amount it is corresponding
Characteristic merge, obtain characteristic data set I3；
(4c) is to I3In characteristic two-by-two carry out distance correlation analysis：
Calculate I3In the distance between characteristic related coefficient two-by-two, and by be less than pre-set threshold alpha apart from phase relation
The corresponding characteristic of number merges, and obtains key feature data set I4；
(5) liquid crystal display response time prediction model is constructed：
(5a) is by key feature data set I4In 70% data as training set data train, remaining data is as test set
Data test；
(5b) is trained training set data train using XGboost algorithm, obtains liquid crystal display response time prediction mould
Type；
(6) the liquid crystal display response time is predicted：
Test set data test is inputted into liquid crystal display response time prediction model, obtains the response time of liquid crystal display.
2. the liquid crystal display response time prediction technique according to claim 1 based on big data, which is characterized in that step
Suddenly in production process described in (1) characteristic and the used equipment of every procedure of liquid crystal display supplemental characteristic,
In, the characteristic of liquid crystal display includes display sizes, the liquid crystal coefficient of viscosity, liquid crystal dielectric coefficient, voltage, electric current, liquid
Brilliant inclination angle, thickness of metal film, coating thickness, box data processed, substrate distortion value, liquid crystal infusion volume, thickness of liquid crystal box and etching are inclined
Angle；The supplemental characteristic of equipment includes the voltage, power and the time of consumption of the used equipment of every procedure.
3. the liquid crystal display response time prediction technique according to claim 1 based on big data, which is characterized in that step
Suddenly calculating I described in (4c)3In the distance between characteristic related coefficient two-by-two, realize that step is：
(4c1) calculates I3The distance between characteristic matrix a in middle X columnj,kIt is obtained with the distance between characteristic matrix in Y column
To distance matrix bj,k, calculation formula is respectively：
aj,k=| | Xj-Xk| |, j, k=1,2 ..., n
bj,k=| | Yj-Yk| |, j, k=1,2 ..., n
Wherein j indicates I3Jth row, k indicate I3Row k, Xj、XkRespectively indicate in X row k in the characteristic of jth row and X
Characteristic, Yj、YkRespectively indicate the characteristic of kth column in the characteristic of jth row and Y in Y；
(4c2) calculates I3The centre distance matrix A of characteristic in middle X columnj,kWith the centre distance matrix of characteristic in Y column
Bj,k, calculation formula is respectively：
WhereinIt is aj,kThe average value of middle jth row,It is aj,kThe average value of middle kth column,It is matrix aj,kGrand mean,
It is bj,kThe average value of middle jth row,It is bj,kThe average value of middle kth column,It is matrix bj,kGrand mean；
(4c3) calculates I3The distance between characteristic and characteristic in Y column covariance in middle X columnIt calculates public
Formula is：
(4c4) calculates I3The distance between characteristic and characteristic in Y column variance in middle X columnCalculation formula
For：
(4c5) calculates I3The distance between characteristic and characteristic in Y column related coefficient in middle X columnCalculation formula is：
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810650117.2A CN108873401B (en) | 2018-06-22 | 2018-06-22 | Liquid crystal display response time prediction method based on big data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810650117.2A CN108873401B (en) | 2018-06-22 | 2018-06-22 | Liquid crystal display response time prediction method based on big data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108873401A true CN108873401A (en) | 2018-11-23 |
| CN108873401B CN108873401B (en) | 2020-10-09 |
Family
ID=64340848
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810650117.2A Active CN108873401B (en) | 2018-06-22 | 2018-06-22 | Liquid crystal display response time prediction method based on big data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108873401B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111768045A (en) * | 2020-07-03 | 2020-10-13 | 上海积成能源科技有限公司 | Method for supplementing resident electricity consumption missing data by applying multiple interpolation in resident electricity consumption management |
| CN116416884A (en) * | 2023-06-12 | 2023-07-11 | 深圳市彤兴电子有限公司 | Testing device and testing method for display module |
| CN116484269A (en) * | 2023-06-25 | 2023-07-25 | 深圳市彤兴电子有限公司 | Parameter processing method, device and equipment of display screen module and storage medium |
| WO2023184258A1 (en) * | 2022-03-30 | 2023-10-05 | 京东方科技集团股份有限公司 | Model training method, performance prediction method and apparatus, device, and medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102361014A (en) * | 2011-10-20 | 2012-02-22 | 上海大学 | State monitoring and fault diagnosis method for large-scale semiconductor manufacture process |
| CN103236383A (en) * | 2013-05-04 | 2013-08-07 | 四川虹欧显示器件有限公司 | Method for dynamically determining and optimizing parameter importance in manufacturing process of plasma display screen |
| CN104239722A (en) * | 2014-09-18 | 2014-12-24 | 山东大学 | Forecasting method based on recognition of correlational relationship between factors |
| CN105354208A (en) * | 2015-09-21 | 2016-02-24 | 江苏讯狐信息科技有限公司 | Big data information mining method |
| CN105469108A (en) * | 2015-11-17 | 2016-04-06 | 深圳先进技术研究院 | Clustering method, clustering system, clustering result evaluation method and clustering result evaluation system based on biological data |
