CN108053266A - A kind of patent value predictor method and device - Google Patents
A kind of patent value predictor method and device Download PDFInfo
- Publication number
- CN108053266A CN108053266A CN201711482711.7A CN201711482711A CN108053266A CN 108053266 A CN108053266 A CN 108053266A CN 201711482711 A CN201711482711 A CN 201711482711A CN 108053266 A CN108053266 A CN 108053266A
- Authority
- CN
- China
- Prior art keywords
- assessed
- future time
- probability
- assessment
- grade
- 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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0206—Price or cost determination based on market factors
-
- 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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0283—Price estimation or determination
Abstract
This application provides a kind of patent value predictor method and device, wherein, this method includes:Patent to be assessed is obtained respectively in the patent assessment grade of multiple historical periods;Wherein, the patent assessment grade is determined according to the estimated value of the patent to be assessed;Based on the patent to be assessed in the patent assessment grade of each historical period, structure hidden Markov model obtains the probability mutually shifted between each patent assessment grade within the period of each historical period composition;Based on the probability mutually shifted between each patent assessment grade, patent assessment grade of the patent to be assessed in future time period is estimated.In this way, the future value that can be directed to patent is estimated.
Description
Technical field
This application involves patent assessment field, in particular to a kind of patent value predictor method and device.
Background technology
With 2006《National Program for Medium-to Long-term Scientific and Technological Development (2006-2020)》Promulgation, country will
Build theme of the development strategy of innovation-oriented country as development.And with deepening constantly for development strategy of innovation-oriented country is built, pass through patent
Independent intellectual property right is protected to become an important process of current each enterprise.Patent is as a kind of important intangible asset, valency
The assessment of value is also increasingly taken seriously in practice.
Law parameter, economic parameters, technical parameter, production mainly currently are based on by professional person to the assessment of patent value
Industry risk, risk of infringement etc. carry out value assessment for a patent, and assessment result is gone through only for a patent a certain
The special value at history moment is assessed.But in fact, over time, the value of patent also becomes constantly
Change, and current patent value assessment method can not realize the prediction to future patent value.
The content of the invention
In view of this, the embodiment of the present application is designed to provide a kind of patent value predictor method and device, can
It is estimated for the future value of patent.
In a first aspect, the embodiment of the present application provides a kind of patent value predictor method, this method includes:
Patent to be assessed is obtained respectively in the patent assessment grade of multiple historical periods;Wherein, the patent assessment grade
It is determined according to the estimated value of the patent to be assessed;
Based on the patent to be assessed each historical period patent assessment grade, build hidden Markov model, obtain
The probability mutually shifted between each patent assessment grade in the period formed in each historical period;
Based on the probability mutually shifted between each patent assessment grade, the patent to be assessed is estimated in the special of future time period
Sharp evaluation grade.
Second aspect, the embodiment of the present application also provide a kind of patent value estimating device, which includes:
Evaluation module, for obtaining patent to be assessed respectively in the patent assessment grade of multiple historical periods;Wherein, it is described
Patent assessment grade is determined according to the estimated value of the patent to be assessed;
Module is built, in the patent assessment grade of each historical period, building hidden horse based on the patent to be assessed
Er Kefu models obtain mutually shifting between each patent assessment grade in the period formed in each historical period general
Rate;
Module is estimated, for based on the probability mutually shifted between each patent assessment grade, estimating the patent to be assessed
In the patent assessment grade of future time period.
The patent value predictor method and device that the embodiment of the present application is provided are obtaining patent to be assessed respectively more
After the patent assessment grade of a historical period, based on patent to be assessed each historical period patent assessment grade, structure it is hidden
Markov model obtains mutually shifting between each patent assessment grade in the period formed in each historical period general
Rate, the probability then mutually shifted according to each patent assessment grade stent, to patent to be assessed future time period patent assessment
Grade is estimated, it is achieved thereby that being estimated to the future value of patent.
Description of the drawings
It, below will be to needed in the embodiment attached in order to illustrate more clearly of the technical solution of the embodiment of the present application
Figure is briefly described, it should be understood that the following drawings illustrates only some embodiments of the application, therefore is not construed as pair
The restriction of scope, for those of ordinary skill in the art, without creative efforts, can also be according to this
A little attached drawings obtain other relevant attached drawings.
Fig. 1 shows a kind of flow chart of patent value predictor method provided by the embodiments of the present application;
Fig. 2 shows the specific method flow chart for the structure hidden Markov model that the embodiment of the present application is provided;
Fig. 3 shows that the patent to be assessed of estimating that the embodiment of the present application is provided has in the patent assessment grade of future time period
Body method flow chart;
Fig. 4 shows the flow chart for another patent value predictor method that the embodiment of the present application is provided;
Fig. 5 shows definite assessment weight of each evaluation index under the historical period that the embodiment of the present application is provided
Specific method flow chart;
Fig. 6 shows that the embodiment of the present application provided obtains next round sign based on the first marking result for currently seeking the opinion of table
Ask the specific method flow chart of the default weighted value of table;
Fig. 7 show that the embodiment of the present application provided according to all the first marking for seeking the opinion of table as a result, calculating the evaluation
The specific method flow chart of the assessment weight of index;
Fig. 8 shows that the value assessment for the quantitative target for calculating patent to be assessed respectively that the embodiment of the present application is provided obtains
Point and qualitative index value assessment score specific method specific method flow chart;
Fig. 9 shows a kind of structure diagram for patent value estimating device that the embodiment of the present application is provided;
Figure 10 shows that the embodiment of the present invention provides a kind of structure diagram of computer equipment.
Specific embodiment
Current patent value assessment method can only assess the current value of patent, and can not be to patent not
It is estimated to be worth.Based on this, a kind of patent value predictor method and device that the application provides can be directed to patent
Future value is estimated.
For ease of understanding the present embodiment, first to a kind of patent valve estimating side disclosed in the embodiment of the present application
Method describes in detail.
Patent value predictor method shown in Figure 1, that the embodiment of the present application is provided, including:
S101:Patent to be assessed is obtained respectively in the patent assessment grade of multiple historical periods;Wherein, patent assessment grade
It is determined according to the estimated value of patent to be assessed.
When specific implementation, patent to be assessed refers in the historical period in the estimated value of some historical period
Under, according to the current state of patent, and value assessment is carried out to patent to be assessed according to certain evaluation criterion and is obtained, and it is same
The standard that patent to be assessed carries out value assessment under multiple historical periods is consistent.In multiple historical periods to be assessed special
Profit carries out value assessment, the estimated value of the patent to be assessed under each historical period can be obtained, then according to certain
The estimated value of patent to be assessed is converted into patent assessment grade by transformation standard.
