CN107087160A - A kind of Forecasting Methodology of the user experience quality based on BP Adaboost neutral nets - Google Patents
A kind of Forecasting Methodology of the user experience quality based on BP Adaboost neutral nets Download PDFInfo
- Publication number
- CN107087160A CN107087160A CN201710291022.1A CN201710291022A CN107087160A CN 107087160 A CN107087160 A CN 107087160A CN 201710291022 A CN201710291022 A CN 201710291022A CN 107087160 A CN107087160 A CN 107087160A
- Authority
- CN
- China
- Prior art keywords
- adaboost
- data
- user experience
- training
- kpi
- 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
- 238000000034 method Methods 0.000 title claims abstract description 44
- 230000007935 neutral effect Effects 0.000 title claims abstract description 15
- 238000012549 training Methods 0.000 claims abstract description 29
- 238000013528 artificial neural network Methods 0.000 claims abstract description 13
- 239000000284 extract Substances 0.000 claims abstract description 4
- 238000003062 neural network model Methods 0.000 claims abstract description 4
- 238000004422 calculation algorithm Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000013480 data collection Methods 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 3
- 230000003466 anti-cipated effect Effects 0.000 abstract description 2
- 230000007547 defect Effects 0.000 abstract 1
- 238000012706 support-vector machine Methods 0.000 description 10
- 238000004458 analytical method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 230000008447 perception Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000002045 lasting effect Effects 0.000 description 1
- 210000004218 nerve net Anatomy 0.000 description 1
- 230000006641 stabilisation Effects 0.000 description 1
- 238000011105 stabilization Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N17/00—Diagnosis, testing or measuring for television systems or their details
- H04N17/004—Diagnosis, testing or measuring for television systems or their details for digital television systems
Abstract
The invention discloses a kind of Forecasting Methodology of the user experience quality based on BP Adaboost neutral nets, this method is used to solve the not high defect of the forecasting accuracy for user experience quality in existing IPTV video traffics, and implementing procedure of the invention includes：First KPI data is extracted from the data that IPTV set top box is collected, and extract feature, user's viewing rate has been specifically designed as key character therein, then it is based on Adaboost frameworks, and be embedded BP neural network, as Weak Classifier, the training of BP Adaboost neural network models is completed, then the KPI data of unknown subscriber's Quality of experience is predicted.Using method of the invention, it is possible to contribute to from user's subjective feeling, user experience quality is better anticipated, and designed new model training and Forecasting Methodology can more accurately and efficiently predict user experience quality.
Description
Technical field
BP- is based on the present invention relates to the user experience quality analysis technical field in video traffic, more particularly to one kind
The Forecasting Methodology of the user experience quality of Adaboost neutral nets.
Background technology
With the rise and communication of the business such as wired 4K, mobile 2K, AR/VR and developing rapidly for network technology, script
Very powerful and exceedingly arrogant Video service is even more extremely popular, and industrial chain enthusiasm is greatly lighted, including Video service business, operation
Links including business, ISP, equipment vendor etc. all actively throw oneself into big video upsurge.In big video
Generation under, in various networks flourish Video service emphasize be not only speed, be bandwidth, be video quality, be even more
The impression and experience of user.In face of the Video service of magnanimity, the final basis of user or the experience of itself go to judge and
Selection.For video service provider and Virtual network operator, the quality of Consumer's Experience directly influences the general of Video service
And degree.Only become passive O＆M into active perception, lift service quality, so as to realize the degree of recognition of lasting lifting user and stick
Degree, by Internet resources and operation activity orientation into the user group of most worthy, realize customer group stabilization and resource it is excellent
Change, the sustainable growth of industry income could be realized in keen competition.
Scientifically define measure user experience standard, for lifting experience for be firstly the need of solve the problem of.When
Before, Quality of experience (Quality of Experience, QoE) is commonly used to evaluate and described the subjective feeling of user.QoE is got over
It is good, illustrate that the Video service that IPTV is provided is better.So, the prediction for user experience quality becomes particularly important.Although pin
It is many to the user QoE machine learning models predicted and method, but often there is the adaptability underaction of model, ignore not
With the independence between feature, cause the shortcomings of degree of accuracy of prediction is not high enough.Accordingly, it would be desirable to design efficiently and accurately user
QoE Forecasting Methodologies.And the present invention the problem of can solve above well.
