CN107733705A - A kind of user experience quality assessment models method for building up and equipment - Google Patents

A kind of user experience quality assessment models method for building up and equipment Download PDF

Info

Publication number
CN107733705A
CN107733705A CN201710933787.0A CN201710933787A CN107733705A CN 107733705 A CN107733705 A CN 107733705A CN 201710933787 A CN201710933787 A CN 201710933787A CN 107733705 A CN107733705 A CN 107733705A
Authority
CN
China
Prior art keywords
data sample
training data
user experience
sample set
experience quality
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
Application number
CN201710933787.0A
Other languages
Chinese (zh)
Other versions
CN107733705B (en
Inventor
李溪林
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Ruijie Networks Co Ltd
Original Assignee
Ruijie Networks Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Ruijie Networks Co Ltd filed Critical Ruijie Networks Co Ltd
Priority to CN201710933787.0A priority Critical patent/CN107733705B/en
Publication of CN107733705A publication Critical patent/CN107733705A/en
Application granted granted Critical
Publication of CN107733705B publication Critical patent/CN107733705B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

The invention discloses a kind of user experience quality assessment models method for building up and equipment, the user experience quality for the WLAN to enterprise-level is assessed.This method includes：Training data sample set is obtained, the wireless environment parameter of wireless communication parameters and AP when the user experience quality label, AP when training data sample set includes carrying out network experience by multiple AP diverse locations of the different types of applied business in different types of user equipment in WLAN coverage are communicated with user equipment；The algorithm model used according to training data sample set corresponding to each applied business and each applied business needs, establish two disaggregated models of each applied business；Wireless parameter includes the wireless environment parameter of the wireless communication parameters and/or AP when AP is communicated with user equipment；User experience quality assessment models are obtained according to two disaggregated models of each applied business.