| CN107193876A (en) * | 2017-04-21 | 2017-09-22 | 美林数据技术股份有限公司 | A kind of missing data complementing method based on arest neighbors KNN algorithms |
| CN107909433A (en) * | 2017-11-14 | 2018-04-13 | 重庆邮电大学 | A kind of Method of Commodity Recommendation based on big data mobile e-business |
-
2018
- 2018-06-22 CN CN201810650117.2A patent/CN108873401B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102361014A (en) * | 2011-10-20 | 2012-02-22 | 上海大学 | State monitoring and fault diagnosis method for large-scale semiconductor manufacture process |
| CN103236383A (en) * | 2013-05-04 | 2013-08-07 | 四川虹欧显示器件有限公司 | Method for dynamically determining and optimizing parameter importance in manufacturing process of plasma display screen |
| CN104239722A (en) * | 2014-09-18 | 2014-12-24 | 山东大学 | Forecasting method based on recognition of correlational relationship between factors |
| CN105354208A (en) * | 2015-09-21 | 2016-02-24 | 江苏讯狐信息科技有限公司 | Big data information mining method |
| CN105469108A (en) * | 2015-11-17 | 2016-04-06 | 深圳先进技术研究院 | Clustering method, clustering system, clustering result evaluation method and clustering result evaluation system based on biological data |
| CN107193876A (en) * | 2017-04-21 | 2017-09-22 | 美林数据技术股份有限公司 | A kind of missing data complementing method based on arest neighbors KNN algorithms |
| CN107909433A (en) * | 2017-11-14 | 2018-04-13 | 重庆邮电大学 | A kind of Method of Commodity Recommendation based on big data mobile e-business |
Non-Patent Citations (1)
| Title |
|---|
| 夏振平等: "基于小波变换的闪烁背光LCD响应时间估算", 《东南大学学报》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111768045A (en) * | 2020-07-03 | 2020-10-13 | 上海积成能源科技有限公司 | Method for supplementing resident electricity consumption missing data by applying multiple interpolation in resident electricity consumption management |
| WO2023184258A1 (en) * | 2022-03-30 | 2023-10-05 | 京东方科技集团股份有限公司 | Model training method, performance prediction method and apparatus, device, and medium |
| CN116416884A (en) * | 2023-06-12 | 2023-07-11 | 深圳市彤兴电子有限公司 | Testing device and testing method for display module |
| CN116416884B (en) * | 2023-06-12 | 2023-08-18 | 深圳市彤兴电子有限公司 | Testing device and testing method for display module |
| CN116484269A (en) * | 2023-06-25 | 2023-07-25 | 深圳市彤兴电子有限公司 | Parameter processing method, device and equipment of display screen module and storage medium |
| CN116484269B (en) * | 2023-06-25 | 2023-09-01 | 深圳市彤兴电子有限公司 | Parameter processing method, device and equipment of display screen module and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108873401B (en) | 2020-10-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108873401A (en) | Liquid crystal display response time prediction technique based on big data | |
| CN107194872B (en) | Remote sensed image super-resolution reconstruction method based on perception of content deep learning network | |
| CN109948471B (en) | Traffic haze visibility detection method based on improved IncepotionV 4 network | |
| CN103440654B (en) | A kind of LCD foreign body defect detection method | |
| CN106127741B (en) | Non-reference picture quality appraisement method based on improvement natural scene statistical model | |
| CN108171209A (en) | A kind of face age estimation method that metric learning is carried out based on convolutional neural networks | |
| CN101976444B (en) | Pixel type based objective assessment method of image quality by utilizing structural similarity | |
| CN107133654A (en) | A kind of method of monitor video accident detection | |
| CN102422324B (en) | Age estimation device and method | |
| CN114022770A (en) | Mountain crack detection method based on improved self-attention mechanism and transfer learning | |
| CN109214470A (en) | Image visibility detection method based on coding network fine adjustment | |
| CN108256462A (en) | A kind of demographic method in market monitor video | |
| CN110298829A (en) | A kind of lingual diagnosis method, apparatus, system, computer equipment and storage medium | |
| CN107944049A (en) | A kind of film based on deep learning recommends method | |
| CN115440346A (en) | Acne grading method, system, equipment and storage medium based on semi-supervised learning | |
| Gao et al. | A novel face feature descriptor using adaptively weighted extended LBP pyramid | |
| CN111259833A (en) | Vehicle counting method based on traffic images | |
| CN104506852A (en) | Objective quality assessment method facing video conference encoding | |
| Chen et al. | Real-time defect detection of TFT-LCD displays using a lightweight network architecture | |
| Qiao et al. | No-reference quality assessment for neutron radiographic image based on a deep bilinear convolutional neural network | |
| CN112488165A (en) | Infrared pedestrian identification method and system based on deep learning model | |
| CN108776958B (en) | Mix the image quality evaluating method and device of degraded image | |
| CN104156312B (en) | A kind of method for assessing software reliability | |
| Ji et al. | No-reference image quality assessment for dehazed images | |
| CN113255789B (en) | Video quality evaluation method based on confrontation network and multi-tested electroencephalogram signals |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20230724 Address after: East side of 1st floor, Building 6, Ruiyun, No. 99 Furong Middle Third Road, Xishan District, Wuxi City, Jiangsu Province, 214191 Patentee after: Wuxi Qigong Data Technology Co.,Ltd. Address before: 710071 Taibai South Road, Yanta District, Xi'an, Shaanxi Province, No. 2 Patentee before: XIDIAN University |
|
| TR01 | Transfer of patent right |