The content of the invention
Present invention aims at solve above-mentioned the deficiencies in the prior art, it is proposed that one kind is based on BP-Adaboost nerve nets
The Forecasting Methodology of the user experience quality of network, this method is applied to solve to check the quality for user's body in existing IPTV video traffics
The problem of forecasting accuracy of amount is not high, this method extracts KPI data from the data that IPTV set top box is collected first, and
Feature is extracted, user's viewing rate has been specifically designed as key character therein, then based on Adaboost frameworks, and by BP
Neutral net is embedded, and as Weak Classifier, completes the training of BP-Adaboost neural network models, is then used unknown
The KPI data of family Quality of experience is predicted.Using method of the invention, it is possible to contribute to from user's subjective feeling, more
User experience quality is predicted well, and designed new model training and Forecasting Methodology can be more accurately and efficiently
Predict user experience quality.
The technical scheme adopted by the invention to solve the technical problem is that：It is a kind of based on BP-Adaboost neutral nets
The Forecasting Methodology of user experience quality, this method comprises the following steps：
Step 1：Data prediction, it is determined that the factor of influence user satisfaction.
In Key Performance Indicator (KPI) original record for the video traffic that (1-1) is collected from IPTV set top box, fixation is selected
5 attributes are included in the KPI data of time span, every KPI data：Equipment propagation delay time df, equipment packet loss lp, media
Start_time, the end time end_time of viewing video between at the beginning of Loss Rate lm, user's viewing video；
(1-2) calculates program viewing rate Vr, its calculation formula is as follows：
In above formula, program_time is the total duration of the program, and it can be by No. id of program, by inquiring about video
The programme of supplier, mapping is obtained.Finally the feature of each KPI data is:{df,lp,lm,Vr}；KPI data collection isIts corresponding label is Y={ y1,...,yi,...,yN, wherein working as yi=1 represents Consumer's Experience
It is not good, yi=0 represents that Consumer's Experience is normal.
(1-3) is standardized to KPI data collection, i.e. first obtainedMean μ and variance Σ, the data after standardization
For：X={ x1,...,xi,...,xN}。
Step 2：Training pattern.Pretreated data are inputted, training obtains BP_Adaboost models, and detailed process is such as
Under：
(2-1) initializes the weight D of training data1=(w11,…,w1i,…,w1N), wherein w1i=1/N, i=1,2 ...,
N, N represent data volume；In addition, making iterations m=1, total iterations is set as M；
(2-2) starts iteration, and using one three-layer neural network of selection, it includes input layer-hidden layer-output layer, respectively
The nodes of layer are respectively 4,8,2, and using reverse conduction (BP) algorithm of standard, training obtains Weak Classifier Gm(X), it is necessary to
That illustrate is Gm(X) it is output as 0 or 1.
(2-3) calculates training data in current class device Gm(X) error rate under：
Wherein I () is indicator function, when the formula in bracket is set up, and its value is 1；When invalid, its value is 0.wmiCome from
Weight distribution set D after the m times iterationm。
(2-4) calculates Gm(X) factor alpham：
αmRepresent Gm(X) significance level in final classification device, that is, the basic classification device G obtainedm(X) in final classification
Shared proportion in device.
(2-5) updates weight distribution collection and is combined into Dm+1=(wm1,…,wmi,…,wmN), wherein wmiCalculation formula it is as follows：
Wherein,For standardizing to above formula.
(2-6) judges whether to terminate iteration.If m<M, then jump to step (2-2), and iterations adds 1 (m=m+1),
Continue next iteration；Otherwise, iteration, output are terminatedWithComplete BP-Adaboost neutral nets point
The training process of class device.
Step 3：Complete the prediction of the user experience quality of KPI data.
The KPI data of (3-1) for unknown subscriber's Quality of experience label y'Completed to pre-process according to step 1 first, obtained
To x'.
X' as input, is substituted into the BP-Adaboost neural network classifiers trained by (3-2), as follows：
Wherein,WithFor the output obtained after step 2 training.
Beneficial effect：
1. the present invention is pre-processed to data, can be well by traditional network performance configured transmission and Consumer's Experience
Correlative factor is combined, particularly using program viewing rate as important characteristic attribute, and the present invention can aid in from householder
Perception is set out, and user experience quality is better anticipated.
2. the present invention well can be organically combined BP neural network and Adaboost algorithm, designed is new
Model training method can more accurately and efficiently predict user experience quality.
Brief description of the drawings
Fig. 1 is flow chart of the method for the present invention.