Description

A kind of user experience quality assessment models method for building up and equipment
Technical field
The present invention relates to communication technical field, more particularly to a kind of user experience quality assessment models method for building up and set It is standby.
Background technology
At present, the WLAN (Wireless Local Area Network, WLAN) of enterprise-level is widely used in business The network struction in the places such as field, enterprise, hospital, school, wherein, the wireless controller that the WLAN of enterprise-level includes (Wireless Access Point Controller, AC), WAP (Access Point, AP) are large number of. , it is necessary to which the optimal effectiveness reached is the user experience quality for each position for allowing user in WLAN coverage in these places (QoE, Quality of Experience, QoE) is all more good, therefore needs to comment in real time after wlan network deployment Estimate each user and use the user experience quality of wireless network (Wireless-FIdelity), and then enter traveling one to network design The optimization of step.
Wherein, influenceing the factor of user experience quality mainly includes three parts, the i.e. original quality of business, data transmission network The presentation ability of network quality, user equipment to business.Wherein, the original quality of business refers to that service provider provides business Quality, such as when business is Video service, then the original quality then refers to the definition of video, such as high definition, super can be divided into Clearly, blue light is super clear etc.；Data transmission network quality is then the quality of the transmission network during business progress, such as moves fortune Seek the quality of the network of business；Presentation ability of the user equipment to business then refers to the performance height of user equipment.
Due to the numerous types for the user equipment that user uses, such as can be mobile phone, IPAD and personal computer (personal computer, PC) etc., and it is corresponding, and the user equipment of each type can also be further including each Brand styles, meanwhile, the type of service huge number of Consumer's Experience, such as can be the networking telephone (Voice over Internet Protocol, VoIP), video calling, mobile TV, mobile phone games etc., and the AP quantity in the WLAN of enterprise-level It is huge, and dispose more intensive therefore clearly unrealistic by user feedback mode with reference to as artificial operation maintenance 's.At present, the appraisal procedure for user experience quality mainly includes commenting for the user experience quality of specific business scenario Estimate method, and for operator's cable network environment or the user experience quality appraisal procedure of operator mobile network.Example Such as mobile video method for evaluating quality, but this method is only applicable to specific a certain business scenario, and for other industry It is engaged in and does not apply to；Or for operator's cable network environment or the user experience quality assessment side of operator mobile network Method, but due to the difference of operator's cable network environment or operator mobile network and the WLAN of enterprise-level deployment way, Therefore this user experience quality appraisal procedure is not suitable for the WLAN of enterprise-level yet.It can be seen that how to the WLAN's of enterprise-level User experience quality is assessed in real time, and then is improved optimization to the WLAN of enterprise-level according to assessment result, is current Technical problem urgently to be resolved hurrily.
The content of the invention
The embodiment of the present invention provides a kind of user experience quality assessment models method for building up and equipment, for enterprise-level WLAN user experience quality is accurately assessed.
First aspect, there is provided a kind of user experience quality assessment models method for building up, this method include：
Training data sample set is obtained, the training data sample set includes at least one training data sample, and described Training data sample set is included by different types of applied business in different types of user equipment in WLAN coverage User experience quality label when interior multiple AP diverse locations carry out network experience, nothing when AP is communicated with user equipment The wireless environment parameter of line messaging parameter and AP, wherein, training data sample include a user experience quality label, At least one wireless communication parameters and at least one wireless environment parameter；
The training data sample set according to corresponding to the training data sample set obtains each applied business；
The algorithm used according to training data sample set corresponding to each applied business and each applied business needs Model, two disaggregated models of each applied business are established, two disaggregated models of each applied business are for characterizing The model of mapping relations between wireless parameter and user experience quality；The wireless parameter is led to including the AP with user equipment The wireless environment parameter of wireless communication parameters and/or AP during letter；
The user experience quality assessment models are obtained according to two disaggregated models of each applied business.
Optionally, the training data sample set according to corresponding to the training data sample set obtains each applied business, Including：
The training data sample set is standardized so that all data in the training data sample set Dimension is unified；
The training number of each applied business is determined from the training data sample set according to the priority of applied business According to sample set；Network quality of the priority of the applied business according to needed for the applied business is set；
Data and user experience quality mark in each training data sample included according to the training data sample set The correlation of label, the data of the predetermined quantity in each training data sample are selected, it is final to obtain each applied business Training data sample set.
Optionally, the priority according to applied business determines each applied business from the training data sample Training data sample set, including：
Positive training data sample of the priority higher than the other application business of the first applied business is added to described first In the positive training data sample set of applied business, and by priority less than the other application business of first applied business Negative training data sample is added in the negative training data sample set of first applied business；Positive training data sample after addition This collection and the training data sample set for bearing training data sample set composition first applied business；Wherein, the positive training number It is that the satisfaction that user experience quality tag characterization user experiences to network is more than or equal to default satisfaction threshold according to sample Data during value, and the negative training data sample are that user experience quality tag characterization user is satisfied with journey to what network was experienced Degree is less than data during default satisfaction threshold value.
Optionally, the algorithm model that described each applied business needs to use is from advance by k- folding crossover algorithms The algorithm model of storage concentrates the minimum algorithm model of the error rate determined.
Optionally, the user experience quality assessment models are obtained according to two disaggregated models of each applied business, wrapped Include：
Two disaggregated models of each applied business are combined according to the priority of applied business, to obtain State user experience quality assessment models.
Optionally, merged by two disaggregated models of each applied business, to obtain user experience quality assessment After model, methods described also includes：
Acquisition updates the data sample set, and the sample set that updates the data updates the data sample including at least one, and described Update the data sample set be included in wireless communication parameters when the AP uploaded in AP runnings is communicated with user equipment with And AP wireless environment parameter；Wherein, one updates the data sample and includes at least one wireless communication parameters and at least one nothing Thread environment parameter；
Any one in sample set is updated the data described in calculating to update the data in sample and the training data sample set often The similarity of one training data sample, by the user of similarity highest training data sample in the training data sample set Quality of experience label as it is described any one update the data the user experience quality label of sample, to obtain carrying Consumer's Experience Quality tab updates the data sample set；
Mould is assessed to the user experience quality according to the sample set that updates the data of the carrying user experience quality label Type is updated.
Optionally, methods described also includes：
Receive the feedback message for the user experience quality difference that user equipment is sent；
Determine whether the user experience quality assessment models detect network failure corresponding to the feedback message；
If it is determined that the user experience quality assessment models be not detected by the feedback message corresponding to network failure, then obtain Take service data corresponding to the feedback message；
The user experience quality assessment models are corrected according to the service data.
Optionally, according to two disaggregated models of each applied business obtain the user experience quality assessment models it Afterwards, methods described also includes：
Obtain current network supplemental characteristic of first user equipment when carrying out network experience, the network parametric data bag Include the current radio environment parameter of the Current wireless communication parameter and AP when AP is communicated with first user equipment；
According to the current network supplemental characteristic and the user experience quality assessment models, determine that first user sets Standby active user's Quality of experience.
Optionally, the user equipment is being determined by user experience quality assessment models according to the network parametric data After user experience quality when carrying out network experience, methods described also includes：
When it is determined that result shows active user's Quality of experience difference of first user equipment, pass through the Consumer's Experience Evaluation Model on Quality calculates the weight of each network parametric data in the current network supplemental characteristic；The weight is worked as to be described Influence degree of each network parametric data to the determination result in preceding network parametric data；
Determined to cause the user's body of first user equipment to check the quality according to the weight of each network parametric data of calculating Measure the reason for poor.
Second aspect, there is provided a kind of user experience quality assessment models establish equipment, and the equipment includes：
First data capture unit, for obtaining training data sample set, the training data sample set includes at least one Individual training data sample, and the training data sample set include pass through different types of user equipment in it is different types of should User experience quality label, AP and the use during network experience are carried out with multiple AP diverse locations of the business in WLAN coverage The wireless environment parameter of wireless communication parameters and AP when family equipment is communicated, wherein, a training data sample includes One user experience quality label, at least one wireless communication parameters and at least one wireless environment parameter；
First data processing unit, for instruction corresponding to obtaining each applied business according to the training data sample set Practice set of data samples；
Model establishes unit, and industry is applied with each for the training data sample set according to corresponding to each applied business Business needs the algorithm model that uses, establishes two disaggregated models of each applied business, two points of each applied business Class model is the model for characterizing mapping relations between wireless parameter and user experience quality；The wireless parameter includes the AP The wireless environment parameter of wireless communication parameters and/or AP when being communicated with user equipment；
The model is established unit and is additionally operable to, and the Consumer's Experience is obtained according to two disaggregated models of each applied business Evaluation Model on Quality.
Optionally, first data processing unit obtains each applied business pair according to the training data sample set The training data sample set answered, including：
The training data sample set is standardized so that all data in the training data sample set Dimension is unified；
The training number of each applied business is determined from the training data sample set according to the priority of applied business According to sample set；Network quality of the priority of the applied business according to needed for the applied business is set；
Data and user experience quality mark in each training data sample included according to the training data sample set The correlation of label, the data of the predetermined quantity in each training data sample are selected, it is final to obtain each applied business Training data sample set.
Optionally, first data processing unit according to the priority of applied business from the training data sample really The training data sample set of each fixed applied business, including：
Positive training data sample of the priority higher than the other application business of the first applied business is added to described first In the positive training data sample set of applied business, and by priority less than the other application business of first applied business Negative training data sample is added in the negative training data sample set of first applied business；Positive training data sample after addition This collection and the training data sample set for bearing training data sample set composition first applied business；Wherein, the positive training number It is that the satisfaction that user experience quality tag characterization user experiences to network is more than or equal to default satisfaction threshold according to sample Data during value, and the negative training data sample are that user experience quality tag characterization user is satisfied with journey to what network was experienced Degree is less than data during default satisfaction threshold value.
Optionally, the algorithm model that each described applied business needs to use passes through for first data processing unit K- folds crossover algorithm and concentrates the minimum algorithm model of the error rate determined from the algorithm model prestored.
Optionally, the model establishes unit and obtains the Consumer's Experience according to two disaggregated models of each applied business Evaluation Model on Quality, including：
The model establishes unit according to the priority of applied business by two disaggregated models of each applied business It is combined, to obtain the user experience quality assessment models.
Optionally, the equipment also includes the first model modification unit；
The data capture unit is additionally operable to, and acquisition updates the data sample set, and the sample set that updates the data is included at least One updates the data sample, and the sample set that updates the data is included in the AP uploaded in AP runnings and user equipment progress The wireless environment parameter of wireless communication parameters and AP during communication；Wherein, one updates the data sample and includes at least one nothing Line messaging parameter and at least one wireless environment parameter；
First data processing unit is additionally operable to, and any one in sample set is updated the data described in calculating and updates the data sample Originally, will be similar in the training data sample set with the similarity of each training data sample in the training data sample set The user experience quality label of degree highest training data sample be used as described in any one update the data the Consumer's Experience of sample Quality tab, sample set is updated the data to obtain carrying user experience quality label；
The first model modification unit is used to update the data sample set according to the carrying user experience quality label The user experience quality assessment models are updated.
Optionally, the equipment also includes the second model modification unit；
The data capture unit is additionally operable to, and receives the feedback message for the user experience quality difference that user equipment is sent；
