CN106384197A - Service quality evaluation method and device based on big data - Google Patents
Service quality evaluation method and device based on big data Download PDFInfo
- Publication number
- CN106384197A CN106384197A CN201610818704.9A CN201610818704A CN106384197A CN 106384197 A CN106384197 A CN 106384197A CN 201610818704 A CN201610818704 A CN 201610818704A CN 106384197 A CN106384197 A CN 106384197A
- Authority
- CN
- China
- Prior art keywords
- quality
- feature
- industry
- characteristic
- business
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/211—Selection of the most significant subset of features
- G06F18/2113—Selection of the most significant subset of features by ranking or filtering the set of features, e.g. using a measure of variance or of feature cross-correlation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Educational Administration (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Development Economics (AREA)
- Evolutionary Computation (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Strategic Management (AREA)
- Game Theory and Decision Science (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a service quality evaluation method and device based on big data. The method comprises the steps: feature data related to the service is acquired, and the storage mode of the feature data is determined according to the service features; preprocessing is performed on the feature data; a basic marking model of the service quality is determined, and a feature weight vector in the basic marking model is to-be-determined; an industry weighting coefficient is determined for each feature of the feature data; the feature data is sampled according to industry, and training samples are obtained; a machine learning stochastic gradient descent algorithm is adopted to the training samples to obtain an optimal feature weight vector which can achieve local minimum in the error function of the machine learning stochastic gradient descent algorithm; the industry weighting coefficient and the optimal feature weight vector obtained by the machine learning are combined to obtain a final service quality model; and according to a quality ranking of different services organized manually, quantitative evaluation is performed on the final service quality model.
Description
Technical field
The present invention relates to machine learning field is and in particular to a kind of method of the evaluation the quality based on big data and dress
Put.
Background technology
With the aggravation of the level of informatization, big data industry is arisen at the historic moment and development speed is surprising and has 4V characteristic,
I.e. Volume (a large amount of), Velocity (at a high speed), Variety (various), Value (value).According to authoritative institution's prediction, the whole world is big
Data Market scale is expected to reach 53,000,000,000 dollars in 2017.With deepening continuously of big data application, it is also applied to commonly
Internet firm in, such as:The aspects such as business intelligence, data operation.Due to the characteristic of Internet firm, the operation shape of business
Condition can by the quality on line (for example:Day any active ues etc.) synthetically reflect, but in prior art, lack how base
Method and apparatus in mass data accurate evaluation quality of service.
Content of the invention
In view of this, embodiments provide a kind of evaluation the quality method and apparatus based on big data, with
More accurately assess quality of service.
A kind of first aspect, there is provided evaluation the quality method based on big data, including:
Step one, obtains the characteristic related to business, and determines described characteristic according to the feature of business itself
Storage mode;
Step 2, carries out pretreatment to described characteristic;
Step 3, determines the basic scoring model of quality of service, and the feature weight vector in the scoring model of described basis is
To be determined;Industry weight coefficient is determined to each feature in described characteristic;Described characteristic is taken out by industry
After sample, obtain training sample;Machine learning stochastic gradient descent algorithm is used to obtain optimal characteristics weight described training sample
Vector, described optimal characteristics weight vectors meet the error function Local Minimum of described machine learning stochastic gradient descent algorithm;
The described optimal characteristics weight vectors obtaining in conjunction with described industry weight coefficient and machine learning obtain final service quality model:
Step 4, the quality ranking of the different business being gone out according to manual sorting, to the described final industry obtaining in step 3
Business quality model carries out quantitative evaluation.
In conjunction with a first aspect, in the first possible implementation method of first aspect, described step 2 includes:To described
Characteristic carries out data scrubbing, data integration, data normalization data stipulations.
In conjunction with the first possible implementation method of first aspect or first aspect, possible in the second of first aspect
In implementation method, the basic scoring model of described quality of service is h (X)=θ X=θ1x1+θ2x2+…+θnxn, wherein, h (X) generation
Table quality of service, X=(x1..., xn) vectorial for input feature vector value, n represents n-th feature, θnRepresent the weight of n-th feature
Value, xnRepresent the input feature vector value of n-th feature, θ=(θ1..., θn) it is feature weight vector to be determined.