For example, carrying out value assessment to patent to be assessed in multiple periods, patent to be assessed can be obtained in each history
Estimated value under period is as shown in table 1 below:
Table 1
Time | 2008Q1 | 2008Q2 | 2008Q3 | 2008Q4 |
Estimated value | 0.90 | 0.85 | 0.66 | 0.43 |
Time | 2009Q1 | 2009Q2 | 2009Q3 | 2009Q4 |
Estimated value | 0.77 | 0.89 | 0.30 | 0.65 |
Time | 2010Q1 | 2010Q2 | 2010Q3 | 2010Q4 |
Estimated value | 0.97 | 0.76 | 0.40 | 0.85 |
Time | 2011Q1 | 2011Q2 | 2011Q3 | 2011Q4 |
Estimated value | 0.65 | 0.45 | 0.91 | 0.55 |
Time | 2012Q1 | 2012Q2 | 2012Q3 | 2012Q4 |
Estimated value | 0.80 | 0.40 | 0.44 | 0.88 |
Time | 2013Q1 | 2013Q2 | 2013Q3 | 2013Q4 |
Estimated value | 0.35 | 0.37 | 0.76 | 0.92 |
Time | 2014Q1 | 2014Q2 | 2014Q3 | 2014Q4 |
Estimated value | 0.94 | 0.45 | 0.65 | 0.66 |
Time | 2015Q1 | 2015Q2 | 2015Q3 | 2015Q4 |
Estimated value | 0.92 | 0.80 | 0.88 | 0.33 |
Time | 2016Q1 | 2016Q2 | 2016Q3 | 2016Q4 |
Estimated value | 0.55 | 0.88 | 0.89 | 0.76 |
Time | 2017Q1 | 2017Q2 | 2017Q3 | 2017Q4 |
Estimated value | 0.85 | 0.42 | 0.95 | 0.73 |
In table 1 above, using from 2008 to 2017 years each season as a historical period, it is obtained with it is every
The estimated value of the corresponding patent to be assessed of one historical period is as shown in table 1.The estimated value of patent to be assessed is converted
It is for the standard of evaluation grade:If estimated value is more than 0.8, the corresponding patent assessment grade of the historical period to be advanced,
It is represented with e1;If estimated value is less than or equal to 0.8, and is more than 0.5, then the corresponding patent assessment grade of the historical period
For middle rank, represented with e2;If estimated value is less than or equal to 0.5, the corresponding patent assessment grade of the historical period is
It is rudimentary, it is represented with e3.After the estimated value in upper table 1 then is converted into patent assessment grade according to the transformation standard, specifically such as
Shown in the following table 2:
Table 2
Time | 2008Q1 | 2008Q2 | 2008Q3 | 2008Q4 |
Patent assessment grade | e1 | e1 | e2 | e3 |
Time | 2009Q1 | 2009Q2 | 2009Q3 | 2009Q4 |
Patent assessment grade | e2 | e1 | e3 | e2 |
Time | 2010Q1 | 2010Q2 | 2010Q3 | 2010Q4 |
Patent assessment grade | e1 | e2 | e3 | e1 |
Time | 2011Q1 | 2011Q2 | 2011Q3 | 2011Q4 |
Patent assessment grade | e2 | e3 | e1 | e2 |
Time | 2012Q1 | 2012Q2 | 2012Q3 | 2012Q4 |
Patent assessment grade | e1 | e3 | e3 | e1 |
Time | 2013Q1 | 2013Q2 | 2013Q3 | 2013Q4 |
Patent assessment grade | e3 | e3 | e2 | e1 |
Time | 2014Q1 | 2014Q2 | 2014Q3 | 2014Q4 |
Patent assessment grade | e1 | e3 | e2 | e2 |
Time | 2015Q1 | 2015Q2 | 2015Q3 | 2015Q4 |
Patent assessment grade | e1 | e2 | e1 | e3 |
Time | 2016Q1 | 2016Q2 | 2016Q3 | 2016Q4 |
Patent assessment grade | e2 | e1 | e1 | e2 |
Time | 2017Q1 | 2017Q2 | 2017Q3 | 2017Q4 |
Patent assessment grade | e2 | e3 | e1 | e2 |
S102:Based on patent to be assessed each historical period patent assessment grade, build hidden Markov model, obtain
The probability mutually shifted between each patent assessment grade in the period formed in each historical period.
Hidden Markov model is the probabilistic model on sequential, describes to be generated at random by a hiding Markov chain
Unobservable state random sequence, then an observation is generated by each state and generates the process of observation random sequence.Ma Er
Can husband's chain, be in mathematics have Markov property discrete time stochastic process;During being somebody's turn to do, in given current knowledge or letter
In the case of breath, the past (i.e. current pervious historic state), (i.e. current later future state) was unrelated in the future for prediction
's.The sequence for the state that hiding Markov chain generates at random is known as status switch (state sequence);Each state
One observation of generation, and the random sequence of resulting observation, are known as observation sequence (observation sequence).Sequence
Each position of row can be regarded as a moment again.
If Q is the set of all possible state, V is the set of all possible observation.
Then Q can be represented as:Q={ q1,q2,…,qN, wherein, N is the quantity of possible state.
V can be represented as:V={ v1,v2,…,vM, wherein, M is the quantity of possible observation.
A is state transition probability matrix, and A is expressed as:A=[aij]N×N;
Wherein, aij=P (it+1=qj|it=qi), i=1,2 ..., N;J=1,2 ..., N.
State transition probability matrix A represents to be in state q in moment tiUnder conditions of in moment t+1 be transferred to state qj
Probability.
B is observation probability matrix, and B is expressed as:B=[bj(k)]N×M;
Wherein, bj(k)=P (ot=vk|it=qj), k=1,2 ..., M;J=1,2 ..., N.
The observation probability matrix B represents to be in state q in moment tjUnder conditions of generation observation vkProbability.
π is initial state probability vector, and π is expressed as:π=(πi);
Wherein, πi=P (i1=qi), i=1,2 ..., N;
The initial state probability vector represents that moment t+1 is in probability qiProbability.
Hidden Markov model is determined by initial state probability vector π, state transition probability matrix A and observation probability matrix B
Fixed, i.e. π and A determine status switch, and B determines observation sequence.Hidden Markov model λ can be represented with ternary symbol as a result,
I.e.:
λ=(A, B, π).
That is state transition probability matrix A and initial state probability vector π determines hiding Markov Chain, and generation can not
The status switch of observation.Observation probability matrix B determines how generation observation, determines how that generation is seen with status switch synthesis
Sequencing row.
Hidden Markov model has made two basic assumptions:
(1) before Markov property is it is assumed that assume that the state of hiding Markov chain t at any time only depends on it
The state at one moment, it is unrelated with the state at other moment and observation.
According to the markov property of markoff process and the condition probability formula of Bayes, have
The recurrence formula of state probability is gradually calculated with this:
In above formula (1), π (0)=[π1(0),π2(0),...,πn(0)] it is initial state probability vector.The k-th moment
State probability predicts, if a certain event the 0th moment original state it is known that i.e. π (0) it is known that recursion using top
Formula (1), it is possible to acquire after k state shifts, the probability of various possible states is in k-th of moment.It can
The state probability that the event is obtained at the k-th moment is predicted.
Shown in Figure 2 based on above-mentioned hidden Markov model, the embodiment of the present application also provides a kind of based on to be assessed special
Profit each historical period patent assessment grade, build hidden Markov model, obtain each historical period form when
The method of the probability mutually shifted between each patent assessment grade in section, this method include:
S201:Each patent assessment grade is determined as a kind of state;And by the patent assessment grade of each historical period
It is determined as the state observed in corresponding historical period, obtains the corresponding status switch of each historical period.
Wherein, the row and column of state transition probability matrix characterizes each state;The list of elements in state transition probability matrix
Levy the element be expert at characterization state be transferred to the element column and characterize shape probability of state or the element column institute
Characterization state is transferred to the element and is expert at characterized shape probability of state.