First data processing unit is additionally operable to, and it is described to determine whether the user experience quality assessment models detect Network failure corresponding to feedback message；Corresponded to if it is determined that the user experience quality assessment models are not detected by the feedback message Network failure, then obtain service data corresponding to the feedback message；
The second model modification unit is used for, and the user experience quality assessment models are entered according to the service data Row correction.
Optionally, including：
Second data capture unit, for obtaining current network parameter number of first user equipment when carrying out network experience According to the network parametric data includes the Current wireless communication parameter and AP when AP is communicated with first user equipment Current radio environment parameter；
Assessment unit, for according to the current network supplemental characteristic and the user experience quality assessment models, it is determined that Active user's Quality of experience of first user equipment.
Optionally, the equipment also includes the second data processing unit；
Second data processing unit is used for, when it is determined that result shows active user's experience of first user equipment When of poor quality, each network parameter number in the current network supplemental characteristic is calculated by the user experience quality assessment models According to weight；The weight is influence of each network parametric data to the determination result in the current network supplemental characteristic Degree；Determined to cause the user experience quality of first user equipment poor according to the weight of each network parametric data of calculating The reason for.
The third aspect, there is provided a kind of computer installation, described device include processor, and the processor is used to perform storage The user experience quality assessment models method for building up and second provided such as first aspect is provided during the computer program stored in device The step of any one of the user experience quality appraisal procedure that aspect provides methods described.
Fourth aspect, there is provided a kind of computer-readable recording medium, be stored thereon with computer program, the computer journey Realize that user experience quality assessment models method for building up and second aspect as first aspect provides carry when sequence is executed by processor The step of any one of the user experience quality appraisal procedure of confession methods described.
In embodiments of the present invention, user is not only included in progress network for establishing the training sample data of assessment models The operational factor and AP of AP during experience and the interaction parameter of user equipment, include the real Quality of experience label of user, So, the assessment models of foundation then not just pass through the web service parameter of acquisition, i.e. AP operational factor and AP and user The interaction parameter of equipment is come the usage experience quality of user predicted, but the actual experience closer to user is experienced, then assesses Result it is more accurate.Meanwhile in embodiments of the present invention, collection training sample data can include a variety of user equipmenies and The Consumer's Experience supplemental characteristic of a variety of applied business under various circumstances, that is, consider a variety of user equipmenies or a variety of apply industry The deviation of user experience quality caused by between business so that assessment of the invention is not limited in a kind of user equipment or one kind Specific applied business, and then the scope of application is wider.
Brief description of the drawings
In order to illustrate the technical solution of the embodiments of the present invention more clearly, it will make below to required in the embodiment of the present invention Accompanying drawing is briefly described, it should be apparent that, accompanying drawing described below is only some embodiments of the present invention, for For those of ordinary skill in the art, on the premise of not paying creative work, other can also be obtained according to these accompanying drawings Accompanying drawing.
Fig. 1 is the schematic flow sheet of user experience quality assessment models method for building up provided in an embodiment of the present invention；
Fig. 2 is a kind of hierarchical structure schematic diagram of user experience quality assessment models provided in an embodiment of the present invention；
Fig. 3 is a kind of structural representation that user experience quality assessment models provided in an embodiment of the present invention establish equipment；
Fig. 4 is a kind of structural representation of computer installation provided in an embodiment of the present invention.
Embodiment
To make the purpose, technical scheme and advantage of the embodiment of the present invention clearer, below in conjunction with the embodiment of the present invention In accompanying drawing, the technical scheme in the embodiment of the present invention is clearly and completely described.
The technical background of the embodiment of the present invention is described below.
At present, the appraisal procedure for user experience quality mainly includes checking the quality for the user's body of specific business scenario Appraisal procedure is measured, and for operator's cable network environment or the user experience quality assessment side of operator mobile network Method.Wherein, the user experience quality appraisal procedure for specific business scenario refers to be assessed for a certain specific business, Such as mobile video method for evaluating quality, but this method is only applicable to specific a certain business scenario, and for others Business does not apply to simultaneously；And for operator's cable network environment or the user experience quality assessment side of operator mobile network Method, due to the difference of operator's cable network environment or operator mobile network and the WLAN of enterprise-level deployment way, because This this user experience quality appraisal procedure is not suitable for the WLAN of enterprise-level yet.It can be seen that how to the WLAN of enterprise-level use Family Quality of experience is accurately assessed, and then is improved optimization to the WLAN of enterprise-level according to assessment result, be at present urgently Technical problem to be solved.
In consideration of it, the embodiment of the present invention provides a kind of user experience quality appraisal procedure, in the method, commented for establishing Estimating the training sample data of model not only includes AP of the user when carrying out network experience operational factor and AP and user equipment Interaction parameter, include the real Quality of experience label of user, so, the assessment models of foundation are then not just by obtaining The operational factor and AP of the web service parameter taken, i.e. AP and the interaction parameter of user equipment are come the usage experience of user predicted Quality, but the actual experience closer to user is experienced, then the result assessed is more accurate.Meanwhile in embodiments of the present invention, The Consumer's Experience ginseng of a variety of user equipmenies and a variety of applied business under various circumstances can be included by gathering training sample data Number data, that is, consider the deviation of user experience quality caused by between a variety of user equipmenies or a variety of applied business so that The assessment of the present invention is not limited in a kind of user equipment or a kind of specific applied business, and then the scope of application is wider.
Technical scheme provided in an embodiment of the present invention is introduced below in conjunction with the accompanying drawings.
Fig. 1 is referred to, one embodiment of the invention provides a kind of user experience quality assessment models method for building up, and this method can To establish equipment by user experience quality assessment models provided in an embodiment of the present invention to perform, the user experience quality is assessed Model establishes equipment can realize that this method includes by monitor supervision platform of PC, server or operation maintenance personnel etc.：
Step 101：Training data sample set is obtained, training data sample set includes at least one training data sample, and Training data sample set is included by different types of applied business in different types of user equipment in WLAN coverage User experience quality label when interior multiple AP diverse locations carry out network experience, nothing when AP is communicated with user equipment The wireless environment parameter of line messaging parameter and AP, wherein, training data sample include a user experience quality label, At least one wireless communication parameters and at least one wireless environment parameter；
Step 102：The training data sample set according to corresponding to training data sample set obtains each applied business；
Step 103：Need to adopt according to training data sample set corresponding to each applied business and each applied business Algorithm model, establishes two disaggregated models of each applied business, two disaggregated models of each applied business be for Characterize the model of mapping relations between wireless parameter and user experience quality；When wireless parameter is communicated including AP with user equipment Wireless communication parameters and/or AP wireless environment parameter；
Step 104：User experience quality assessment models are obtained according to two disaggregated models of each applied business.
It is to obtain the training data sample set for establishing assessment models first in the embodiment of the present invention.Specifically, instruction Practice set of data samples to be made up of multiple training data samples, and training data sample set mainly includes the data of three parts, It is used to characterizing the good or poor user experience quality label of user experience quality, is when AP is communicated with user equipment wireless Messaging parameter data with AP wireless environment parameter, wherein, each training data sample includes the data of this three parts.
In the embodiment of the present invention, the relatively common user equipment of in the market and applied business can be selected to be trained number According to the collection of sample, such as user equipment can be mobile phone, IPAD, PC or internet of things equipment of mainstream vendor etc., using industry Business can be wechat, VOIP, HD video, web page browsing etc., it is of course also possible to be other possible user equipmenies or application Business, the present invention are not limited herein.
After it have selected the user equipment and applied business needed for data acquisition, then it can pass through the user equipment of selection Usage experience is carried out with multiple APs of the applied business in WLAN coverage diverse location, and then is gathered in usage experience mistake User experience quality label in journey.Wherein, actual experience of user experience quality tag characterization when using the wlan network, When carrying out gathering user experience quality label, the evaluation that collector is experienced to wlan network can be received, even if It is preferably or poor with experiencing.Meanwhile when gathering user experience quality label, in order to avoid the judgement of collector's subjectivity is led The error of cause, experience parameter and standard can also be set according to the different of applied business, be used to help judge user experience quality It is good and bad.For example, when applied business is specially HD video, then can according to video cardton number, video buffer speed, regard Frequency speed of download sets experience parameter and standard, when video cardton number is excessive or video speed of download is slower, and adopting now When the subjective feeling that collection personnel submit is poor, then it can determine that user experience quality now is poor, then corresponding Consumer's Experience Quality tab is then 0；If video-see is smooth, video speed of download is also very fast, and the subjectivity that collector now submits When experiencing preferably, then it can determine that user experience quality now is good, then corresponding user experience quality label is then 1.When So, experience parameter and standard and user it is widely different when, can also be judged according to actual conditions, such as regarding when detection Frequency speed of download is very slow, but video-see is but very smooth always, and at this moment can is defined according to the subjective judgement of user, i.e., this When user experience quality it is good.
While user experience quality label is gathered, AP can also report now AP wireless environment parameter and AP and use Wireless communication parameters when family equipment is communicated, the wireless environment parameter and AP of AP here are communicated with user equipment When wireless communication parameters be it is corresponding with the user experience quality label of above-mentioned collection, it is above-mentioned to adopt for a certain angle The user experience quality label of collection can be then the label of these parameters.Wherein, the user experience quality label of above-mentioned collection can To be usage experience of the collector within a period of time, then the AP that the collector is connected in this period upload with The label of the wireless environment parameter of wireless communication parameters and AP when the collector is communicated is to be received in this period Collector user experience quality label, that is to say, that wireless communication parameters and wireless environment parameter be can and user Quality of experience label is bound one by one, then then can upload wireless communication parameters and wireless environment parameter in AP will also use in the lump Family Quality of experience label uploads or after server end obtains wireless communication parameters and wireless environment parameter together, then Carried out according to time and AP position etc. with user experience quality label corresponding.
Specifically, AP wireless environment parameter can include signal to noise ratio, channel utilization, air interference rate etc. one kind or It is a variety of；Wireless communication parameters when AP is communicated with user equipment can include time delay, packet loss, received signal strength One kind of RSSI, upstream rate and downstream rate, air message retransmission rate, air message effective percentage, up-downgoing flow etc. or It is a variety of, it is, of course, also possible to without limitation including other possible parameters, the embodiment of the present invention.Wherein, AP can will be above-mentioned These parameters are sent to server end again and are handled and stored by reporting parameters to AC, AC.
After training data sample set is obtained, then user experience quality label can be established according to training data sample set The user experience quality of the wireless environment parameter mapping of wireless communication parameters and AP when being communicated with AP with user equipment is commented Estimate model.
In the embodiment of the present invention, training data sample set includes the substantial amounts of training data sample of collection, therefore first may be used To be pre-processed to the training data sample set.
Specifically, each of which training data sample in training data sample set all includes user experience quality mark The data of wireless communication parameters and AP wireless environment parameter this three parts of label, AP when being communicated with user equipment, by The wireless environment parameter of wireless communication parameters and AP when AP is communicated with user equipment may also further include Polytype data, but the dimension of each data may be it is different, such as time delay dimension for millisecond (ms) or Second (s), and the dimension of upstream rate is then bps (bits per second, bps), the difference of dimension causes different numbers It is compared according to that can not be placed in same aspect, therefore in order to which what follow-up data can be unified treats and handle, may be used also first To remove the dimension of all data.Specifically, training data sample set can be standardized by z-score (z-score) Processing so that the dimension of all data is unified in training data sample set, naturally it is also possible to united by the method for other standards One dimension, the present invention are without limitation.Wherein, because the method for z-score standardization is known to those skilled in the art Standardized method, therefore this process belongs to the category of prior art, will not be repeated here.
Due to reasons such as network transmission or data acquisitions, some numbers are there may be in the training data sample set of acquisition According to the sample data of missing, such as one or more parameter number in the wireless environment parameter of AP in some sample data According to missing, such sample data can not be used, it is necessary to removed from training data sample set, therefore can entered While the operation of the unified dimension of row or after which, the sample data that shortage of data be present is removed.