In conjunction with the possible implementation method of any of the above, in the third possible implementation method of first aspect, institute
State and industry weight coefficient is determined to each feature in characteristic, including:For different industries, manual sorting goes out to meet difference
The industry weight coefficient of industrial characteristic, the industry weight coefficient being gone out according to described manual sorting, determine in described characteristic
The industry weight coefficient of each feature.
In conjunction with the possible implementation method of any of the above, in the 4th kind of possible implementation method of first aspect, institute
State and characteristic is sampled by industry, including:Different industries is sampled according to predetermined ratio, according to concrete business
Emphasis adjust industry shared by training sample ratio.
In conjunction with the possible implementation method of the second of first aspect, in the 5th kind of possible implementation method of first aspect
In, the error function of described machine learning stochastic gradient descent algorithm isWherein θ is to treat really
Fixed feature weight vector, the sum of p representative sample, h (Xi) represent the discreet value of i-th sample, yiRepresent i-th sample
Actual value;
Described to training sample use machine learning stochastic gradient descent algorithm obtain optimal characteristics weight vectors, including:
(1) random initializtion θ vector;(2) learning rate that setting declines(3) value of θ is updated toJ (θ) is become
Little, whereinIt is the partial derivative to θ for the J;(4) until J (θ) convergence, after convergence, the value of θ is weighed as optimal characteristics for circulation execution (3)
Weight vector θ *, it meets described error function Local Minimum.
In conjunction with the 5th kind of possible implementation method of first aspect, in the 6th kind of possible implementation method of first aspect
In, the optimal characteristics weight vectors that described combination industry weight coefficient is obtained with machine learning obtain final service quality model,
Including:According to described, each feature in described characteristic is determined under industry m obtaining in industry weight coefficient step
Parameter θ 'mWith described optimal characteristics weight vectors θ *, and input feature vector value vector X=(x1..., xn), draw final service
Quality model is:h*(X)=θ*·θ′m·XT.
In conjunction with the possible implementation method of any of the above, in the 7th kind of possible implementation method of first aspect, its
It is characterised by, described step 4 is specially:Manual sorting goes out the quality ranking of different business, true according to step 3 to each business
Fixed final service quality model calculates quality of service, then carries out the mass value of the mass value of each business and other business
Relatively, if comparative result is consistent with the quality of service ranking that manual sorting goes out, for positive example, otherwise for negative example, positive example total
Quantity is P, and the total quantity of negative example is N, then evaluation quality L is defined as:
A kind of second aspect, there is provided evaluation the quality device based on big data, including:
First module, for obtaining the characteristic related to business, and determines described spy according to the feature of business itself
Levy the storage mode of data;
Second module, for carrying out pretreatment to described characteristic;
Three module, for determining the basic scoring model of quality of service, the feature weight in the scoring model of described basis
Vector is to be determined;Industry weight coefficient is determined to each feature in described characteristic;Industry is pressed to described characteristic
After being sampled, obtain training sample;Described training sample is used machine learning stochastic gradient descent algorithm obtain optimum special
Levy weight vectors, described optimal characteristics weight vectors meet the error function local of described machine learning stochastic gradient descent algorithm
Minimum;The described optimal characteristics weight vectors obtaining in conjunction with described industry weight coefficient and machine learning obtain final service quality
Model;
4th module, the quality ranking of the different business for being gone out according to manual sorting, to described in three module acquisition
Final service quality model carries out quantitative evaluation.
A kind of third aspect, there is provided evaluation the quality device based on big data, including:
Memorizer, have program stored therein in described memorizer instruction;
At least one processor, for executing described program instruction;
Described program instruction by during described computing device so that the method for described computing device first aspect.