When specific implementation, by taking the example shown in above-mentioned table 2 as an example, due in above-mentioned 2 Patent evaluation grade of table
Including:Therefore each patent assessment grade, is determined as a kind of state, the state set Q of composition is by totally three kinds of e1, e2 and e3:
Q={ e1,e2,e3}。
It is obtained by the state that the patent assessment grade of each historical period is determined as observing in corresponding historical period
The corresponding status switch of each historical period is:
e1→e1→e2→e3→e2→e1→e3→e2→
e1→e2→e3→e1→e2→e3→e1→e2→
e1→e3→e3→e1→e3→e3→e2→e1→
e1→e3→e2→e2→e1→e2→e1→e3→
e2→e1→e1→e2→e2→e3→e1→e2
S202:Based on status switch structural regime transition probability matrix.
Herein, in above-mentioned table 2 corresponds to example, the state transition probability matrix A of constructed hidden Markov model can
To be expressed as:
From above-mentioned status switch:In the state set out in 15 e1,3 are to be transferred to e1 from e1,7 be from
E1 is transferred to e2, and 5 are to be transferred to e3 from e1;So:
It can similarly obtain:In the state set out in 13 e2,7 are to be transferred to e1 from e2, and 2 are transferred to from e2
E2,4 are to be transferred to e3 from e2;So:
In the state set out in 11 e3,4 are to be transferred to e1 from e3, and 5 are to be transferred to e2 from e3, and 2 are
E3 is transferred to from e3;So:
It can thus be concluded that above-mentioned state-transition matrix A is:
The probability then mutually shifted between identified each patent assessment grade is as shown in above-mentioned state-transition matrix.
S103:Based on the probability mutually shifted between each patent assessment grade, patent to be assessed is estimated in future time period
Patent assessment grade.
It is shown in Figure 3 when specific implementation, it provides a kind of based on mutually shifting between each patent assessment grade
Probability, estimates specific method of the patent to be assessed in the patent assessment grade of future time period, and this method includes:
S301:Patent to be assessed is predefined in multiple future time periods in a nearest future time period in each patent assessment etc.
The probability of grade.
S302:For each future time period in addition to a nearest future time period, based on what is abutted with the future time period
In the period that patent to be assessed is formed in the probability of each patent assessment grade and each historical period in a upper future time period
The probability mutually shifted between each patent assessment grade determines that patent to be assessed is in each patent assessment grade in the future time period
Probability.
S303:According to patent to be assessed in the future time period in the probability of each patent assessment grade, patent to be assessed is determined
Patent assessment grade in the future time period.
Specifically, patent to be assessed is commented in each patent in a nearest future time period in multiple future time periods are predefined
When estimating the probability of grade, since patent to be assessed includes multiple patent assessment grades, can be based in future time period
A nearest future time period corresponding state probability vector estimates patent to be assessed.
When specific implementation, by taking 2 corresponding example of upper table as an example, it is assumed that the period residing for current time is
2017Q4 determines that in multiple future time periods a nearest future time period is 2018Q1.
Be e2 to the patent assessment grade forecast that patent to be assessed carries out assuming that in the 2018Q1 periods, then it is to be assessed special
Profit forms a probability vector in the probability of three patent assessment grades e1, e2 and e3, and probability vector π meets:π (0)=
[0,1,0]。
Herein, in 2 corresponding example of upper table, obtained state-transition matrix A is:
Determine that patent to be assessed is in the probability of each patent assessment grade in the future time period, in one embodiment,
Each future time period in addition to a nearest future time period can be directed to, based on the upper future abutted with the future time period
The product of period corresponding state probability vector and state probability matrix determines that patent to be assessed is special at each in the future time period
The probability of sharp evaluation grade.
For example, probability vector π (0)=[0,1,0] and probability transfer matrix A are updated to shown in above-mentioned formula (1)
In recurrence formula, then it can obtain the patent to be assessed and be likely to occur the general of various states in remaining three season in 2018
Rate:
In the 2018Q2 periods:
In the 2018Q3 periods:
In the 2018Q4 periods:
In the 2019Q1 periods:
Then period second quarter 2018Q2 in 2018, period third season 2018Q3, period fourth quarter 2018Q4 with
And period first quarter 2019Q1 in 2019, the result estimated to patent to be assessed are as shown in table 3 below:
Table 3
After obtaining such as the estimation results of four future time periods in upper table 3, existed according to patent to be assessed in the future time period
The probability of each patent assessment grade estimates patent assessment grade of the patent to be assessed in the future time period.
For example, using the highest state of probability of occurrence in each period as patent to be assessed in patent assessment of the period etc.
Grade, i.e., 2018Q2 periods, the patent assessment grade of patent to be assessed are:e1;In the 2018Q3 periods, the patent of patent to be assessed is commented
Grade is estimated for e2;In the 2018Q4 periods, the patent assessment grade of patent to be assessed is e1;In the 2019Q1 periods, patent to be assessed
Patent assessment grade be e2.
The patent value predictor method that the embodiment of the present application is provided, obtain patent to be assessed respectively in multiple history when
After the patent assessment grade of section, based on patent to be assessed in the patent assessment grade of each historical period, structure hidden Markov
Model obtains the probability mutually shifted between each patent assessment grade in the period formed in each historical period, then
According to the probability that each patent assessment grade stent mutually shifts, patent assessment grade of the patent to be assessed in future time period is carried out
It estimates, it is achieved thereby that being estimated to the future value of patent.
It is shown in Figure 4, in an alternative embodiment of the invention, patent to be assessed is being obtained respectively in the special of multiple historical periods
Before sharp evaluation grade, the patent value predictor method that the embodiment of the present application is provided further includes:
S401:Determine each evaluation index included in the assessment indicator system built in advance in multiple and different history
The assessment weight of section.
When specific implementation, structure indicator evaluation system is first had to.It is based on evaluation that assessment is carried out to patent value
Included a variety of evaluation indexes are evaluated in index system.Evaluation index generally comprises two kinds:Qualitative index and quantitative
Index, the property of each evaluation index is any one in qualitative index and quantitative target.
Wherein, quantitative target is generally all more specific and with good intuitive, passes through actual numerical value
It calculates, and formulates evaluation criterion and also compare clearly, pass through simple quantificational description, it is possible to direct, clearly expression assessment
As a result.Current, the assessment of patent value generally requires to consider attribute itself, legal status and the market economy of patent to this
Influence situation of patent etc..But patent valve estimating is the system of a complex multi-dimensional, not all can reflect specially
The factor and index of profit value can quantify what is weighed, therefore design qualitative index is particularly important.These qualitative indexes
In covering on width and range for information, quantitative target will be far longer than, but also the evaluation result of patent valve estimating is more
Add with guide property and comprehensive.
In an alternative embodiment, constructed assessment indicator system is as shown in table 4:
Table 4
After assessment indicator system is constructed, to be referred to according to each evaluation included in the assessment indicator system of structure
Mark determines assessment weight of the patent to be assessed in multiple and different historical periods.
Specifically, the embodiment of the present application also provides one kind under any one historical period, determines in assessment indicator system
Assessment weight of each evaluation index under the historical period specific method, shown in Figure 5, this method includes:
S501:For each evaluation index in assessment indicator system, generate more wheels for the evaluation index and seek the opinion of table;Its
In, seeking the opinion of table includes:Multiple default weighted values that evaluation index corresponding with seeking the opinion of table uses in being seeked the opinion of when front-wheel.