Fig. 2 contrasts for method BP_Ada and BP, SVM of the invention under three kinds of data sets mean absolute error.
Fig. 3 contrasts for method BP_Ada and BP, SVM of the invention under three kinds of data sets mean square error.
Fig. 4 contrasts for method BP_Ada and BP, SVM of the invention under three kinds of data sets absolute error.
Embodiment
The invention is described in further detail with reference to Figure of description.
As shown in figure 1, the invention provides a kind of prediction of the user experience quality based on BP-Adaboost neutral nets
Method, this method comprises the following steps：
Step 1：Data prediction, it is determined that the factor of influence user satisfaction.
In Key Performance Indicator (KPI) original record for the video traffic that (1-1) is collected from IPTV set top box, fixation is selected
5 attributes are included in the KPI data of time span, every KPI data：Equipment propagation delay time df, equipment packet loss lp, media
Start_time, the end time end_time of viewing video between at the beginning of Loss Rate lm, user's viewing video；
(1-2) calculates program viewing rate Vr, its calculation formula is as follows：
In above formula, program_time is the total duration of the program, and it can be by No. id of program, by inquiring about video
The programme of supplier, mapping is obtained.Finally the feature of each KPI data is:{df,lp,lm,Vr}；KPI data collection isIts corresponding label is Y={ y1,...,yi,...,yN, wherein working as yi=1 represents Consumer's Experience not
It is good, yi=0 represents that Consumer's Experience is normal.
(1-3) is standardized to KPI data collection, i.e. first obtainedMean μ and variance Σ, the data after standardization
For：X={ x1,...,xi,...,xN}。
Step 2：Training pattern.Pretreated data are inputted, training obtains BP_Adaboost models, and detailed process is such as
Under：
(2-1) initializes the weight D of training data1=(w11,…,w1i,…,w1N), wherein w1i=1/N, i=1,2 ...,
N, N represent data volume；In addition, making iterations m=1, total iterations is set as M；
(2-2) starts iteration, and using one three-layer neural network of selection, it includes input layer-hidden layer-output layer, respectively
The nodes of layer are respectively 4,8,2, and using reverse conduction (BP) algorithm of standard, training obtains Weak Classifier Gm(X), it is necessary to
That illustrate is Gm(X) it is output as 0 or 1.
(2-3) calculates training data in current class device Gm(X) error rate under：
Wherein I () is indicator function, when the formula in bracket is set up, and its value is 1；When invalid, its value is 0.wmiCome from
Weight distribution set D after the m times iterationm。
(2-4) calculates Gm(X) factor alpham：
αmRepresent Gm(X) significance level in final classification device, that is, the basic classification device G obtainedm(X) in final classification
Shared proportion in device.
(2-5) updates weight distribution collection and is combined into Dm+1=(wm1,…,wmi,…,wmN), wherein wmiCalculation formula it is as follows：
Wherein,For standardizing to above formula.
(2-7) judges whether to terminate iteration.If m<M, then jump to step (2-2), and iterations adds 1 (m=m+1),
Continue next iteration；Otherwise, iteration, output are terminatedWithComplete BP-Adaboost neutral nets point
The training process of class device.
Step 3：Complete the prediction of the user experience quality of KPI data.
The KPI data of (3-1) for unknown subscriber's Quality of experience label y'Completed to pre-process according to step 1 first, obtained
To x'.
X' as input, is substituted into the BP-Adaboost neural network classifiers trained by (3-2), as follows：
Wherein,WithFor the output obtained after step 2 training.
Performance evaluation：
The present invention is tested by above-mentioned flow, and data prediction is carried out first, is selected characteristic attribute parameter, is then used
BP-Adaboost neural network models complete training and predicted.Data set used includes the different use that IPTV set top box is collected
100000 records at family, are randomly divided into three data sets, to each data set, will pass through step wherein 90% as training data
Rapid 2 train forecast model.Remaining passes through step 3 and completes prediction task 10% as data to be predicted.For analysis result,
There is representative using mean absolute error (MAE), mean square error (MSE) and absolute error to contrast the inventive method and two kinds
Existing method --- the estimated performance of SVMs (SVM) and BP neural network of property.