Beneficial effects of the present invention are as follows:
The present invention according to the feature of different industries it is intended that the sampling proportion of training sample so that machine learning algorithm
More accurately more stable evaluation the quality model can be trained.It has been simultaneously introduced the feature weight correcting mode by industry,
Ensure that the quality of service of model evaluation can be contrasted between industry.
Brief description
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
Have technology description in required use accompanying drawing be briefly described it should be apparent that, drawings in the following description be only this
Some embodiments of invention, for those of ordinary skill in the art, on the premise of not paying creative work, acceptable
Other accompanying drawings are obtained according to these accompanying drawings.
Fig. 1 is a kind of flow chart of evaluation the quality method based on big data provided in an embodiment of the present invention;
Fig. 2 is the example flow diagram of characteristic pretreatment provided in an embodiment of the present invention;
Fig. 3 is the flow chart determining quality of service model provided in an embodiment of the present invention.
Specific embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Site preparation description is it is clear that described embodiment is only a part of embodiment of the present invention, rather than whole embodiments.It is based on
Embodiment in the present invention, those of ordinary skill in the art obtained on the premise of not making creative work all its
His embodiment, broadly falls into the scope of protection of the invention.
Fig. 1 is a kind of flow chart of evaluation the quality method based on big data provided in an embodiment of the present invention, including
Following steps:
S101, the acquisition characteristic related to business, and described characteristic is determined according to the feature of business itself
Storage mode.
Wherein, the mode obtaining the characteristic related to business includes:1) associated nets are obtained by data crawlers
Stand data;2) own service data accumulation;3) third party's data company provides data.To protect during obtaining characteristic
The accuracy of card overall data is with real-time it is ensured that timing data being capable of continuous updating.
Because different business has different characteristics, such as business can be divided into computation-intensive, I/O intensive type etc., so not
The business of same type has different storage organization Technology Selections.It is thus desirable to according to the feature of business itself, selecting different
Characteristic storage mode, carries out efficient storage, renewal, modification and the purpose applied to reach to the characteristic getting.
Wherein, it is not very big for data volume and updates business infrequently, can be using the file storage characteristic number related to this business
According to;For high access frequency and the business slightly higher to requirement of real-time, optional memory database (for example, redis etc.) storage
The characteristic related to this business;But data volume slightly larger, data structure complicated business general for access frequency, may be selected
Relevant database (for example, mysql etc.) the storage characteristic related to this business;Very big (for example, every time for data volume
Data updates magnitude in more than GB) and the business not high to requirement of real-time, optional distributed data processing warehouse is (for example,
Hadoop etc.) the storage characteristic related to this business.
S102, characteristic pretreatment.This step is used for the characteristic obtaining described in S101 step and store is entered
The pretreatment such as row data scrubbing, data integration, data normalization, hough transformation, to obtain the structure more complete, accuracy is higher
Change characteristic, be conducive to subsequently carrying out more accurately quality evaluation.
S103, determine the basic scoring model of quality of service, the feature weight vector in the scoring model of described basis is for treating
Determine;Industry weight coefficient is determined to each feature in described characteristic;Described characteristic is sampled by industry
After obtain training sample;To described training sample use machine learning stochastic gradient descent algorithm obtain optimal characteristics weight to
Amount, it meets the error function Local Minimum of described machine learning stochastic gradient descent algorithm;In conjunction with described industry weight coefficient
Obtain final service quality model with the described optimal characteristics weight vectors that machine learning obtains.
S104, the quality ranking of the different business being gone out according to manual sorting, to the final service matter obtaining in S103 step
Amount model carries out quantitative evaluation.In this step, manual sorting go out different business quality ranking (only focus on order, for example:Before
100 etc.).According to the final service quality model that S103 determines, quality of service is calculated to each business, then by each business
Mass value is compared with the mass value of other business, if comparative result is consistent with the quality of service ranking of manual sorting,
For positive example, otherwise for bearing example, the total quantity of positive example is P, and the total quantity of negative example is N, then evaluation quality L is defined as:
Can there is the assessment of a quantization according to evaluation quality L to the effect of final service quality model, thus being follow-up
Optimize the standard that basis is provided.