Specifically, when generation wheel seeks the opinion of table, it is necessary to which to seeking the opinion of the problem of is described, for example, to patent to be assessed
In value assessment, the problem of can will seeking the opinion of, is described as:Evaluation index is occupied in constructed assessment indicator system
Assessment weight;After problem description is carried out, to provide corresponding background material, as expert marking when part marking according to
According to.In this application, multiple default weighted values are provided with for each evaluation index, the first expert can be asked based on what is seeked the opinion of
Topic, gives a mark to multiple default weighted values of each evaluation index, is then based on all first experts and same evaluation is referred to
Target is taken turns more seeks the opinion of table marking as a result, obtaining the assessment weight of this kind of evaluation index.
Herein, the multiple default weighted values used in table are seeked the opinion of in the first round, can be carried out based on the background material of collection
Preliminary estimation draws or is obtained according to certain algorithm.
S502:Table is seeked the opinion of for every wheel, obtains what the first expert of multidigit used the evaluation index in the wheel seeks the opinion of table
First marking result of multiple default weights;Wherein, each round seeks the opinion of the first marking result there are one table correspondences;And next round is levied
The default weighted value of inquiry table is determined based on the first marking result for currently seeking the opinion of table.
Herein, the first expert is that have certain representativeness and authoritative in the corresponding technical field of patent to be assessed
Multidigit expert by seeking the opinion of multidigit first with expertise, and counts the first expertise of multidigit, is handled, being divided
Analysis and conclusion, objectively integrate the first expertise and subjective judgement, to being largely difficult to carry out quantitative analysis using technical method
Factor make reasonable estimation, after excessively taking turns opinion and seeking the opinion of, feed back and adjust, obtain the assessment weight of each evaluation index.
After being determined that each evaluation index is corresponding and seeking the opinion of table, the first expert can seek the opinion of table to the evaluation index in the wheel
The middle multiple default weighted values used are given a mark, and obtain corresponding first marking result.
It is noted herein that since default weighted value that next round seeks the opinion of table is based on currently seeking the opinion of first dozen of table
What point result determined, therefore, it is not disposably to be generated before carry out problem is seeked the opinion of that more wheels, which seek the opinion of table, but first generates first
Wheel seeks the opinion of table;Obtaining the first expert the first marking result that table gives a mark was seeked the opinion of the first round and then based on the first round
The first marking of table is seeked the opinion of as a result, obtaining the second wheel seeks the opinion of the corresponding default weighted value of table, the second wheel of generation seeks the opinion of table;By same
The method of sample can obtain third round and seek the opinion of the corresponding default weighted value of table, and generation third round seeks the opinion of table;Until it obtains final
Seek the opinion of result.
In addition, the corresponding wheel number for seeking the opinion of table of different evaluation indexes, can be the same or different, as long as last is taken turns
Seek the opinion of until table can obtain satisfactory result.
Specifically, it is shown in Figure 6, seek the opinion of the pre- of table obtaining next round based on the first marking result for currently seeking the opinion of table
If it during weighted value, specifically includes:
S601:For each default weighted value of each evaluation index in table is currently seeked the opinion of, all first experts are calculated
The average value that weighted value gives a mark is preset to this;
S602:According to the corresponding average value of all default weighted values of the evaluation index, the evaluation index pair is calculated
Answer the desired value of weight;
S603:Based on the desired value of the evaluation index respective weights, determine that next round seeks the opinion of the more of the evaluation index in table
A default weighted value.
Here, it is assumed that it is 0.293 calculating the corresponding weight desired value of certain evaluation index, in the case where being worth to based on the expectation
It, can be by default weighted value in 0.293 left and right settings when one wheel seeks the opinion of multiple default weighted values corresponding with evaluation index in table;
For example, if this kind of evaluation index corresponding default weighted value in next round seeks the opinion of table can be respectively set to:0.28、
0.29、0.3、0.31、0.32。
S503:According to all the first marking for seeking the opinion of table as a result, calculating the assessment weight of the evaluation index.
Herein, seek the opinion of carrying out more wheels to certain evaluation index, obtain after more wheels seek the opinion of the corresponding first marking result of table,
It can be according to corresponding first marking of this kind of evaluation index as a result, calculating the corresponding marking weight of this kind of evaluation index.
Specifically, shown in Figure 7, the embodiment of the present application also provides a kind of according to all the first marking knots for seeking the opinion of table
Fruit, calculates the specific method of the assessment weight of the evaluation index, and this method includes:
S701:Determine the expectation that the first marking result based on the evaluation index in last wheel seeks the opinion of table is calculated
Value, and using the desired value as the final weight of the evaluation index;
S702:The corresponding final weight of all evaluation indexes is normalized, obtains each evaluation index correspondence
Assessment weight.
Herein, the corresponding final weight of each evaluation index is the weight that determines of marking result based on the first expert, institute
Have the final weight of evaluation index and it is likely larger than 1, it is also possible to less than 1, it is therefore desirable to corresponding to all evaluation indexes
Final weight is normalized, to obtain the corresponding assessment weight of each evaluation index;The appraisal right of all evaluation indexes
Weight and equal to 1.
Each evaluation index is different to the influence degree (weight) of patent to be assessed, and score value varies, and passes through the first
Expert estimation, quantization weight and score value simultaneously calculate score, so as to judge the height of patented technology dimension scores.
When determining the weight to certain evaluation index, it is assumed that the first expert has 15 people:
The first round seeks the opinion of:
First marking result of multiple default weights that the first expert uses in being seeked the opinion of in the first round the evaluation index is whole
Reason such as the following table 5:
Table 5
Pass through formulaThe corresponding probability for the various default weighted values filled in 15 experts is averaged, and obtains
The average value given a mark to the first expert to 5 kinds of default weighted values of the evaluation index:
……
The probability corresponding to this kind of evaluation index 5 kinds of default weighted values of correspondence can be finally obtained, it is as shown in table 6 below:
Table 6
Default weight | 0.1 | 0.2 | 0.3 | 0.4 | 0.5 |
Probability (%) | 7.3 | 16.2 | 58 | 13.2 | 5.3 |
According to upper table 3, the desired value E (x) that the evaluation index seeks the opinion of corresponding weight in table in the wheel is calculated:
E (x)=7.3% × 0.1+16.2% × 0.2+58.0% × 0.3+13.2% × 0.4+5.3% × 0.5=
0.293;
The variance D (x) of the corresponding weight of the evaluation index is calculated, variance is used to describe evaluation index in the wheel seeks the opinion of table
Corresponding weight deviates the index of desired value size:
Standard deviation S is calculated according to method, standard deviation S is the square root of variance D (x):
According to standard deviation S and desired value E (x), coefficient of dispersion β is calculated,
Wherein, in the embodiment, variance is used to weigh the metric of metadata and desired value difference;Standard deviation is used to reflect
The dispersion degree of data set;Coefficient of dispersion is used to reflect the dispersion degree in unit average, and in this application, coefficient of dispersion is got over
Small, the accuracy that result is obtained by expert estimation is also higher.