In the embodiment of the present invention, because user equipment class is various used in user, user device type can not be directed to To assess user experience quality, and user equipment finally also needs to carry out network experience by way of applied business, therefore It is this to be assessed for user device type, for being assessed relative to different applied business types, actual meaning It is adopted little.Therefore, after being pre-processed to data, then training data sample set can be entered according to the difference of applied business Row division, to obtain training data sample set corresponding to different application business.
In order that the sample data for obtaining each applied business is more abundant, can be by for gathering training data sample Applied business is divided into different priority, and is handled the training data sample of different application business according to priority.
Wherein, network quality required during this kind of applied business of the height of priority and experience is relevant, such as can be body The network quality tested needed for this kind of applied business is higher, then the priority of the applied business is higher, and the division of priority can be by Divided according to experience or experimental result.If for example, applied business include HD video, web page browsing, wechat and VOIP, then according to abundant experimental results it is recognised that HD video user experience quality preferably when, other application The user experience quality of business also all preferably, but the user experience quality of other application business preferably when, the user of HD video Quality of experience is but not necessarily preferably, therefore the network quality required for HD video is above other application business, i.e. high definition The highest priority of video.The order of the priority obtained according to abundant experimental results from high to low be followed successively by HD video, VOIP, web page browsing, wechat.
In specific embodiment, training data sample can be divided into positive training data sample and and negative training data sample, its In, positive training data sample is more than or equal to default for the satisfaction that user experience quality tag characterization user experiences to network Data during satisfaction threshold value, for example, user experience quality label be 1 when training data sample；And negative training data sample It is less than data during default satisfaction threshold value for the satisfaction that user experience quality tag characterization user experiences to network, such as Training data sample when user experience quality label is 0.It is determined that each application corresponding to training data sample set when, specifically It is described by taking the first applied business as an example.Specifically, priority can be higher than to the other application business of the first applied business Positive training data sample added in the first applied business originally positive training data sample set that collects, obtain the first application Positive training data sample set corresponding to business；And priority is less than to the negative training of the other application business of the first applied business In the negative training data sample set that data sample collects originally added to the first applied business, the first applied business pair is obtained The negative training data sample set answered, so both can group by the positive training data sample set after adding and negative training data sample set Into the training data sample set of the first applied business.
Such as when the first applied business is VOIP, because VOIP priority is less than HD video, when HD video User experience quality label be 1, i.e., when user experience quality is good, then VOIP user experience quality generally also preferably, because The positive training data sample of this HD video can also be as VOIP positive training data sample, you can with by HD video just Training data sample is added in VOIP positive training data sample set；Because VOIP priority is higher than web page browsing or micro- Letter, when the user experience quality label of web page browsing or wechat is 0, i.e., user experience quality difference when, then VOIP use Family Quality of experience is generally also to be poor, therefore the negative training data sample of web page browsing or wechat can also be as VOIP negative instruction Practice data sample, you can so that the negative training data sample of web page browsing or wechat to be added to VOIP negative training data sample Concentrate.
In the embodiment of the present invention, due in the training data sample set of each applied business, not all number According to deciding factor is all played to user experience quality, i.e., data larger to the influence degree of user experience quality only include master The several data wanted, therefore after the training data sample set of each applied business is obtained, the instruction of each applied business Practicing set of data samples all includes positive training data sample and negative training data sample, can also be included according to training data sample set Each training data sample in data and user experience quality label correlation, select predetermined in each sample data The data of quantity, to obtain the final training data sample set of each applied business.
Specifically, it can be calculated by PCA (Principal Components Analysis, PCA) every One data and the correlation of user experience quality, then the larger data of relevance values of wherein predetermined quantity are filtered out as special Data are levied, characteristic is to play the key data of deciding factor in the applied business to user experience quality, here Predetermined quantity can be then configured according to actual conditions, or select the optimal quantity of test result, example according to test result It such as could be arranged to 8；And other can then be deleted with the less data of user experience quality correlation, to avoid information redundancy, And influence the accuracy of final result.
, then can root after the final training data sample set of each applied business is obtained in the embodiment of the present invention User experience quality assessment models are established according to the final training data sample set of each applied business of acquisition.
Specifically, the algorithm model for being adapted to each applied business can be selected from pre-stored algorithm model set, And then establish a kind of algorithm model for each applied business.Wherein, can specifically be wrapped in pre-stored algorithm model set Include a variety of machine learning algorithm models, for example, can include algorithm of support vector machine (Support Vector Machine, SVM), linear discriminent analysis (Linear Discriminant Analysis, LDA), logistic regression analysis models, people Artificial neural networks (Artificial Neural Networks, ANNs).When specifically being selected, it can be rolled over and handed over by k- The method of checking is pitched to select the optimal algorithm model for each applied business.Wherein, k- roll over cross validation refer to by The sample data for the 1/K that the training sample data collection of each applied business includes is as test data, and remaining a part of number According to as data are verified, each algorithm model in optional algorithm model set is trained K times according to test data, then By verifying that data are verified to the model for training to obtain each time, finally take average be used as of the error rate of K checking should The terminal error rate of kind algorithm model, after the completion of the terminal error rate of all algorithm models calculates, selection wherein error rate is most Algorithm model of the low algorithm model as corresponding applied business.
, then can be according to the instruction of each applied business after the algorithm model that each applied business needs to use determines Practice set of data samples and the algorithm model that determines establishes two sorting algorithm models of each applied business, two points mentioned here Class algorithm model refers to that the model that each applied business is established finally all is used to carry out two classification, that is, is used to determine user's body The amount of checking the quality is preferably still to be poor.
, then can basis after two sorting algorithm models of each applied business are successfully established in the embodiment of the present invention Two sorting algorithm models of all applied business are combined by the priority relationship between applied business, to obtain final use Family Quality of experience assessment models.Specifically, according to the priority relationship of different applied business, by the application industry that priority is minimum Two sorting algorithm models of business are put in the bottom, and up then by priority ranking from low to high in the way of by different application industry Two sorting algorithm models of business are ranked up.Such as shown in Fig. 2, show for a kind of hierarchical structure of user experience quality assessment models It is intended to, wherein, it is followed successively by two points of HD video according to the algorithm model hierarchical relationship of each applied business of the order of priority Class algorithm model, VOIP two sorting algorithm models, two sorting algorithm models of web page browsing, two sorting algorithm moulds of wechat Type.When needing by the user experience quality assessment models of foundation to determine user experience quality, then from the wechat of the bottom Two sorting algorithm models proceed by determination, if two sorting algorithm models of wechat determine user experience quality preferably, then Continue the determination of last layer, until the user experience quality that a certain layer determines is poor or number of plies highest applied business two points Untill class algorithm model, for example, when it is determined that to VOIP two sorting algorithm models when, the user experience quality of the applied business is When poor, then illustrate the user experience quality of applied business of the priority less than VOIP all preferably；Or when HD video is arrived in determination Two sorting algorithm models when, the user experience quality of the applied business still preferably, then illustrates the user of all applied business Quality of experience is all preferably.
In the embodiment of the present invention, due in the training data sample for establishing user experience quality assessment models only Sample data including in the market mainstream user equipment and applied business, and it is actual also there is other kinds of user equipment and Applied business, in order to improve the compatibility of user experience quality assessment models, therefore also need to user experience quality assessment models Carry out further optimization renewal.
In the embodiment of the present invention, it can also obtain and update the data sample set, to be carried out to user experience quality assessment models Further optimization renewal.
In the embodiment of the present invention, in AP actual moving process, understand and ceaselessly report service data to AC, AC can then be incited somebody to action These service datas upload onto the server end, therefore can be obtained by server end and update the data sample set.But due to obtaining What is taken updates the data data when sample set is AP usual operation, only includes the service data of AP uploads, i.e. AP and user are set For the wireless communication parameters and AP wireless environment parameter when being communicated, and now include the experience of collector Impression, that is to say, that not including user experience quality label, then obtain update the data that sample set includes update the data sample This is all no label, it is impossible to is directly used in and carries out further optimization more to these data user's Quality of experience assessment models Newly, therefore, after acquisition updates the data sample set, can combine semi-supervised learning algorithm will update the data sample without label What collection was converted into label updates the data sample set.
Specifically, after acquisition updates the data sample set, characteristic therein can be selected, this characteristic It can be the characteristic determined above by PCA algorithms, then sample is updated the data therewith by updating the data in sample set The characteristic of the training data sample of preceding acquisition is contrasted, and is calculated similarity between the two, then will be with training number According to the user experience quality label of similarity highest training data sample in sample as updating the data without label accordingly The label of sample, in this manner it is possible to sample is converted into label updates the data sample by updating the data without label, so as to Sample set is updated the data to have a label.
Obtain have label update the data sample set after, then can update the data sample set to building according to have a label Vertical user experience quality assessment models optimize renewal.Specifically, can be by combining the renewal number for having label obtained Renewal is optimized to the user experience quality assessment models of foundation according to sample set and the training data sample set of collection.
In the embodiment of the present invention, in order to ensure the accuracy for the user experience quality assessment models established, in Consumer's Experience Before Evaluation Model on Quality comes into operation, it is required for verifying the model, if finding that the degree of accuracy is low during checking, it is also necessary to Reason is found to repair model.Specifically, the method that user experience quality assessment models are verified can include but It is not limited to following two methods：
(1) user experience quality assessment models are verified by the new training data sample of collection.Specifically, will Wireless parameter data in the new training data sample of collection are then corresponding for the wireless parameter data of input as input, output User experience quality, and user experience quality label is as judging data, and the user experience quality that will be exported is with actually adopting The user experience quality label of collection is compared, and then determines whether user experience quality assessment models are correct, is carrying out repeatedly After checking, the accuracy of multiple authentication is calculated, whether judging nicety rate reaches predetermined accuracy, then can be with if reaching Illustrate that the user experience quality assessment models accuracy is higher, actual use can be put into.
(2) data that report of AP when by obtaining that user is actual to carry out network experience from server, then assessment is sampled, The data sample of extraction is determined corresponding user experience quality by user experience quality assessment models, then obtains extraction Sample data corresponding to user actual user's Quality of experience, by model determine user experience quality and user actual use Family Quality of experience is contrasted, and then verifies the accuracy of user experience quality assessment models.
In the embodiment of the present invention, in the user experience quality evaluation process of reality, when user is actually carrying out network body When testing, it is understood that there may be user experience quality difference situation, now user the situation can then be fed back.Accordingly, connecing After the feedback message for receiving user, then it can determine whether user experience quality assessment models detect that the feedback of the user disappears Network failure corresponding to breath, if detecting, then it will not make a response；But if user experience quality assessment models are not detected by The network failure, then illustrate that current user experience quality assessment models are not perfect, then it is anti-can to obtain feedback message institute The service data during network failure of feedback, and then user experience quality assessment models are corrected according to the service data, with So that user experience quality assessment models are more accurate.
In the embodiment of the present invention, when user experience quality assessment models verify that reality can be put into by repeatedly determining that accuracy reaches After the standard that border uses, then the user experience quality assessment models can be actually used in user experience quality assessment.
In the embodiment of the present invention, when the first user equipment carries out network experience by the WLAN of enterprise-level, then it can obtain Current network supplemental characteristic of first user equipment when carrying out network experience is taken, the current network parametric data can include AP The wireless environment parameter of Current wireless communication parameter and AP when being communicated with user equipment.Specifically, can be from The AP that one user equipment is communicated obtains current network supplemental characteristic；Or obtain current network ginseng from the AC for managing this AP Number data；Or obtain current network supplemental characteristic from server end.