Fig. 2 is the example flow diagram of characteristic pretreatment provided in an embodiment of the present invention.As shown in Fig. 2 characteristic
Pretreatment specifically may include following steps:
S201, data scrubbing.Data scrubbing includes removing noise data and extraneous data, processes missing data, identification simultaneously
Delete isolated point etc. to process.For example, for noise data, can be sentenced according to the extent of deviation with expected value (or meansigma methodss)
Disconnected.Isolated point is that this distance to be measured by given threshold value apart from far point with the meansigma methodss of corresponding stochastic variable, leads to
It is often the integral multiple of standard deviation.For missing data, using default value (difference be might have according to the business implication of data), all
Value, mode, median etc. carry out completion process.Carry out filtering removal for repeating the noise datas such as record.
S202, data integration.Data entity expresses possibility incomplete same it is also possible to lead in the data of different data sources
Cause the problem of data redundancy, need to carry out redundancy deletion data integration as the case may be.
S203, data normalization.Namely dimensionless process is carried out to characteristic, the value making characteristic is specified
In span.
S204, hough transformation.Stipulations are carried out to the mass data in data base, for example, Data Dimensionality Reduction, restriction data value
Scope etc..Data after stipulations remains close to keep the integrity of former data, but relatively small many of data volume, so dug
The performance of pick and efficiency can be greatly improved.
Fig. 3 is the flow chart determining quality of service model provided in an embodiment of the present invention, and concrete steps are described in detail as follows:
S301, determine the basic scoring model of quality of service.
In one example, using scoring model based on linear model:
H (X)=θ X=θ1x1+θ2x2+…+θnxn
Wherein, h (X) represents quality of service, X=(x1..., xn) vectorial for input feature vector value, n represents n-th feature, θn
Represent the weighted value of n-th feature, xnRepresent the input feature vector value of n-th feature, θ=(θ1..., θn) it is feature to be determined
Weight vectors.
S302, determine the industry weight coefficient of each feature.For different industries, manual sorting goes out a set of to meet the sector
The weight coefficient of feature.Assume that industry is expressed as m, then the weight vector of industry m is expressed as θ 'm.
S303, the characteristic to acquisition are sampled by industry, obtain training sample.To different industries according to prior
Designated ratio is sampled, can according to shared by the emphasis of concrete business adjusts industry training sample ratio.Its benefit exists
In:For different business, the impact effect of different industries can be controlled.
S304, using machine learning stochastic gradient descent algorithm obtain optimal characteristics weight vectors.In order to draw in S301
The feature weight vector to be determined of basic scoring model, needs using the training sample in S303, using machine learning boarding steps
Degree descent algorithm is learnt.Described stochastic gradient descent algorithm needs to build error function Wherein θ is characterized weight vectors, the sum of p representative sample, h (Xi) represent the discreet value of i-th sample, yiRepresent i-th
The actual value of bar sample.Accordingly, it would be desirable to ask meet the θ that J (θ) is minimum for optimal characteristics weight vectors.S304 specifically may include
Following steps:
1) random initializtion θ vector.
2) learning rate that setting declines
3) value of θ is updated toJ (θ) is diminished.WhereinFor setting learning rate,It is J to θ
Partial derivative, i.e. the gradient direction of J (θ).
4) circulate execution 3) until J (θ) convergence, after convergence, as optimal characteristics weight vectors θ *, it meets error to the value of θ
Function Local Minimum.