In order to make evaluation result more accurate, the default weighted value in the second wheel seeks the opinion of table, in desired value E (x) left and right
It resets, is respectively:0.28th, 0.29,0.3,0.31,0.32, and generate the second wheel and seek the opinion of table.
Second wheel is seeked the opinion of:
First marking result of multiple default weights that the first expert uses the evaluation index in the second wheel is seeked the opinion of is whole
Reason such as the following table 7:
Table 7
Pass through formulaThe corresponding probability for the various default weighted values filled in 15 experts is averaged, and obtains
The average value given a mark to the first expert to 5 kinds of default weighted values of the evaluation index:
……
The probability corresponding to this kind of evaluation index 5 kinds of default weighted values of correspondence can be finally obtained, it is as shown in table 8 below:
Table 8
Default weighted value | 0.28 | 0.29 | 0.3 | 0.31 | 0.32 |
Probability (%) | 12.3 | 11.2 | 51.7 | 18.5 | 5.3 |
According to upper table 3, the desired value E (x) that the evaluation index seeks the opinion of corresponding weight in table in the wheel is calculated:
E (x)=12.3% × 0.28+11.2% × 0.29+51.7% × 0.30+18.5% × 0.31+5.3% × 0.32
=0.296;
The variance D (x) of the corresponding weight of the evaluation index is calculated, method is used to describe evaluation index in the wheel seeks the opinion of table
Corresponding weight deviates the index of desired value size:
Standard deviation S is calculated according to method, standard deviation S is the square root of variance D (x):
According to standard deviation S and desired value E (x), coefficient of dispersion β is calculated,
In order to make evaluation result more accurate, third round is seeked the opinion of to the default weighted value in table, in desired value E (x) left and right
It resets, and generates the second wheel and seek the opinion of table.
Similarly, table structure third round can be seeked the opinion of by the second wheel and seeks the opinion of table, finally seeked the opinion of in table and calculated according to each wheel
The desired value arrived obtains the final weight of the evaluation index.
Specifically, it is every wheel seek the opinion of when, can be determined according to the size that calculated coefficient of dispersion is seeked the opinion of when front-wheel be
The no next round that carry out is seeked the opinion of for example, after each round is seeked the opinion of, which is seeked the opinion of to the coefficient of dispersion being calculated with presetting
Threshold value be compared, if coefficient of dispersion be less than default threshold value, the desired value E (x) for the weight which is seeked the opinion of
Final weight as the evaluation index.It is compared furthermore it is also possible to which more wheels are seeked the opinion of obtained coefficient of dispersion, by minimum
Final weights of the corresponding desired value E (x) for seeking the opinion of the weight that table is calculated of coefficient of dispersion as the evaluation index.
After the final weight of all evaluation indexes is calculated based on method same as described above, to all evaluation indexes
Final weight is normalized, and calculates the assessment weight of each evaluation index:
Assuming that there are four types of evaluation index, and the corresponding final weight of four kinds of evaluation indexes is followed successively by:0.293,0.328,
0.456,0.536, according to formula:Calculate the assessment weight w of each evaluation indexi:
That is the corresponding assessment weight of four kinds of evaluation indexes is followed successively by:0.182,0.203,0.283,0.332.
In assessment indicator system is determined after the assessment weight of each evaluation index, also to draw patent to be assessed at each
The value assessment score of evaluation index, referring specifically to following S402:
S402:Based on each evaluation index multiple and different historical periods assessment weight, calculate patent to be assessed
The value assessment score of each historical period.
Specifically, since the actual conditions of each patent to be assessed are all different, although being determined by above-mentioned steps
When assessing this patent in used assessment indicator system each evaluation index assessment weight, to be also directed to specific
Patent, obtain the value assessment score of each evaluation index of the patent to be assessed.Herein, due to the property bag of evaluation index
Two kinds of qualitative index and quantitative target are included, and qualitative index and quantitative target are obtaining its value assessment due to property qualitative difference
Used method is different during index, therefore, to calculate the quantitative target of patent to be assessed respectively based on different methods
Value assessment score and qualitative index value assessment score.
Shown in Figure 8, the embodiment of the present application also provides a kind of value for the quantitative target for calculating patent to be assessed respectively
The specific method of the value assessment score of score and qualitative index is assessed, this method includes:
S801:For qualitative index, the evaluation score of each qualitative index is calculated;And the evaluation based on each qualitative index
Score determines the qualitative index value assessment score of patent to be assessed.
S802:For quantitative target, the quantitative target value that patent to be assessed is calculated using fuzzy overall evaluation algorithm is commented
Estimate score.
When S801 is implemented, the qualitative index valency of the qualitative index of patent to be assessed is calculated using following step
Value assessment score:The evaluation score of each qualitative index is calculated using following step S901-S903 first, secondly uses following steps
Rapid S904 according to the evaluation score of all qualitative indexes, calculates the quantitative target value assessment score of patent to be assessed.Specifically
Ground:
S901:Respectively each qualitative index determines opinion rating and scoring scope corresponding with each opinion rating;
Each qualitative index corresponds to multiple opinion ratings.
When specific implementation, in order to be scored qualitative index, it is necessary to determine evaluation etc. for each qualitative index
Grade, and to determine the corresponding scoring scope of each opinion rating.Wherein, the opinion rating of each qualitative index all according to its with
The difference of other qualitative indexes, and the difference with the opinion rating of other qualitative indexes is caused, and different qualitative indexes corresponds to
Opinion rating quantity may be different, it is also possible to it is identical.Can be that each qualitative index determines to distinguish the score of qualitative index
Multiple opinion ratings.
S902:Obtain second marking result of multiple second experts to each qualitative index;Wherein, the second expert is according to every
The corresponding opinion rating of kind qualitative index and scoring scope corresponding with each opinion rating give a mark to qualitative index.
Specific implementation when, in order to accurately score a variety of qualitative criterias, obtain and each
The corresponding qualitative index value assessment score of qualitative criteria will obtain second marking of the second expert of multidigit to each qualitative index
As a result.Second marking is the result is that the second expert is based on standards of grading corresponding with each opinion rating and based on objective fact
And the judgement of oneself experience, according to the corresponding opinion rating of each qualitative index and scoring corresponding with each opinion rating
Scope obtains qualitative index marking.
S903:According to commenting for the second marking result, each corresponding Weight of Expert of second expert and each qualitative index
Estimate weight, calculate the evaluation score of corresponding qualitative index.
After the second expert of multidigit is obtained to the second marking result of each qualitative index, the second marking result is based on
And the assessment weight of each qualitative index, and based on the corresponding Weight of Expert of each second expert, calculate corresponding qualitative index
Evaluation score.
Specifically, according to the second marking result, each corresponding Weight of Expert of second expert and each qualitative index
Assessment weight, when calculating the evaluation score of corresponding qualitative index:
The special of each qualitative index is calculated according to the second marking result, the corresponding Weight of Expert of each second expert first
Family's evaluation score, wherein, following formula may be employed in expert opinion score:
M=a1×b1+a2×b2+…+ai×bi
Wherein, M is the corresponding expert opinion score of each qualitative index;a1To aiIt is qualitative to each for i the second experts
The second marking result that index is given a mark;b1To biThe Weight of Expert being corresponding in turn to for i the second experts.