Fig. 2 compared for BP_Ada (BP-Adaboost abbreviation, the inventive method), and BP and SVM algorithm are in three kinds of data
Mean absolute error (MAE) under collection, it is seen that the algorithm that the present invention is used is than other two methods performances more preferably,
Its mean absolute error is lower.Traditional machine learning method SVM is higher than the MAE values of other two kinds of algorithms, this explanation BP and
BP_Ada improves the degree of accuracy of prediction to a certain extent, and the estimated performance that SVM does not have neutral net is good.In addition, and BP
Performance compare, BP_Ada performance has further lifting, and its reason is mainly BP_Ada and have adjusted BP neural network
Weight, further increasing predictablity rate.
Fig. 3 compared for BP_Ada, BP, mean square error (MSE) of the SVM algorithm under three kinds of data sets, it can be seen that this
Invention has minimum mean square error, that is, illustrates BP_Ada prediction accuracy highest.
Fig. 4 compared for BP_Ada, BP, SVM absolute error.Abscissa is 21 users randomly selected out in figure.From
It can be seen from the figure that, method of the invention is effectively improved prediction by the way that BP neural network is being embedded in into Adaboost frameworks
The degree of accuracy.
Claims (5)
1. a kind of Forecasting Methodology of the user experience quality based on BP-Adaboost neutral nets, it is characterised in that methods described
Comprise the following steps：
Step 1：Data prediction, it is determined that the factor of influence user satisfaction；
Step 2：Training pattern, inputs pretreated data, and training obtains BP_Adaboost models；
Step 3：Complete the prediction of the user experience quality of KPI data.
2. a kind of Forecasting Methodology of user experience quality based on BP-Adaboost neutral nets according to claim 1,
Characterized in that, the step 1 includes：
In Key Performance Indicator (KPI) original record for the video traffic that (1-1) is collected from IPTV set top box, the set time is selected
5 attributes are included in the KPI data of length, every KPI data：Equipment propagation delay time df, equipment packet loss lp, media loss rate
Start_time, the end time end_time of viewing video between at the beginning of lm, user's viewing video；
(1-2) calculates program viewing rate Vr, its calculation formula is as follows：
In above formula, start_time and end_time are respectively the beginning and end moment of program, and program_time is should
The total duration of program, it is according to No. id of program, and by inquiring about the programme of video supplier, mapping is obtained；Final each KPI
The feature of data is:{df,lp,lm,Vr}；KPI data collection isIts corresponding label is Y=
{y1,...,yi,...,yN, wherein working as yi=1 represents that Consumer's Experience is not good, yi=0 represents that Consumer's Experience is normal；
(1-3) is standardized to KPI data collection, i.e. first obtainedMean μ and variance Σ, the data after standardization are：X={ x1,...,xi,...,xN}。
3. a kind of Forecasting Methodology of user experience quality based on BP-Adaboost neutral nets according to claim 1,
Characterized in that, the step 2 includes：
(2-1) initializes the weight D of training data1=(w11,…,w1i,…,w1N), wherein w1i=1/N, i=1,2 ..., N, N
Represent data volume；In addition, making iterations m=1, total iterations is set as M；
(2-2) starts iteration, and using one three-layer neural network of selection, it includes input layer-hidden layer-output layer, each layer
Nodes are respectively 4,8,2, and using reverse conduction (BP) algorithm of standard, training obtains Weak Classifier Gm(X), it is necessary to explanation
It is Gm(X) it is output as 0 or 1；
(2-3) calculates training data in current class device Gm(X) error rate under：
Wherein I () is indicator function, when the formula in bracket is set up, and its value is 1；When invalid, its value is 0, wmiFrom m
Weight distribution set D after secondary iterationm；
(2-4) calculates Gm(X) factor alpham：
αmRepresent Gm(X) significance level in final classification device, that is, the basic classification device G obtainedm(X) in final classification device
Shared proportion；
(2-5) updates weight distribution collection and is combined into Dm+1=(wm1,…,wmi,…,wmN), wherein wmiCalculation formula it is as follows：
Wherein,For standardizing to above formula；
(2-6) judges whether to terminate iteration, if m<M, then jump to step (2-2), and iterations adds 1 (m=m+1), continues
Next iteration；Otherwise, iteration, output are terminatedWithComplete BP-Adaboost neural network classifiers
Training process；Wherein M is total iterations set in advance, and the arbitrary integer in 10~20 can be taken in the present invention.
4. a kind of Forecasting Methodology of user experience quality based on BP-Adaboost neutral nets according to claim 1,
Characterized in that, the step 3 includes：
The KPI data of (3-1) for unknown subscriber's Quality of experience label y'Completed to pre-process according to step 1 first, obtain x'；
X' as input, is substituted into the BP-Adaboost neural network classifiers trained by (3-2), as follows：
Wherein,WithFor the output obtained after step 2 training.