S305, the optimal characteristics weight vectors obtaining with reference to industry weight coefficient and machine learning obtain final service quality
Model.According to the parameter θ ' under industry m that S302 obtainsmWith the optimal characteristics weight vectors θ * drawing in S304, and defeated
Enter feature value vector X=(x1..., xn) it can be deduced that final service quality model is:
h*(X)=θ*·θ′m·XT
The embodiment of the present invention additionally provides a kind of evaluation the quality device based on big data, can be used for executing aforementioned
Method in embodiment and described in accompanying drawing 1-3.The described evaluation the quality device based on big data, including:
First module, for obtaining the characteristic related to business, and determines described spy according to the feature of business itself
Levy the storage mode of data;
Second module, for carrying out pretreatment to described characteristic;
Three module, for determining the basic scoring model of quality of service, the feature weight in the scoring model of described basis
Vector is to be determined;Industry weight coefficient is determined to each feature in described characteristic;Industry is pressed to described characteristic
After being sampled, obtain training sample;Described training sample is used machine learning stochastic gradient descent algorithm obtain optimum special
Levy weight vectors, described optimal characteristics weight vectors meet the error function local of described machine learning stochastic gradient descent algorithm
Minimum;The described optimal characteristics weight vectors obtaining in conjunction with described industry weight coefficient and machine learning obtain final service quality
Model;
4th module, the quality ranking of the different business for being gone out according to manual sorting, to described in three module acquisition
Final service quality model carries out quantitative evaluation.
The embodiment of the present invention additionally provides a kind of evaluation the quality device based on big data, can be used for executing aforementioned
Method in embodiment and described in accompanying drawing 1-3.Described memorizer and at least is included based on the evaluation the quality device of big data
One processor, have program stored therein in described memorizer instruction, and at least one processor described refers to for executing described program
Order, described program instruction by during described computing device so that in described computing device previous embodiment and accompanying drawing 1-3 described in
Method.
Those skilled in the art can be understood that, for convenience and simplicity of description, the system of foregoing description,
Device and the specific work process of unit, may be referred to the corresponding process in preceding method embodiment, will not be described here.
Those of ordinary skill in the art are it is to be appreciated that combine the mould of each example of the embodiments described herein description
Block and method and step, being capable of being implemented in combination in electronic hardware or computer software and electronic hardware.These functions are actually
To be executed with hardware or software mode, the application-specific depending on technical scheme and design constraint.Professional and technical personnel
Each specific application can be used different methods to realize described function, but this realization is it is not considered that exceed
The scope of the present invention.
One of ordinary skill in the art will appreciate that it is permissible for realizing all or part of step in above-described embodiment method
Completed by program class instruction processing unit.If described function is realized and as independent product using in the form of SFU software functional unit
When selling or using, can be stored in a computer read/write memory medium.Based on such understanding, the technology of the present invention
Part that scheme substantially contributes to prior art in other words or this technical scheme partly can be with software product
Form embodies, and this computer software product is stored in a storage medium, including some instructions with so that one is counted
Calculate machine equipment (can be personal computer, server, or network equipment etc.) execution each embodiment methods described of the present invention
All or part of step.And aforesaid storage medium includes:USB flash disk, portable hard drive, read only memory (ROM, Read-Only
Memory), random access memory (RAM, Random Access Memory), magnetic disc or CD etc. are various can store journey
The medium of sequence code.
More than, the specific embodiment of the only present invention, but protection scope of the present invention is not limited thereto, any it is familiar with
Those skilled in the art the invention discloses technical scope in, change or replacement can be readily occurred in, all should cover
Within protection scope of the present invention.Therefore, protection scope of the present invention should be defined by scope of the claims.
Claims (10)
1. a kind of evaluation the quality method based on big data is it is characterised in that include:
Step one, obtains the characteristic related to business, and determines depositing of described characteristic according to the feature of business itself
Storage mode;
Step 2, carries out pretreatment to described characteristic;
Step 3, determines the basic scoring model of quality of service, the feature weight vector in the scoring model of described basis is for treating really
Fixed;Industry weight coefficient is determined to each feature in described characteristic;After described characteristic is sampled by industry,
Obtain training sample;Machine learning stochastic gradient descent algorithm is used to obtain optimal characteristics weight vectors described training sample,
Described optimal characteristics weight vectors meet the error function Local Minimum of described machine learning stochastic gradient descent algorithm;In conjunction with institute
State industry weight coefficient and obtain final service quality model with the described optimal characteristics weight vectors that machine learning obtains;
Step 4, the quality ranking of the different business being gone out according to manual sorting, to the described final service matter obtaining in step 3
Amount model carries out quantitative evaluation.