Secondly, according to the expert opinion score of each qualitative index and the corresponding assessment weight of each qualitative index,
Determine the evaluation score of corresponding qualitative index.The specific following manner that may be employed calculates that each qualitative index is corresponding to be evaluated
Point:Ni=Mi×wi;
Wherein:NiFor the corresponding evaluation score of i-th kind of qualitative index;MiIt is obtained for the corresponding expert opinion of i-th kind of qualitative index
Point;wiFor the corresponding assessment weight of i-th kind of qualitative index.
After obtaining the corresponding evaluation score of each qualitative index, to be determined based on the evaluation score of each qualitative index to be evaluated
Estimate the qualitative index value assessment score of patent.It specifically includes:
S904:The qualitative index value assessment that the sum of evaluation score by each qualitative index is determined as patent to be assessed obtains
Point.
When S802 is implemented, the quantitative target valency of the quantitative target of patent to be assessed is calculated using following step
Value assessment score:
S1001:It establishes quantitative target collection and index corresponding with each quantitative target that quantitative target is concentrated is commented
Valency collection;Metrics evaluation concentration includes:The grade and different basis weights index set for quantitative target is corresponding respectively with different brackets
Evaluation criterion.
When specific implementation, first have to establish quantitative target collection and metrics evaluation collection;Wherein, quantitative target concentration includes
Multiple quantitative targets.Quantitative target collection in this application refers to the evaluation index established during patent valve estimating
The set of all quantitative targets in system, wherein each quantitative target are all corresponding, and there are one assessment weights.Metrics evaluation collection refers to
The grade and the corresponding evaluation criterion of different brackets set for each quantitative target that quantitative target is concentrated.
S1002:According to quantitative target collection and metrics evaluation collection, obtain the 3rd expert of multidigit and each quantitative target is carried out
The class information that grade is evaluated.
When specific implementation, be based on above-mentioned quantitative target collection and with metrics evaluation collection, obtain the 3rd expert of multidigit
Carry out the class information evaluated of grade, actually the 3rd expert of multidigit to each quantitative target, objective fact and itself
Experience, and based on metrics evaluation collection, grade evaluation is carried out to each quantitative target, obtains the corresponding grade letter of each quantitative target
Breath.
It should be noted that every the 3rd expert all can to all corresponding multiple grades of quantitative target, according to etc.
The corresponding evaluation criterion of grade carries out level ratings, obtains all 3rd experts and is corresponded to respectively for each quantitative target when evaluating
Class information.
S1003:It is patent to be assessed according to grade scoring table and each definite quantitative target corresponding grade information
Quantitative target generation fuzzy evaluating matrix;Each grade of grade scoring table characterization quantitative target is respectively in multiple predetermined levels
Probability under scoring.
S1004:The weight vectors that the weight of each quantitative target is formed with fuzzy evaluating matrix are multiplied, obtain quantitative target
Score vector;And index score vector is multiplied with subordinated-degree matrix, the value score of quantitative target is obtained, wherein, degree of membership
Matrix element is the score value of multiple predetermined levels scoring.
Due in the assessment indicator system of structure, not only containing quantitative target, but also qualitative index is contained, and above-mentioned
Identified assessment weight is normalized for the final weight of all quantitative target and qualitative index in step S402
Obtained from processing, so if directly use assessment weight corresponding with each quantitative target, for calculating quantitative target
Quantitative target value assessment score is unfavorable, it is necessary to which the corresponding assessment weight of all quantitative targets is normalized,
So as to obtain the weight vectors of the weight of each quantitative target composition.
After the weight vectors formed in the weight for obtaining each quantitative target, weight vectors are multiplied with fuzzy evaluating matrix,
Quantitative target score vector is obtained, and index score vector is multiplied with subordinated-degree matrix, obtains the value score of quantitative target.
S1005:The product for being worth the sum of score and each quantitative target weight is determined as to the quantitative target of patent to be assessed
Value assessment score.
After value assessment score of the patent to be assessed in each historical period is calculated, this method further includes:
S403:According to pre-set patent assessment grade classification foundation and patent to be assessed in each historical period
Value assessment score determines patent to be assessed respectively in the patent assessment grade of multiple historical periods.
If for example, speed it is definite patent to be assessed it is as listed in Table 1 in the value assessment score of each historical period, and
The estimated value of patent to be assessed is also converted into the standard of evaluation grade by pre-set patent assessment grade classification foundation
For:If estimated value is more than 0.8, the corresponding patent assessment grade of the historical period is advanced, is represented with e1;If assessment
Value is less than or equal to 0.8, and more than 0.5, then the corresponding patent assessment grade of the historical period is middle rank, is represented with e2;
If estimated value is less than or equal to 0.5, the corresponding patent assessment grade of the historical period is rudimentary, is represented with e3.Then
Identified patent to be assessed respectively multiple historical periods patent assessment grade as shown in Table 2 above.
Based on same inventive concept, patent valency corresponding with patent value assessment method is additionally provided in the embodiment of the present application
It is worth assessment system, since the principle that the system in the embodiment of the present application solves the problems, such as and the above-mentioned patent value of the embodiment of the present application are commented
Estimate that method is similar, therefore the implementation of system may refer to the implementation of method, overlaps will not be repeated.
Shown in Figure 9, the patent value estimating device that the embodiment of the present application is provided specifically includes:
Acquisition module, for obtaining patent to be assessed respectively in the patent assessment grade of multiple historical periods;Wherein, patent
Evaluation grade is determined according to the estimated value of patent to be assessed;
Module is built, in the patent assessment grade of each historical period, building hidden Ma Erke based on patent to be assessed
Husband's model obtains the probability mutually shifted between each patent assessment grade in the period formed in each historical period;
Module is estimated, for based on the probability mutually shifted between each patent assessment grade, estimating patent to be assessed not
Carry out the patent assessment grade of period.
Optionally, module is estimated, is specifically used for:
Patent to be assessed is predefined in multiple future time periods in a nearest future time period in each patent assessment grade
Probability;
For each future time period in addition to a nearest future time period, based on upper one abutted with the future time period
Each patent in the period that patent to be assessed is formed in the probability of each patent assessment grade and each historical period in future time period
The probability mutually shifted between evaluation grade determines that patent to be assessed is in the probability of each patent assessment grade in the future time period;
According to patent to be assessed in the future time period in the probability of each patent assessment grade, determine patent to be assessed at this not
Carry out the patent assessment grade in the period.
Optionally, module is built, is specifically used for:
Each patent assessment grade is determined as a kind of state;And
By the state that the patent assessment grade of each historical period is determined as observing in corresponding historical period, respectively gone through
History period corresponding status switch;
Based on status switch structural regime transition probability matrix;
Wherein, the row and column of state transition probability matrix characterizes each state;The list of elements in state transition probability matrix
Levy the element be expert at characterization state be transferred to the element column and characterize shape probability of state or the element column institute
Characterization state is transferred to the element and is expert at characterized shape probability of state.
Optionally, module is estimated, is specifically used for:
Predefine the corresponding state probability vector of a nearest future time period in multiple future time periods;
For each future time period in addition to a nearest future time period, based on upper one abutted with the future time period
The product of the corresponding state probability vector of future time period and state probability matrix, determines that patent to be assessed is every in the future time period
The probability of kind patent assessment grade;And
According to patent to be assessed in the future time period in the probability of each patent assessment grade, patent to be assessed is estimated at this
The patent assessment grade of future time period;
Wherein, the element number that state probability vector includes is patent assessment number of levels;The state probability vector list of elements
Patent to be assessed is in the probability of corresponding patent assessment grade in the corresponding future time period of sign;The step-length of future time period and historical period
Step-length is identical.