5. a kind of Forecasting Methodology of user experience quality based on BP-Adaboost neutral nets according to claim 1,
It is characterized in that：Methods described extracts KPI data from the data that IPTV set top box is collected first, and extracts feature, especially
User's viewing rate is devised as key character therein, then based on Adaboost frameworks, and BP neural network it is embedded in
In, as Weak Classifier, the training of BP-Adaboost neural network models is completed, then to the KPI of unknown subscriber's Quality of experience
Data are predicted.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710291022.1A CN107087160A (en) | 2017-04-28 | 2017-04-28 | A kind of Forecasting Methodology of the user experience quality based on BP Adaboost neutral nets |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710291022.1A CN107087160A (en) | 2017-04-28 | 2017-04-28 | A kind of Forecasting Methodology of the user experience quality based on BP Adaboost neutral nets |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107087160A true CN107087160A (en) | 2017-08-22 |
Family
ID=59613041
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710291022.1A Pending CN107087160A (en) | 2017-04-28 | 2017-04-28 | A kind of Forecasting Methodology of the user experience quality based on BP Adaboost neutral nets |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107087160A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107483511A (en) * | 2017-10-10 | 2017-12-15 | 山东大学 | A kind of Streaming Media QoE control systems based on software defined network SDN |
| CN107563394A (en) * | 2017-09-19 | 2018-01-09 | 广东工业大学 | A kind of method and system of predicted pictures popularity |
| CN107733705A (en) * | 2017-10-10 | 2018-02-23 | 锐捷网络股份有限公司 | A kind of user experience quality assessment models method for building up and equipment |
| CN109905382A (en) * | 2019-02-15 | 2019-06-18 | 南京邮电大学 | The subjective and objective comprehensive estimation method of IPTV video stream traffic user experience quality |
| CN109934627A (en) * | 2019-03-05 | 2019-06-25 | 中国联合网络通信集团有限公司 | Establish the method and device of satisfaction degree estimation model |
| CN110446112A (en) * | 2019-07-01 | 2019-11-12 | 南京邮电大学 | IPTV user experience prediction technique based on two-way LSTM-Attention |
| CN111343484A (en) * | 2018-12-19 | 2020-06-26 | 飞思达技术（北京）有限公司 | IPTV/OTT intelligent quality alarm method based on artificial intelligence |
| CN111369091A (en) * | 2018-12-26 | 2020-07-03 | 中国移动通信集团四川有限公司 | Method, apparatus, device and medium for user perceptual portrait analysis |
| CN112231621A (en) * | 2020-10-13 | 2021-01-15 | 电子科技大学 | Method for reducing element detection limit based on BP-adaboost |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102946613A (en) * | 2012-10-10 | 2013-02-27 | 北京邮电大学 | Method for measuring QoE |
| US20130148525A1 (en) * | 2010-05-14 | 2013-06-13 | Telefonica, S.A. | Method for calculating perception of the user experience of the quality of monitored integrated telecommunications operator services |
| WO2016109916A1 (en) * | 2015-01-05 | 2016-07-14 | 华为技术有限公司 | Quality of experience (qoe) prediction apparatus, network device and method |
| CN105897736A (en) * | 2016-05-17 | 2016-08-24 | 北京邮电大学 | Method and device for assessing quality of experience (QoE) of TCP (Transmission Control Protocol) video stream service |
| CN106534976A (en) * | 2016-10-12 | 2017-03-22 | 南京邮电大学 | Intelligent prediction method of user satisfaction in IPTV video business |
-
2017
- 2017-04-28 CN CN201710291022.1A patent/CN107087160A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130148525A1 (en) * | 2010-05-14 | 2013-06-13 | Telefonica, S.A. | Method for calculating perception of the user experience of the quality of monitored integrated telecommunications operator services |
| CN102946613A (en) * | 2012-10-10 | 2013-02-27 | 北京邮电大学 | Method for measuring QoE |
| WO2016109916A1 (en) * | 2015-01-05 | 2016-07-14 | 华为技术有限公司 | Quality of experience (qoe) prediction apparatus, network device and method |
| CN105897736A (en) * | 2016-05-17 | 2016-08-24 | 北京邮电大学 | Method and device for assessing quality of experience (QoE) of TCP (Transmission Control Protocol) video stream service |
| CN106534976A (en) * | 2016-10-12 | 2017-03-22 | 南京邮电大学 | Intelligent prediction method of user satisfaction in IPTV video business |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107563394A (en) * | 2017-09-19 | 2018-01-09 | 广东工业大学 | A kind of method and system of predicted pictures popularity |