2. the evaluation the quality method based on big data as claimed in claim 1 is it is characterised in that described step 2 bag
Include:Described characteristic is carried out with data scrubbing, data integration, data normalization data stipulations.
3. the evaluation the quality method based on big data as claimed in claim 1 or 2 is it is characterised in that described business matter
The basic scoring model of amount is h (X)=θ X=θ1x1+θ2x2+…+θnxn, wherein, h (X) represents quality of service, X=(x1...,
xn) vectorial for input feature vector value, n represents n-th feature, θnRepresent the weighted value of n-th feature, xnRepresent the defeated of n-th feature
Enter eigenvalue, θ=(θ1..., θn) it is feature weight vector to be determined.
4. as claim 1-3 any one described evaluation the quality method based on big data it is characterised in that described right
Each feature in characteristic determines industry weight coefficient, including:For different industries, manual sorting goes out to meet different industries
The industry weight coefficient of feature, the industry weight coefficient being gone out according to described manual sorting, determine each in described characteristic
The industry weight coefficient of feature.
5. as claim 1-4 any one described evaluation the quality method based on big data it is characterised in that described right
Characteristic is sampled by industry, including:Different industries is sampled according to predetermined ratio, according to the side of concrete business
Emphasis adjusts the ratio of training sample shared by industry.
6. the evaluation the quality method based on big data as claimed in claim 3 it is characterised in that described machine learning with
The error function of machine gradient descent algorithm isWherein θ is feature weight vector to be determined,
The sum of p representative sample, h (Xi) represent the discreet value of i-th sample, yiRepresent the actual value of i-th sample;
Described to training sample use machine learning stochastic gradient descent algorithm obtain optimal characteristics weight vectors, including:(1) with
Machine initialization θ vector;(2) learning rate that setting declines(3) value of θ is updated toJ (θ) is diminished, whereinIt is the partial derivative to θ for the J;(4) until J (θ) convergence, after convergence, the value of θ is as optimal characteristics weight vectors for circulation execution (3)
θ *, it meets described error function Local Minimum.
7. the evaluation the quality method based on big data as claimed in claim 6 is it is characterised in that described combination industry adds
The optimal characteristics weight vectors that weight coefficient is obtained with machine learning obtain final service quality model, including:According to described to institute
State each feature in characteristic and determine the parameter θ ' under industry m obtaining in industry weight coefficient stepmWith described optimum
Feature weight vector θ *, and input feature vector value vector X=(x1..., xn), show that final service quality model is:h*(X)=
θ*·θ′m·XT.
8. as claim 1-7 any one described evaluation the quality method based on big data it is characterised in that described step
Rapid four are specially:Manual sorting goes out the quality ranking of different business, the final service matter that each business is determined according to step 3
Amount model calculates quality of service, is then compared the mass value of the mass value of each business and other business, if compared
Result is consistent with the quality of service ranking that manual sorting goes out, then for positive example, otherwise for bearing example, the total quantity of positive example is P, negative example
Total quantity is N, then evaluation quality L is defined as:
9. a kind of evaluation the quality device based on big data is it is characterised in that described commented based on the quality of service of big data
Estimate device to include:
First module, for obtaining the characteristic related to business, and determines described characteristic number according to the feature of business itself
According to storage mode;
Second module, for carrying out pretreatment to described characteristic;
Three module, for determining the basic scoring model of quality of service, the feature weight vector in the scoring model of described basis
For to be determined;Industry weight coefficient is determined to each feature in described characteristic;Described characteristic is carried out by industry
After sampling, obtain training sample;Described training sample is used machine learning stochastic gradient descent algorithm obtain optimal characteristics power
Weight vector, described optimal characteristics weight vectors meet the error function local of described machine learning stochastic gradient descent algorithm
Little;The described optimal characteristics weight vectors obtaining in conjunction with described industry weight coefficient and machine learning obtain final service quality mould
Type;
4th module, the quality ranking of the different business for being gone out according to manual sorting, what three module was obtained is described final
Quality of service model carries out quantitative evaluation.