Optionally, evaluation module, before acquisition patent to be assessed respectively the patent assessment grade in multiple historical periods,
It is additionally operable to:
Determine each evaluation index included in the assessment indicator system built in advance in multiple and different historical periods
Assessment weight;
Based on each evaluation index in the assessment weight of multiple and different historical periods, being gone through each for patent to be assessed is calculated
The value assessment score of history period;
According to pre-set patent assessment grade classification according to and patent to be assessed each historical period value
Score is assessed, determines patent to be assessed respectively in the patent assessment grade of multiple historical periods.
The patent value estimating device that the embodiment of the present invention is provided, obtain patent to be assessed respectively in multiple history when
After the patent assessment grade of section, based on patent to be assessed in the patent assessment grade of each historical period, structure hidden Markov
Model obtains the probability mutually shifted between each patent assessment grade in the period formed in each historical period, then
According to the probability that each patent assessment grade stent mutually shifts, patent assessment grade of the patent to be assessed in future time period is carried out
It estimates, it is achieved thereby that being estimated to the future value of patent.
Corresponding to the patent value predictor method in Fig. 1, the embodiment of the present application additionally provides a kind of computer equipment, such as schemes
Shown in 10, which includes memory 1000, processor 2000 and is stored on the memory 1000 and can be in the processor
The computer program run on 2000, wherein, above-mentioned processor 2000 realizes above-mentioned patent valency when performing above computer program
The step of being worth predictor method.
Specifically, above-mentioned memory 1000 and processor 2000 can be general memory and processor, not do here
It is specific to limit, when the computer program of 2000 run memory 1000 of processor storage, it is pre- to be able to carry out above-mentioned patent value
Estimate method, so as to solve the problems, such as that the future value that can not currently treat assessment patent is estimated, and then can be directed to
The effect that the future value of patent is estimated.
Corresponding to the patent value predictor method in Fig. 1, the embodiment of the present application additionally provides a kind of computer-readable storage
Medium is stored with computer program on the computer readable storage medium, is performed when which is run by processor
The step of stating patent value assessment method.
Specifically, which can be general storage medium, such as mobile disk, hard disk, on the storage medium
Computer program when being run, be able to carry out above-mentioned patent value predictor method, it is special so as to solve currently treat assessment
The problem of future value of profit is estimated, and then the effect that the future value of patent is estimated can be directed to.
The patent value predictor method and the computer program product of system that the embodiment of the present application is provided, including storage
The computer readable storage medium of program code, the instruction that program code includes can be used for performing in previous methods embodiments
Method, specific implementation can be found in embodiment of the method, and details are not described herein.
It is apparent to those skilled in the art that for convenience and simplicity of description, the system of foregoing description
With the specific work process of system, the corresponding process in preceding method embodiment is may be referred to, details are not described herein.
If function is realized in the form of SFU software functional unit and is independent production marketing or in use, can store
In a computer read/write memory medium.Based on such understanding, the technical solution of the application is substantially in other words to existing
The part for having part that technology contributes or the technical solution can be embodied in the form of software product, the computer
Software product is stored in a storage medium, is used including some instructions so that a computer equipment (can be personal meter
Calculation machine, server or network equipment etc.) perform each embodiment method of the application all or part of step.It is and foregoing
Storage medium includes:USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory
The various media that can store program code such as (RAM, Random Access Memory), magnetic disc or CD.
More than, the only specific embodiment of the application, but the protection domain of the application is not limited thereto is any to be familiar with
In the technical scope that those skilled in the art disclose in the application, change or replacement can be readily occurred in, should all be covered
Within the protection domain of the application.Therefore, the protection domain of the application should be subject to the protection scope in claims.
Claims (10)
1. a kind of patent value predictor method, which is characterized in that this method includes:
Patent to be assessed is obtained respectively in the patent assessment grade of multiple historical periods;Wherein, the patent assessment grade according to
The estimated value of the patent to be assessed determines;
Based on the patent to be assessed each historical period patent assessment grade, build hidden Markov model, obtain
The probability mutually shifted between each patent assessment grade in the period that each historical period is formed;
Based on the probability mutually shifted between each patent assessment grade, estimate the patent to be assessed and commented in the patent of future time period
Estimate grade.
It is 2. according to the method described in claim 1, it is characterized in that, general based on mutually being shifted between each patent assessment grade
Rate is estimated patent assessment grade of the patent to be assessed in future time period, is specifically included:
The patent to be assessed is predefined in multiple future time periods in a nearest future time period in each patent assessment grade
Probability;
For each future time period in addition to a nearest future time period, based on the upper future abutted with the future time period
Each patent in the period that the patent to be assessed is formed in the probability of each patent assessment grade and each historical period in period
The probability mutually shifted between evaluation grade determines that the patent to be assessed is in the general of each patent assessment grade in the future time period
Rate;
According to the patent to be assessed in the future time period in the probability of each patent assessment grade, determine that the patent to be assessed exists
Patent assessment grade in the future time period.
3. according to the method described in claim 1, it is characterized in that, based on the patent to be assessed in the special of each historical period
Sharp evaluation grade builds hidden Markov model, obtains each patent assessment etc. in the period formed in each historical period
The probability mutually shifted between grade, specifically includes:
Each patent assessment grade is determined as a kind of state;And
By the state that the patent assessment grade of each historical period is determined as observing in corresponding historical period, when obtaining each history
The corresponding status switch of section;
Based on the status switch structural regime transition probability matrix;
Wherein, the row and column of the state transition probability matrix characterizes each state;Member in the state transition probability matrix
Element characterize the element be expert at characterization state be transferred to the element column characterize shape probability of state or the element place
Row institute characterization state is transferred to the element and is expert at characterized shape probability of state.
It is 4. according to the method described in claim 3, it is characterized in that, general based on mutually being shifted between each patent assessment grade
Rate is estimated patent assessment grade of the patent to be assessed in future time period, is specifically included:
Predefine the corresponding state probability vector of a nearest future time period in multiple future time periods;
For each future time period in addition to a nearest future time period, based on the upper future abutted with the future time period
The product of period corresponding state probability vector and the state probability matrix, determines the patent to be assessed in the future time period
In the probability of each patent assessment grade;And
According to the patent to be assessed in the future time period in the probability of each patent assessment grade, the patent to be assessed is estimated
In the patent assessment grade of the future time period;
Wherein, the element number that the state probability vector includes is patent assessment number of levels;The state probability vector member
The patent to be assessed is in the probability of corresponding patent assessment grade in the corresponding future time period of element characterization;The step-length of the future time period
It is identical with the step-length of historical period.
5. according to claim 1-4 any one of them methods, which is characterized in that gone through respectively multiple obtaining patent to be assessed
Before the patent assessment grade of history period, further include:
Determine assessment of each evaluation index included in the assessment indicator system built in advance in multiple and different historical periods
Weight;
Based on each evaluation index multiple and different historical periods assessment weight, calculate patent to be assessed in each history
The value assessment score of section;
According to pre-set patent assessment grade classification according to and the patent to be assessed each historical period value
Score is assessed, determines the patent to be assessed respectively in the patent assessment grade of multiple historical periods.