| CN107563394B (en) * | 2017-09-19 | 2021-01-26 | 奇秦科技(北京)股份有限公司 | Method and system for predicting popularity of picture |
| CN107483511A (en) * | 2017-10-10 | 2017-12-15 | 山东大学 | A kind of Streaming Media QoE control systems based on software defined network SDN |
| CN107733705A (en) * | 2017-10-10 | 2018-02-23 | 锐捷网络股份有限公司 | A kind of user experience quality assessment models method for building up and equipment |
| CN111343484A (en) * | 2018-12-19 | 2020-06-26 | 飞思达技术（北京）有限公司 | IPTV/OTT intelligent quality alarm method based on artificial intelligence |
| CN111369091A (en) * | 2018-12-26 | 2020-07-03 | 中国移动通信集团四川有限公司 | Method, apparatus, device and medium for user perceptual portrait analysis |
| CN109905382A (en) * | 2019-02-15 | 2019-06-18 | 南京邮电大学 | The subjective and objective comprehensive estimation method of IPTV video stream traffic user experience quality |
| CN109934627A (en) * | 2019-03-05 | 2019-06-25 | 中国联合网络通信集团有限公司 | Establish the method and device of satisfaction degree estimation model |
| CN110446112A (en) * | 2019-07-01 | 2019-11-12 | 南京邮电大学 | IPTV user experience prediction technique based on two-way LSTM-Attention |
| CN112231621A (en) * | 2020-10-13 | 2021-01-15 | 电子科技大学 | Method for reducing element detection limit based on BP-adaboost |
| CN112231621B (en) * | 2020-10-13 | 2021-09-24 | 电子科技大学 | Method for reducing element detection limit based on BP-adaboost |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107087160A (en) | A kind of Forecasting Methodology of the user experience quality based on BP Adaboost neutral nets | |
| CN107027023B (en) | Based on the VoIP of neural network without reference video communication quality method for objectively evaluating | |
| CN103179592B (en) | QoE (Quality of Experience) comprehensive evaluation method based on hierarchical tree structure | |
| CN109727246A (en) | Comparative learning image quality evaluation method based on twin network | |
| CN102946613B (en) | Method for measuring QoE | |
| CN108696403A (en) | A kind of encrypted video QoE evaluating methods based on the study of network flow latent structure | |
| CN106777024A (en) | Recognize the method and device of malicious user | |
| WO2016101464A1 (en) | Quality of experience estimation method, device, terminal and server | |
| CN103259805B (en) | The domain name access control method evaluated based on user and system | |
| CN104702666A (en) | User experience quality confirmation method and system | |
| CN105207821A (en) | Network comprehensive performance evaluation method oriented to service | |
| CN106487906B (en) | A kind of mobile Web application protocol switching method of context aware | |
| CN106339507A (en) | Method and device for pushing streaming media message | |
| WO2021051623A1 (en) | Grayscale release method and apparatus based on microservice framework, and computer device | |
| CN105100508B (en) | A kind of network voice quality appraisal procedure, device and system | |
| CN111107423A (en) | Video service playing card pause identification method and device | |
| CN109873797A (en) | Conversational video business QoE-QoS parameter mapping method based on statistical analysis | |
| Vieira et al. | A speech quality classifier based on tree-cnn algorithm that considers network degradations | |
| CN106534976A (en) | Intelligent prediction method of user satisfaction in IPTV video business | |
| CN109359686A (en) | A kind of user's portrait method and system based on Campus Network Traffic | |
| CN106127234A (en) | The non-reference picture quality appraisement method of feature based dictionary | |
| CN105245362A (en) | SDN (Software Defined Network) environment important node information acquisition method | |
| CN110008402A (en) | A kind of point of interest recommended method of the decentralization matrix decomposition based on social networks | |
| CN103281555B (en) | Half reference assessment-based quality of experience (QoE) objective assessment method for video streaming service | |
| CN106789349B (en) | Quality of experience modeling analysis and conversation flow classification based method |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170822 |
|
| WD01 | Invention patent application deemed withdrawn after publication |