10. a kind of evaluation the quality device based on big data is it is characterised in that described commented based on the quality of service of big data
Estimate device to include:
Memorizer, have program stored therein in described memorizer instruction;
At least one processor, for executing described program instruction;
Described program instruction by during described computing device so that method as described in claim 1-8 for the described computing device.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610818704.9A CN106384197A (en) | 2016-09-13 | 2016-09-13 | Service quality evaluation method and device based on big data |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610818704.9A CN106384197A (en) | 2016-09-13 | 2016-09-13 | Service quality evaluation method and device based on big data |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106384197A true CN106384197A (en) | 2017-02-08 |
Family
ID=57936469
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610818704.9A Pending CN106384197A (en) | 2016-09-13 | 2016-09-13 | Service quality evaluation method and device based on big data |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106384197A (en) |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107563595A (en) * | 2017-08-02 | 2018-01-09 | 卓智网络科技有限公司 | Colleges and universities' core business index scoring system and method based on dynamic discrete algorithm |
CN107909472A (en) * | 2017-12-08 | 2018-04-13 | 上海壹账通金融科技有限公司 | Management data checking method, device, equipment and computer-readable recording medium |
CN108287928A (en) * | 2018-03-05 | 2018-07-17 | 四川易利数字城市科技有限公司 | A kind of space attribute prediction technique based on local weighted linear regression |
CN109144986A (en) * | 2018-07-30 | 2019-01-04 | 上海电气集团股份有限公司 | A kind of importance appraisal procedure of industrial equipment data |
CN110580490A (en) * | 2018-06-11 | 2019-12-17 | 杭州海康威视数字技术股份有限公司 | Method, device and equipment for determining personnel behavior probability |
CN110781456A (en) * | 2019-09-27 | 2020-02-11 | 上海麦克风文化传媒有限公司 | Sorting weight updating method |
CN110895802A (en) * | 2018-08-23 | 2020-03-20 | 杭州海康威视数字技术股份有限公司 | Image processing method and device |
WO2020077672A1 (en) * | 2018-10-17 | 2020-04-23 | 网宿科技股份有限公司 | Method and device for training service quality evaluation model |
CN112116103A (en) * | 2020-09-17 | 2020-12-22 | 北京大学 | Method, device and system for evaluating personal qualification based on federal learning and storage medium |
CN112650661A (en) * | 2020-12-29 | 2021-04-13 | 北京嘀嘀无限科技发展有限公司 | Data processing quality control method, data processing quality control device, computer equipment and storage medium |
EP3863222A4 (en) * | 2018-10-17 | 2021-12-01 | Wangsu Science & Technology Co., Ltd. | Service quality evaluation model training method and device |
CN115630928A (en) * | 2022-12-01 | 2023-01-20 | 广东省实验动物监测所 | Management method, system and device for administrative permission data of experimental animal |
-
2016
- 2016-09-13 CN CN201610818704.9A patent/CN106384197A/en active Pending
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107563595A (en) * | 2017-08-02 | 2018-01-09 | 卓智网络科技有限公司 | Colleges and universities' core business index scoring system and method based on dynamic discrete algorithm |
CN107909472A (en) * | 2017-12-08 | 2018-04-13 | 上海壹账通金融科技有限公司 | Management data checking method, device, equipment and computer-readable recording medium |
CN107909472B (en) * | 2017-12-08 | 2020-11-03 | 深圳壹账通智能科技有限公司 | Operation data auditing method, device and equipment and computer readable storage medium |
CN108287928A (en) * | 2018-03-05 | 2018-07-17 | 四川易利数字城市科技有限公司 | A kind of space attribute prediction technique based on local weighted linear regression |
CN110580490A (en) * | 2018-06-11 | 2019-12-17 | 杭州海康威视数字技术股份有限公司 | Method, device and equipment for determining personnel behavior probability |