6. a kind of patent value estimating device, which is characterized in that the device includes:
Evaluation module, for obtaining patent to be assessed respectively in the patent assessment grade of multiple historical periods;Wherein, the patent
Evaluation grade is determined according to the estimated value of the patent to be assessed;
Module is built, in the patent assessment grade of each historical period, building hidden Ma Erke based on the patent to be assessed
Husband's model obtains the probability mutually shifted between each patent assessment grade in the period formed in each historical period;
Module is estimated, for based on the probability mutually shifted between each patent assessment grade, estimating the patent to be assessed not
Carry out the patent assessment grade of period.
7. device according to claim 6, which is characterized in that it is described to estimate module, it is specifically used for:
The patent to be assessed is predefined in multiple future time periods in a nearest future time period in each patent assessment grade
Probability;
For each future time period in addition to a nearest future time period, based on the upper future abutted with the future time period
Each patent in the period that the patent to be assessed is formed in the probability of each patent assessment grade and each historical period in period
The probability mutually shifted between evaluation grade determines that the patent to be assessed is in the general of each patent assessment grade in the future time period
Rate;
According to the patent to be assessed in the future time period in the probability of each patent assessment grade, determine that the patent to be assessed exists
Patent assessment grade in the future time period.
8. device according to claim 6, which is characterized in that the structure module is specifically used for:
Each patent assessment grade is determined as a kind of state;And
By the state that the patent assessment grade of each historical period is determined as observing in corresponding historical period, when obtaining each history
The corresponding status switch of section;
Based on the status switch structural regime transition probability matrix;
Wherein, the row and column of the state transition probability matrix characterizes each state;Member in the state transition probability matrix
Element characterize the element be expert at characterization state be transferred to the element column characterize shape probability of state or the element place
Row institute characterization state is transferred to the element and is expert at characterized shape probability of state.
9. device according to claim 8, which is characterized in that it is described to estimate module, it is specifically used for:
Predefine the corresponding state probability vector of a nearest future time period in multiple future time periods;
For each future time period in addition to a nearest future time period, based on the upper future abutted with the future time period
The product of period corresponding state probability vector and the state probability matrix, determines the patent to be assessed in the future time period
In the probability of each patent assessment grade;And
According to the patent to be assessed in the future time period in the probability of each patent assessment grade, the patent to be assessed is estimated
In the patent assessment grade of the future time period;
Wherein, the element number that the state probability vector includes is patent assessment number of levels;The state probability vector member
The patent to be assessed is in the probability of corresponding patent assessment grade in the corresponding future time period of element characterization;The step-length of the future time period
It is identical with the step-length of historical period.
10. according to claim 6-9 any one of them devices, which is characterized in that the evaluation module, it is to be assessed special obtaining
Profit before the patent assessment grade of multiple historical periods, is additionally operable to respectively:
Determine assessment of each evaluation index included in the assessment indicator system built in advance in multiple and different historical periods
Weight;
Based on each evaluation index multiple and different historical periods assessment weight, calculate patent to be assessed in each history
The value assessment score of section;
According to pre-set patent assessment grade classification according to and the patent to be assessed each historical period value
Score is assessed, determines the patent to be assessed respectively in the patent assessment grade of multiple historical periods.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711482711.7A CN108053266A (en) | 2017-12-29 | 2017-12-29 | A kind of patent value predictor method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711482711.7A CN108053266A (en) | 2017-12-29 | 2017-12-29 | A kind of patent value predictor method and device |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108053266A true CN108053266A (en) | 2018-05-18 |
Family
ID=62128697
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711482711.7A Pending CN108053266A (en) | 2017-12-29 | 2017-12-29 | A kind of patent value predictor method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108053266A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109146727A (en) * | 2018-08-01 | 2019-01-04 | 深圳派富知识产权投资咨询有限公司 | Patent value assessment method and system based on AI |
CN109948885A (en) * | 2019-01-21 | 2019-06-28 | 三峡大学 | A kind of legal construction assessment system based on quantum Markov chain |
-
2017
- 2017-12-29 CN CN201711482711.7A patent/CN108053266A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109146727A (en) * | 2018-08-01 | 2019-01-04 | 深圳派富知识产权投资咨询有限公司 | Patent value assessment method and system based on AI |
CN109948885A (en) * | 2019-01-21 | 2019-06-28 | 三峡大学 | A kind of legal construction assessment system based on quantum Markov chain |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Richardson et al. | Nowcasting GDP using machine-learning algorithms: A real-time assessment | |
Leathwick et al. | Comparative performance of generalized additive models and multivariate adaptive regression splines for statistical modelling of species distributions | |
Cai et al. | A new fuzzy time series forecasting model combined with ant colony optimization and auto-regression | |
Jahanshahloo et al. | Extension of the TOPSIS method for decision-making problems with fuzzy data | |
Hansun | A new approach of brown’s double exponential smoothing method in time series analysis | |
Mlakar et al. | GP-DEMO: differential evolution for multiobjective optimization based on Gaussian process models | |
CN105260390B (en) | A kind of item recommendation method based on joint probability matrix decomposition towards group | |
CN108710905B (en) | Spare part quantity prediction method and system based on multi-model combination | |
CN107239477B (en) | Geographic data support vector regression method fusing spatial correlation | |
Aggarwal | Compensative weighted averaging aggregation operators | |
CN104751007A (en) | Building value evaluation based calculation method and device | |
Gautam et al. | A novel moving average forecasting approach using fuzzy time series data set | |
Parida et al. | Multiple attributes decision making approach by TOPSIS technique | |
Zhao et al. | The c‐chart with bootstrap adjusted control limits to improve conditional performance | |
CN109886445A (en) | A kind of tomorrow requirement prediction technique based on material requirements property quantification | |
CN110794308A (en) | Method and device for predicting train battery capacity | |
CN108053266A (en) | A kind of patent value predictor method and device | |
JP4675308B2 (en) | Planning support device and planning support method for media mix plan | |
CN106776757A (en) | User completes the indicating means and device of Net silver operation | |
Prastyo et al. | Survival analysis of companies’ delisting time in Indonesian stock exchange using Bayesian multiple-period logit approach | |
Lu et al. | Application of grey relational analysis for evaluating road traffic safety measures: advanced driver assistance systems against infrastructure redesign | |
CN114372618A (en) | Student score prediction method and system, computer equipment and storage medium | |
CN114154252A (en) | Risk assessment method and device for failure mode of power battery system of new energy automobile | |
CN114648178A (en) | Operation and maintenance strategy optimization method of electric energy metering device based on DDPG algorithm | |
CN110728466B (en) | Method for determining target accessory demand of new product and computer equipment |
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 | ||
CB02 | Change of applicant information |
Address after: 100070, No. 101-8, building 1, 31, zone 188, South Fourth Ring Road, Beijing, Fengtai District Applicant after: Guoxin Youyi Data Co., Ltd Address before: 100070, No. 188, building 31, headquarters square, South Fourth Ring Road West, Fengtai District, Beijing Applicant before: SIC YOUE DATA Co.,Ltd. |
|
CB02 | Change of applicant information | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180518 |
|
RJ01 | Rejection of invention patent application after publication |