, then can be according to user's body after obtaining user experience quality assessment models and obtaining current network supplemental characteristic Test Evaluation Model on Quality and current network supplemental characteristic determines user experience quality corresponding to current network supplemental characteristic.Specifically , the current network supplemental characteristic of acquisition is inputted into user experience quality assessment models, from two points of the applied business of the bottom Class algorithm model proceeds by determination, if two sorting algorithm models of the applied business of the bottom determine that user experience quality is It is good, then continue the determination of last layer, until the user experience quality that a certain layer determines is difference or number of plies highest applied business Two sorting algorithm models untill, such as shown in Fig. 2 when it is determined that to VOIP two sorting algorithm models when, then can be from the bottom Two sorting algorithm models of wechat proceed by determination, if two sorting algorithm models of wechat determine that user experience quality is It is good, then continue the determination of last layer, until the user experience quality that a certain layer determines is difference or number of plies highest applied business Two sorting algorithm models untill, such as when it is determined that to VOIP two sorting algorithm models when, the Consumer's Experience of the applied business When quality is poor, then illustrate the user experience quality of applied business of the priority less than VOIP all preferably；Or when determination to height During two sorting algorithm model of clear video, the user experience quality of the applied business still preferably, then illustrates all applied business User experience quality all preferably.
In the embodiment of the present invention, show the first user equipment now in the determination results of user experience quality assessment models When user experience quality is poor, then it can be calculated by user experience quality assessment models for obtaining the current of the determination result The weight of each network parametric data in network parametric data, weight here then refer to be determined with each network parametric data pair As a result influence degree, that is to say, that in synchronization, each network parametric data is to the user experience quality of user It is influential.For example, when network speed of download is slower, user watches video may not be smooth；Or time delay it is higher when, may It is not smooth to will also result in user's viewing video, but causes to produce user experience quality in each network parametric data of synchronization The influence degree of difference is different, therefore, when it is determined that during user experience quality difference, then can passing through user experience quality assessment models meter The influence degree of each network parametric data is calculated, and then determines the maximum network parametric data of influence degree, the network parameter Data are then the main reason for causing user experience quality difference.And then can also further it be analyzed according to the network parametric data Actual cause corresponding to the network parametric data, such as wireless environment serious interference, AP covering deficiency, terminal capabilities are poor, AP work( One or more in the actual causes such as rate is unreasonable, in this manner it is possible to help running background attendant to quickly find The reason for causing user experience quality difference, and then rapid pin is repaired to the reason to network, and then lift user's body and check the quality Amount.
In summary, in embodiments of the present invention, user is not only included for establishing the training sample data of assessment models The operational factor and the interaction parameter of AP and user equipment of AP when carrying out network experience, includes the real experience of user Quality tab, so, the assessment models of foundation then not just by the web service parameter of acquisition, i.e., AP operational factor and The interaction parameter of AP and user equipment is come the usage experience quality of user predicted, but closer to the actual experience sense of user By the result then assessed is more accurate.Meanwhile in embodiments of the present invention, collection training sample data can include a variety of use The Consumer's Experience supplemental characteristic of family equipment and a variety of applied business under various circumstances, that is, consider a variety of user equipmenies or The deviation of user experience quality caused by between a variety of applied business so that assessment of the invention is not limited in a kind of user and set Standby or a kind of specific applied business, and then the scope of application is wider.
Equipment provided in an embodiment of the present invention is introduced below in conjunction with the accompanying drawings.
Fig. 3 is referred to, the same inventive concept based on the embodiment shown in Fig. 1, one embodiment of the invention provides a kind of use Family Quality of experience assessment models establish equipment 30, and the equipment includes：
First data capture unit 301, for obtaining training data sample set, training data sample set includes at least one Training data sample, and training data sample set includes passing through different types of applied business in different types of user equipment Multiple AP diverse locations in WLAN coverage carry out user experience quality label, AP and the user equipment during network experience The wireless environment parameter of wireless communication parameters and AP when being communicated, wherein, a training data sample includes a use Family Quality of experience label, at least one wireless communication parameters and at least one wireless environment parameter；
First data processing unit 302, for instruction corresponding to obtaining each applied business according to training data sample set Practice set of data samples；
Model establishes unit 303, should with each for the training data sample set according to corresponding to each applied business The algorithm model that uses is needed with business, establishes two disaggregated models of each applied business, two points of each applied business Class model is the model for characterizing mapping relations between wireless parameter and user experience quality；Wireless parameter is set including AP with user For the wireless communication parameters and/or AP wireless environment parameter when being communicated；
Model is established unit 303 and is additionally operable to, and user experience quality is obtained according to two disaggregated models of each applied business Assessment models.
Optionally, the first data processing unit 302 is according to corresponding to training data sample set obtains each applied business Training data sample set, including：
Training data sample set is standardized so that the dimension system of all data in training data sample set One；
The training data sample of each applied business is determined from training data sample set according to the priority of applied business This collection；Network quality of the priority of applied business according to needed for applied business is set；
Data and user experience quality label in each training data sample included according to training data sample set Correlation, the data of the predetermined quantity in each training data sample are selected, to obtain the final instruction of each applied business Practice set of data samples.
Optionally, the first data processing unit 302 determines often according to the priority of applied business from training data sample A kind of training data sample set of applied business, including：
Positive training data sample of the priority higher than the other application business of the first applied business is applied added to first In the positive training data sample set of business, and by priority less than the first applied business other application business negative training number It is added to according to sample in the negative training data sample set of the first applied business；Positive training data sample set and negative training after addition Set of data samples forms the training data sample set of the first applied business；Wherein, positive training data sample is user experience quality The satisfaction that tag characterization user experiences to network is more than or equal to data during default satisfaction threshold value, and negative training When data sample is that the satisfaction that user experience quality tag characterization user experiences to network is less than default satisfaction threshold value Data.
Optionally, the algorithm model that each applied business needs to use is rolled over for the first data processing unit 302 by k- Folded crossover algorithm concentrates the minimum algorithm model of the error rate determined from the algorithm model prestored.
Optionally, model establishes unit 303 and obtains user experience quality according to two disaggregated models of each applied business Assessment models, including：
Model establishes unit 303 and two disaggregated models of each applied business is carried out into group according to the priority of applied business Close, to obtain user experience quality assessment models.
Optionally, equipment also includes the first model modification unit 304；
Data capture unit is additionally operable to, and acquisition updates the data sample set, and updating the data sample set includes at least one renewal Data sample, and update the data sample set be included in it is wireless when the AP uploaded in AP runnings is communicated with user equipment The wireless environment parameter of messaging parameter and AP；Wherein, one update the data sample include at least one wireless communication parameters and At least one wireless environment parameter；
First data processing unit 302 is additionally operable to, calculating update the data any one in sample set update the data sample with The similarity of each training data sample in training data sample set, similarity highest in training data sample set is trained The user experience quality label of data sample updates the data the user experience quality label of sample as any one, to be taken Sample set is updated the data with user experience quality label；
First model modification unit 304 is used to update the data sample set to user according to carrying user experience quality label Quality of experience assessment models are updated.
Optionally, equipment also includes the second model modification unit 304；
Data capture unit is additionally operable to, and receives the feedback message for the user experience quality difference that user equipment is sent；
First data processing unit 302 is additionally operable to, and determines whether user experience quality assessment models detect feedback message Corresponding network failure；If it is determined that user experience quality assessment models are not detected by network failure corresponding to feedback message, then obtain Take service data corresponding to feedback message；
Second model modification unit 304 is used for, and user experience quality assessment models are corrected according to service data.
Optionally, including：
Second data capture unit 306, for obtaining current network ginseng of first user equipment when carrying out network experience Data are counted, the Current wireless communication parameter and AP when network parametric data is communicated including AP with the first user equipment are worked as Preceding wireless environment parameter；
Assessment unit 307, for according to current network supplemental characteristic and user experience quality assessment models, determining the first use Active user's Quality of experience of family equipment.
Optionally, equipment also includes the second data processing unit 308；
Second data processing unit 308 is used for, when it is determined that result shows active user's Quality of experience of the first user equipment When poor, the weight of each network parametric data in current network supplemental characteristic is calculated by user experience quality assessment models；Power Weight is the influence degree that each network parametric data pair determines result in current network supplemental characteristic；According to each network of calculating The reason for weight of supplemental characteristic determines to cause the user experience quality difference of the first user equipment.
The equipment can be used for performing the method that the embodiment shown in Fig. 2 is provided, therefore, for each function of the equipment Function that module can be realized etc. refers to the description of the embodiment shown in Fig. 2, seldom repeats.Wherein, due to unit 304~ Unit 308 is not essential functional module, therefore shown in broken lines in figure 3.
Fig. 4 is referred to, one embodiment of the invention also provides a kind of computer installation, and the computer installation includes processor 401, processor 401 is used to realize Consumer's Experience provided in an embodiment of the present invention when performing the computer program stored in memory The step of Evaluation Model on Quality method for building up and user experience quality appraisal procedure.
Optionally, processor 401 can be specifically central processing unit, ASIC (English： Application Specific Integrated Circuit, referred to as：ASIC), can be that one or more is used to control journey The integrated circuit that sequence performs, can be using field programmable gate array (English：Field Programmable Gate Array, referred to as：FPGA) the hardware circuit of exploitation, can be BBP.
Optionally, processor 401 can include at least one processing core.
Optionally, the computer installation also includes memory 402, and memory 402 can include read-only storage (English： Read Only Memory, referred to as：ROM), random access memory (English：Random Access Memory, referred to as：RAM) And magnetic disk storage.Memory 402 is used to store data required when processor 401 is run.The quantity of memory 402 is one It is or multiple.Wherein, memory 402 is shown in the lump in Fig. 4, but it is understood that memory 402 is not essential function mould Block, thus it is shown in broken lines in Fig. 4.
One embodiment of the invention also provides a kind of computer-readable recording medium, is stored thereon with computer program, described Such as user experience quality assessment models method for building up provided in an embodiment of the present invention is realized when computer program is executed by processor The step of with user experience quality appraisal procedure.
In embodiments of the present invention, it should be understood that disclosed apparatus and method, can be real by another way It is existing.For example, apparatus embodiments described above are only schematical, for example, the division of the unit or unit, is only A kind of division of logic function, can there is an other dividing mode when actually realizing, for example, multiple units or component can combine or Person is desirably integrated into another system, or some features can be ignored, or does not perform.Another, shown or discussed is mutual Between coupling or direct-coupling or communication connection can be INDIRECT COUPLING or communication link by some interfaces, equipment or unit Connect, can be electrical or other forms.
Each functional unit in embodiments of the present invention can be integrated in a processing unit, or unit also may be used To be independent physical module.
If the integrated unit is realized in the form of SFU software functional unit and is used as independent production marketing or use When, it can be stored in a computer read/write memory medium.Based on such understanding, the technical scheme of the embodiment of the present invention All or part can be embodied in the form of software product, the computer software product is stored in a storage medium In, including some instructions are causing a computer equipment, such as can be that personal computer, server, or network are set It is standby etc., or all or part of step of each embodiment methods described of processor (processor) the execution present invention.It is and foregoing Storage medium includes：General serial bus USB (Universal Serial Bus flash drive), mobile hard disk, only Read memory (read-only memory, ROM), random access memory (random access memory, RAM), magnetic disc Or CD etc. is various can be with the medium of store program codes.
Described above, above example is implemented above only the technical scheme of the application to be described in detail The explanation of example is only intended to help the method for understanding the embodiment of the present invention, should not be construed as the limitation to the embodiment of the present invention.This The change or replacement that those skilled in the art can readily occur in, should all cover the embodiment of the present invention protection domain it It is interior.