CN109144986A (en) * | 2018-07-30 | 2019-01-04 | 上海电气集团股份有限公司 | A kind of importance appraisal procedure of industrial equipment data |
CN110895802A (en) * | 2018-08-23 | 2020-03-20 | 杭州海康威视数字技术股份有限公司 | Image processing method and device |
CN110895802B (en) * | 2018-08-23 | 2023-09-01 | 杭州海康威视数字技术股份有限公司 | Image processing method and device |
WO2020077672A1 (en) * | 2018-10-17 | 2020-04-23 | 网宿科技股份有限公司 | Method and device for training service quality evaluation model |
EP3863222A4 (en) * | 2018-10-17 | 2021-12-01 | Wangsu Science & Technology Co., Ltd. | Service quality evaluation model training method and device |
CN110781456A (en) * | 2019-09-27 | 2020-02-11 | 上海麦克风文化传媒有限公司 | Sorting weight updating method |
CN112116103A (en) * | 2020-09-17 | 2020-12-22 | 北京大学 | Method, device and system for evaluating personal qualification based on federal learning and storage medium |
CN112116103B (en) * | 2020-09-17 | 2024-07-09 | 北京大学 | Personal qualification evaluation method, device and system based on federal learning and storage medium |
CN112650661A (en) * | 2020-12-29 | 2021-04-13 | 北京嘀嘀无限科技发展有限公司 | Data processing quality control method, data processing quality control device, computer equipment and storage medium |
CN115630928A (en) * | 2022-12-01 | 2023-01-20 | 广东省实验动物监测所 | Management method, system and device for administrative permission data of experimental animal |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106384197A (en) | Service quality evaluation method and device based on big data | |
CN108763277B (en) | Data analysis method, computer readable storage medium and terminal device | |
CN111444247A (en) | KPI (Key performance indicator) -based root cause positioning method and device and storage medium | |
CN109902222A (en) | Recommendation method and device | |
CN103745273B (en) | Semiconductor fabrication process multi-performance prediction method | |
CN110738564A (en) | Post-loan risk assessment method and device and storage medium | |
CN109472700A (en) | Prediction technique, server and the storage medium of stock price | |
CN111506637B (en) | Multi-dimensional anomaly detection method and device based on KPI (Key Performance indicator) and storage medium | |
CN107563645A (en) | A kind of Financial Risk Analysis method based on big data | |
CN104077303B (en) | Method and apparatus for data to be presented | |
KR102342321B1 (en) | Method and apparatus for deep learning-based stock screening and its use for advancing and automating portfolio management | |
US9501522B2 (en) | Systems and methods for providing a unified variable selection approach based on variance preservation | |
CN109636212B (en) | Method for predicting actual running time of job | |
CN104198463A (en) | Raman spectrum preprocessing method and system | |
CN107992978A (en) | It is a kind of to net the method for prewarning risk and relevant apparatus for borrowing platform | |
CN103236013B (en) | A kind of stock market deep bid data analysing method based on the identification of crucial stock collection | |
CN112785377A (en) | Data distribution-based order completion period prediction model construction method and prediction method | |
CN109800138B (en) | CPU testing method, electronic device and storage medium | |
CN109977977B (en) | Method for identifying potential user and corresponding device | |
CN113837383A (en) | Model training method and device, electronic equipment and storage medium | |
CN116993548A (en) | Incremental learning-based education training institution credit assessment method and system for LightGBM-SVM | |
CN111028086A (en) | Enhanced index tracking method based on clustering and LSTM network | |
JP2006318013A (en) | Evaluation device and computer program | |
CN113656707A (en) | Financing product recommendation method, system, storage medium and equipment | |
CN114266653A (en) | Client loan risk estimation method for integrated learning |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170208 |