Claims (20)

  1. A kind of 1. user experience quality assessment models method for building up, it is characterised in that including：
    Training data sample set is obtained, the training data sample set includes at least one training data sample, and the training Set of data samples is included by different types of applied business in different types of user equipment in WLAN coverage User experience quality label when multiple AP diverse locations carry out network experience, channel radio when AP is communicated with user equipment Believe parameter and AP wireless environment parameter, wherein, a training data sample includes a user experience quality label, at least One wireless communication parameters and at least one wireless environment parameter；
    The training data sample set according to corresponding to the training data sample set obtains each applied business；
    The algorithm model used according to training data sample set corresponding to each applied business and each applied business needs, Two disaggregated models of each applied business are established, two disaggregated models of each applied business are for characterizing wireless ginseng The model of mapping relations between number and user experience quality；The wireless parameter is included when the AP is communicated with user equipment The wireless environment parameter of wireless communication parameters and/or AP；
    The user experience quality assessment models are obtained according to two disaggregated models of each applied business.
  2. 2. the method as described in claim 1, it is characterised in that each is obtained according to the training data sample set and applies industry Training data sample set corresponding to business, including：
    The training data sample set is standardized so that the dimension of all data in the training data sample set It is unified；
    The training data sample of each applied business is determined from the training data sample set according to the priority of applied business This collection；Network quality of the priority of the applied business according to needed for the applied business is set；
    Data and user experience quality label in each training data sample included according to the training data sample set Correlation, the data of the predetermined quantity in each training data sample are selected, to obtain the final instruction of each applied business Practice set of data samples.
  3. 3. method as claimed in claim 2, it is characterised in that the priority according to applied business is from the training data The training data sample set of each applied business is determined in sample, including：
    Positive training data sample of the priority higher than the other application business of the first applied business is applied added to described first In the positive training data sample set of business, and by priority less than first applied business other application business negative instruction Practice data sample to be added in the negative training data sample set of first applied business；Positive training data sample set after addition The training data sample set of first applied business is formed with negative training data sample set；Wherein, the positive training data sample When this is more than or equal to default satisfaction threshold value for the satisfaction that user experience quality tag characterization user experiences to network Data, and the negative training data sample is that the satisfaction experienced to network of user experience quality tag characterization user is small Data when default satisfaction threshold value.
  4. 4. the method as described in claim 1, it is characterised in that described each applied business need the algorithm model that uses for Crossover algorithm is folded by k- and concentrates the minimum algorithm model of the error rate determined from the algorithm model prestored.
  5. 5. the method as described in claim 1, it is characterised in that according to obtaining two disaggregated models of each applied business User experience quality assessment models, including：
    Two disaggregated models of each applied business are combined according to the priority of applied business, to obtain the use Family Quality of experience assessment models.
  6. 6. the method as described in claim 1, it is characterised in that melted by two disaggregated models of each applied business Close, after obtaining user experience quality assessment models, methods described also includes：
    Acquisition updates the data sample set, and the sample set that updates the data updates the data sample, and the renewal including at least one Set of data samples is included in wireless communication parameters and AP when the AP uploaded in AP runnings is communicated with user equipment Wireless environment parameter；Wherein, one updates the data sample and includes at least one wireless communication parameters and at least one wireless ring Border parameter；
    Any one in sample set is updated the data described in calculating and updates the data sample and each in the training data sample set The similarity of training data sample, by the Consumer's Experience of similarity highest training data sample in the training data sample set Quality tab as it is described any one update the data the user experience quality label of sample, to obtain carrying user experience quality Label updates the data sample set；
    The user experience quality assessment models are entered according to the sample set that updates the data of the carrying user experience quality label Row renewal.
  7. 7. the method as described in claim 1, it is characterised in that methods described also includes：
    Receive the feedback message for the user experience quality difference that user equipment is sent；
    Determine whether the user experience quality assessment models detect network failure corresponding to the feedback message；
    If it is determined that the user experience quality assessment models are not detected by network failure corresponding to the feedback message, then institute is obtained State service data corresponding to feedback message；
    The user experience quality assessment models are corrected according to the service data.
  8. 8. the method as described in claim 1~7 is any, it is characterised in that according to the two of each applied business classification moulds After type obtains the user experience quality assessment models, methods described also includes：
    Current network supplemental characteristic of first user equipment when carrying out network experience is obtained, the network parametric data includes AP The current radio environment parameter of Current wireless communication parameter and AP when being communicated with first user equipment；
    According to the current network supplemental characteristic and the user experience quality assessment models, first user equipment is determined Active user's Quality of experience.
  9. 9. method as claimed in claim 8, it is characterised in that user experience quality is being passed through according to the network parametric data After assessment models determine user experience quality of the user equipment when carrying out network experience, methods described also includes：
    When it is determined that result shows active user's Quality of experience difference of first user equipment, pass through the user experience quality Assessment models calculate the weight of each network parametric data in the current network supplemental characteristic；The weight is the current net Influence degree of each network parametric data to the determination result in network supplemental characteristic；
    Determined to cause the user experience quality of first user equipment poor according to the weight of each network parametric data of calculating The reason for.
  10. 10. a kind of user experience quality assessment models establish equipment, it is characterised in that including：
    First data capture unit, for obtaining training data sample set, the training data sample set includes at least one instruction Practice data sample, and the training data sample set includes applying industry by different types of in different types of user equipment User experience quality label, AP and user when multiple AP diverse locations of the business in WLAN coverage carry out network experience set For the wireless communication parameters and AP wireless environment parameter when being communicated, wherein, a training data sample includes one User experience quality label, at least one wireless communication parameters and at least one wireless environment parameter；
    First data processing unit, for training number corresponding to obtaining each applied business according to the training data sample set According to sample set；
    Model establishes unit, for the training data sample set according to corresponding to each applied business and each applied business need The algorithm model to be used, establish two disaggregated models of each applied business, two classification moulds of each applied business Type is the model for characterizing mapping relations between wireless parameter and user experience quality；The wireless parameter includes the AP with using The wireless environment parameter of wireless communication parameters and/or AP when family equipment is communicated；
    The model is established unit and is additionally operable to, and the user experience quality is obtained according to two disaggregated models of each applied business Assessment models.
  11. 11. equipment as claimed in claim 10, it is characterised in that first data processing unit is according to the training data Sample set obtains training data sample set corresponding to each applied business, including：
    The training data sample set is standardized so that the dimension of all data in the training data sample set It is unified；
    The training data sample of each applied business is determined from the training data sample set according to the priority of applied business This collection；Network quality of the priority of the applied business according to needed for the applied business is set；
    Data and user experience quality label in each training data sample included according to the training data sample set Correlation, the data of the predetermined quantity in each training data sample are selected, to obtain the final instruction of each applied business Practice set of data samples.
  12. 12. equipment as claimed in claim 11, it is characterised in that first data processing unit is according to the excellent of applied business First level determines the training data sample set of each applied business from the training data sample, including：
    Positive training data sample of the priority higher than the other application business of the first applied business is applied added to described first In the positive training data sample set of business, and by priority less than first applied business other application business negative instruction Practice data sample to be added in the negative training data sample set of first applied business；Positive training data sample set after addition The training data sample set of first applied business is formed with negative training data sample set；Wherein, the positive training data sample When this is more than or equal to default satisfaction threshold value for the satisfaction that user experience quality tag characterization user experiences to network Data, and the negative training data sample is that the satisfaction experienced to network of user experience quality tag characterization user is small Data when default satisfaction threshold value.
  13. 13. equipment as claimed in claim 10, it is characterised in that each described applied business needs the algorithm model used Crossover algorithm is folded by k- for first data processing unit and concentrates the error rate determined from the algorithm model prestored Minimum algorithm model.
  14. 14. equipment as claimed in claim 10, it is characterised in that the model establishes unit according to each applied business Two disaggregated models obtain the user experience quality assessment models, including：
    The model establishes unit and is carried out two disaggregated models of each applied business according to the priority of applied business Combination, to obtain the user experience quality assessment models.
  15. 15. equipment as claimed in claim 10, it is characterised in that the equipment also includes the first model modification unit；
    The data capture unit is additionally operable to, and acquisition updates the data sample set, and the sample set that updates the data is including at least one Update the data sample, and it is described update the data sample set and be included in the AP uploaded in AP runnings communicated with user equipment When wireless communication parameters and AP wireless environment parameter；Wherein, one updates the data sample and includes at least one channel radio Believe parameter and at least one wireless environment parameter；
    First data processing unit is additionally operable to, updated the data described in calculating any one in sample set update the data sample with The similarity of each training data sample in the training data sample set, by similarity in the training data sample set most The user experience quality label of high training data sample as it is described any one update the data the user experience quality of sample Label, sample set is updated the data to obtain carrying user experience quality label；
    The first model modification unit is used to update the data sample set to institute according to the carrying user experience quality label User experience quality assessment models are stated to be updated.
  16. 16. equipment as claimed in claim 10, it is characterised in that the equipment also includes the second model modification unit；
    The data capture unit is additionally operable to, and receives the feedback message for the user experience quality difference that user equipment is sent；
    First data processing unit is additionally operable to, and determines whether the user experience quality assessment models detect the feedback Network failure corresponding to message；If it is determined that the user experience quality assessment models are not detected by net corresponding to the feedback message Network failure, then obtain service data corresponding to the feedback message；
    The second model modification unit is used for, and the user experience quality assessment models are rectified according to the service data Just.
  17. 17. the equipment as described in claim 10~16 is any, it is characterised in that including：
    Second data capture unit, for obtaining current network supplemental characteristic of first user equipment when carrying out network experience, The network parametric data includes Current wireless communication parameters and AP of the AP when being communicated with first user equipment Current radio environment parameter；
    Assessment unit, for according to the current network supplemental characteristic and the user experience quality assessment models, it is determined that described Active user's Quality of experience of first user equipment.
  18. 18. equipment as claimed in claim 17, it is characterised in that the equipment also includes the second data processing unit；
    Second data processing unit is used for, when it is determined that result shows active user's Quality of experience of first user equipment When poor, each network parametric data in the current network supplemental characteristic is calculated by the user experience quality assessment models Weight；Influence journey of the weight for each network parametric data in the current network supplemental characteristic to the determination result Degree；Determined to cause the user experience quality difference of first user equipment according to the weight of each network parametric data of calculating Reason.
  19. 19. a kind of computer installation, it is characterised in that described device includes processor, and the processor is used to perform memory Realized during the computer program of middle storage as any one of claim 1-9 the step of method.
  20. 20. a kind of computer-readable recording medium, is stored thereon with computer program, it is characterised in that：The computer program Realized when being executed by processor as any one of claim 1-9 the step of method.
CN201710933787.0A 2017-10-10 2017-10-10 User experience quality assessment model establishing method and device Active CN107733705B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710933787.0A CN107733705B (en) 2017-10-10 2017-10-10 User experience quality assessment model establishing method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710933787.0A CN107733705B (en) 2017-10-10 2017-10-10 User experience quality assessment model establishing method and device

Publications (2)

Publication Number Publication Date
CN107733705A true CN107733705A (en) 2018-02-23
CN107733705B CN107733705B (en) 2021-01-15

Family

ID=61208622

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710933787.0A Active CN107733705B (en) 2017-10-10 2017-10-10 User experience quality assessment model establishing method and device

Country Status (1)

Country Link
CN (1) CN107733705B (en)

Cited By (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108550054A (en) * 2018-04-12 2018-09-18 百度在线网络技术（北京）有限公司 A kind of content quality appraisal procedure, device, equipment and medium
CN109451300A (en) * 2018-11-12 2019-03-08 中国联合网络通信集团有限公司 The determination method and apparatus of video quality score
CN109598285A (en) * 2018-10-24 2019-04-09 阿里巴巴集团控股有限公司 A kind of processing method of model, device and equipment
CN109905696A (en) * 2019-01-09 2019-06-18 浙江大学 A kind of recognition methods of the Video service Quality of experience based on encryption data on flows
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
CN110083542A (en) * 2019-05-06 2019-08-02 百度在线网络技术（北京）有限公司 Model test Method, device and electronic equipment in a kind of recommender system
CN110335058A (en) * 2019-04-30 2019-10-15 中国联合网络通信集团有限公司 A kind of sample generating method and device of user satisfaction prediction model
CN111339748A (en) * 2020-02-17 2020-06-26 北京声智科技有限公司 Analytical model evaluation method, analytical model evaluation device, analytical model evaluation equipment and analytical model evaluation medium
CN112383828A (en) * 2019-12-12 2021-02-19 致讯科技(天津)有限公司 Experience quality prediction method, equipment and system with brain-like characteristic
CN112469071A (en) * 2020-11-16 2021-03-09 成都渊数科技有限责任公司 WiFi network quality evaluation method and system
CN112636976A (en) * 2020-12-23 2021-04-09 武汉船舶通信研究所（中国船舶重工集团公司第七二二研究所） Service quality determination method, device, electronic equipment and storage medium
CN113676341A (en) * 2020-05-15 2021-11-19 华为技术有限公司 Quality difference evaluation method and related equipment
CN114389723A (en) * 2020-10-16 2022-04-22 展讯通信（上海）有限公司 Communication method, device and equipment

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102143507A (en) * 2010-11-23 2011-08-03 北京中创信测科技股份有限公司 Method and system for monitoring service quality, and analytical method and system therefor
CN103856959A (en) * 2012-11-30 2014-06-11 华为技术有限公司 Wireless local area network quality monitoring method, device and system
US20140362713A1 (en) * 2013-06-11 2014-12-11 Seven Networks, Inc. Quality of experience enhancement for wireless networks based on received signal strength at a mobile device
CN104702666A (en) * 2015-01-30 2015-06-10 北京邮电大学 User experience quality confirmation method and system
WO2015144211A1 (en) * 2014-03-25 2015-10-01 Telefonaktiebolaget L M Ericsson (Publ) Method and system for monitoring qoe
CN105264907A (en) * 2013-12-30 2016-01-20 华为技术有限公司 Method for predicting quality of experience of mobile video service, and base station
CN106230624A (en) * 2016-07-25 2016-12-14 中国联合网络通信集团有限公司 A kind of network quality appraisal procedure and device
CN107087160A (en) * 2017-04-28 2017-08-22 南京邮电大学 A kind of Forecasting Methodology of the user experience quality based on BP Adaboost neutral nets
US20170244777A1 (en) * 2016-02-19 2017-08-24 Verizon Patent And Licensing Inc. Application quality of experience evaluator for enhancing subjective quality of experience
EP3226472A1 (en) * 2016-04-01 2017-10-04 Thomson Licensing Method for predicting a level of qoe of an application intended to be run on a wireless user equipment

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102143507A (en) * 2010-11-23 2011-08-03 北京中创信测科技股份有限公司 Method and system for monitoring service quality, and analytical method and system therefor
CN103856959A (en) * 2012-11-30 2014-06-11 华为技术有限公司 Wireless local area network quality monitoring method, device and system
US20140362713A1 (en) * 2013-06-11 2014-12-11 Seven Networks, Inc. Quality of experience enhancement for wireless networks based on received signal strength at a mobile device
CN105264907A (en) * 2013-12-30 2016-01-20 华为技术有限公司 Method for predicting quality of experience of mobile video service, and base station
WO2015144211A1 (en) * 2014-03-25 2015-10-01 Telefonaktiebolaget L M Ericsson (Publ) Method and system for monitoring qoe
CN104702666A (en) * 2015-01-30 2015-06-10 北京邮电大学 User experience quality confirmation method and system
US20170244777A1 (en) * 2016-02-19 2017-08-24 Verizon Patent And Licensing Inc. Application quality of experience evaluator for enhancing subjective quality of experience
EP3226472A1 (en) * 2016-04-01 2017-10-04 Thomson Licensing Method for predicting a level of qoe of an application intended to be run on a wireless user equipment
CN106230624A (en) * 2016-07-25 2016-12-14 中国联合网络通信集团有限公司 A kind of network quality appraisal procedure and device
CN107087160A (en) * 2017-04-28 2017-08-22 南京邮电大学 A kind of Forecasting Methodology of the user experience quality based on BP Adaboost neutral nets

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
张乘铭: "无线泛在环境下QoE评价及优化理论研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *

Cited By (21)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108550054A (en) * 2018-04-12 2018-09-18 百度在线网络技术（北京）有限公司 A kind of content quality appraisal procedure, device, equipment and medium
CN108550054B (en) * 2018-04-12 2022-10-14 百度在线网络技术（北京）有限公司 Content quality evaluation method, device, equipment and medium
CN109598285A (en) * 2018-10-24 2019-04-09 阿里巴巴集团控股有限公司 A kind of processing method of model, device and equipment
CN109451300A (en) * 2018-11-12 2019-03-08 中国联合网络通信集团有限公司 The determination method and apparatus of video quality score
CN109905696A (en) * 2019-01-09 2019-06-18 浙江大学 A kind of recognition methods of the Video service Quality of experience based on encryption data on flows
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
CN110335058B (en) * 2019-04-30 2021-09-14 中国联合网络通信集团有限公司 Sample generation method and device of user satisfaction prediction model
CN110335058A (en) * 2019-04-30 2019-10-15 中国联合网络通信集团有限公司 A kind of sample generating method and device of user satisfaction prediction model
CN110083542A (en) * 2019-05-06 2019-08-02 百度在线网络技术（北京）有限公司 Model test Method, device and electronic equipment in a kind of recommender system
CN110083542B (en) * 2019-05-06 2023-11-07 百度在线网络技术（北京）有限公司 Model testing method and device in recommendation system and electronic equipment
CN112383828A (en) * 2019-12-12 2021-02-19 致讯科技(天津)有限公司 Experience quality prediction method, equipment and system with brain-like characteristic
CN112383828B (en) * 2019-12-12 2023-04-25 致讯科技(天津)有限公司 Quality of experience prediction method, equipment and system with brain-like characteristics
CN111339748A (en) * 2020-02-17 2020-06-26 北京声智科技有限公司 Analytical model evaluation method, analytical model evaluation device, analytical model evaluation equipment and analytical model evaluation medium
CN111339748B (en) * 2020-02-17 2023-11-17 北京声智科技有限公司 Evaluation method, device, equipment and medium of analytical model
CN113676341A (en) * 2020-05-15 2021-11-19 华为技术有限公司 Quality difference evaluation method and related equipment
CN113676341B (en) * 2020-05-15 2022-10-04 华为技术有限公司 Quality difference evaluation method and related equipment
US11489904B2 (en) 2020-05-15 2022-11-01 Huawei Technologies Co., Ltd. Poor-QoE assessment method and related device
CN114389723A (en) * 2020-10-16 2022-04-22 展讯通信（上海）有限公司 Communication method, device and equipment
CN114389723B (en) * 2020-10-16 2023-09-26 展讯通信（上海）有限公司 Communication method, device and equipment
CN112469071A (en) * 2020-11-16 2021-03-09 成都渊数科技有限责任公司 WiFi network quality evaluation method and system
CN112636976A (en) * 2020-12-23 2021-04-09 武汉船舶通信研究所（中国船舶重工集团公司第七二二研究所） Service quality determination method, device, electronic equipment and storage medium

Also Published As

Publication number Publication date
CN107733705B (en) 2021-01-15

Similar Documents

Publication Publication Date Title
CN107733705A (en) A kind of user experience quality assessment models method for building up and equipment
CN104392390B (en) A kind of secondary equipment of intelligent converting station appraisal procedure based on TOPSIS models
CN109934704A (en) Information recommendation method, device, equipment and storage medium
CN106484622A (en) Method of testing, apparatus and system
CN107390056B (en) Detect the method and system of charging pile
US10609570B2 (en) Method and system for optimizing network parameters to improve customer satisfaction of network content
CN104702666A (en) User experience quality confirmation method and system
US20190312791A1 (en) Method and system for automating assessment of network quality of experience
Anchuen et al. QOE model in cellular networks based on QOS measurements using Neural Network approach
CN108984516B (en) Online course content evaluation method and system based on bullet screen evaluation cloud data
US11178024B2 (en) Method and scoring node for estimating a user's quality of experience for a delivered service
CN108599995A (en) Network line failure judgment method and server
CN104427547B (en) Business and network associate method of testing, apparatus and system
CN107992410A (en) Software quality monitoring method, device, computer equipment and storage medium
CN107943697A (en) Problem distribution method, device, system, server and computer-readable storage medium
CN106789391A (en) A kind of automated testing method and device of router DHCP functions
CN107483267A (en) A kind of EIGRP routing failures recognition methods
CN113678494A (en) Method and device for providing network experience testing
CN110070452A (en) Model training method, calculates equipment and computer readable storage medium at device
JP2015106220A (en) Sensory communication quality estimation device and sensory communication quality estimation program
CN105915374A (en) Online game type business perceived quality model establishing method
CN102300238A (en) Client perception system and method for broadband network
CN117560267A (en) Fault diagnosis method and rapid recovery method for multimode field network
CN106648927B (en) Abnormal information collection method and server
Liu et al. KQis-driven QoE anomaly detection and root cause analysis in cellular networks

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