CN109726401A - A kind of patent portfolios generation method and platform - Google Patents
A kind of patent portfolios generation method and platform Download PDFInfo
- Publication number
- CN109726401A CN109726401A CN201910004140.9A CN201910004140A CN109726401A CN 109726401 A CN109726401 A CN 109726401A CN 201910004140 A CN201910004140 A CN 201910004140A CN 109726401 A CN109726401 A CN 109726401A
- Authority
- CN
- China
- Prior art keywords
- portfolios
- subset
- index
- sample
- density
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The present invention relates to field of computer technology, a kind of patent portfolios generation method and platform are specifically disclosed, this method comprises: calculating the text similarity in patent set to be grouped between each patent and other patents;Multiple patent subsets are grouped and generated to the patent collection to be grouped according to the text similarity, each patent subset includes multiple patents;Calculate the network correlation density and complementary index of each patent subset;If judging, the network correlation density of the patent subset is greater than pre-set density threshold value and the complementary index of the patent subset is located within the scope of preset complementary index, patent portfolios are generated according to the patent subset, the patent portfolios are that technology correlation patent portfolios or product are associated with patent portfolios.The present invention can be excavated the profound incidence relation between patented technology in unartificial mode and generate patent portfolios, to promote patent transfer and the conversion ratio of the recognition efficiency and enterprise patent of patent portfolios.
Description
Technical field
The present invention relates to field of computer technology, in particular to a kind of patent portfolios generation method and platform.
Background technique
With the arrival of era of knowledge-driven economy, the amount of the application for patent of Chinese Enterprises increases year by year, and enterprise patent pond is constantly rich
It is rich.But then, the patented technology transfer of enterprise patent and conversion ratio are extremely low, and considerable patent is not implemented such that use, real
Test room achievement and technical marketization application between there is huge gaps.For promote patented technology transfer and conversion, need first into
Row patent valve estimating.
Patent value is not embodied in single-piece patent, but is embodied on one group of patent portfolios with internal correlation,
Aggregate value of the value of patent portfolios much larger than single-piece patents all inside patent portfolios.Therefore, patented technology shifts and turns
Change process is frequently not for single-piece or random several patents, but a series of patents with internal correlation is packed into specially
Benefit combination carries out bulk transfers, to obtain maximum economic well-being of workers and staff.Identify patent portfolios to the overall value for promoting enterprise patent
It is of great significance with transfer efficiency.Currently, the identification work of patent portfolios relies primarily on expertise judgement, this mode
It takes time and effort, and is difficult to excavate the profound incidence relation between patented technology, be unfavorable for improving the patent transfer of patented technology
And conversion ratio.
It should be noted that the above description of the technical background be intended merely to it is convenient to technical solution of the present invention carry out it is clear,
Complete explanation, and facilitate the understanding of those skilled in the art and illustrate.Cannot merely because these schemes of the invention
Background technology part is expounded and thinks that above-mentioned technical proposal is known to those skilled in the art.
Summary of the invention
The technical problem to be solved by the present invention is to it is raw to provide a kind of patent portfolios for above-mentioned deficiency in the prior art
At method and platform, the profound incidence relation between patented technology can be excavated in unartificial mode and generates patent group
It closes, to promote patent transfer and the conversion ratio of the recognition efficiency and enterprise patent of patent portfolios.
To achieve the above object, the present invention provides a kind of patent portfolios generation methods, comprising:
Calculate the text similarity in patent set to be grouped between each patent and other patents;
The patent collection to be grouped is grouped according to the text similarity and generates multiple patent subsets, Mei Gezhuan
Sharp subset includes multiple patents;
Calculate the network correlation density and complementary index of each patent subset;
If judging, the network correlation density of the patent subset is greater than pre-set density threshold value and the patent subset
Complementary index is located within the scope of preset complementary index, generates patent portfolios, the patent group according to the patent subset
It is combined into technology correlation patent portfolios or product association patent portfolios.
Optionally, the text similarity calculated in patent set to be grouped between each patent and other patents specifically wraps
It includes:
Theme feature vector corresponding with patent each in patent set to be grouped is generated using LDA model;
Each patent and its are generated in patent set to be grouped according to the theme feature vector sum cosine similarity algorithm
Text similarity between his patent.
Optionally, it is wrapped before the network correlation density of each patent subset of calculating and complementary index
It includes:
Multiple patent portfolios samples are obtained, the patent portfolios sample includes that technology correlation patent portfolios sample and product close
Join patent portfolios sample;
Calculate the network correlation density and complementary index of each patent portfolios sample;
According to the network correlation density of multiple patent portfolios samples and complementary index generation density threshold and mutually
Benefit property indication range, the complementary index range includes that the complementary index range of technology correlation patent portfolios is associated with product
The complementary index range of patent portfolios.
Optionally, density threshold is generated according to the network correlation density of multiple patent portfolios samples and complementary index
Value and complementary index range specifically include:
The network correlation density of multiple patent portfolios samples is averaged or is minimized to generate density threshold
Value;
The complementary index of patent portfolios is associated with according to the complementary index generation technique of technology correlation patent portfolios sample
Range;
The complementary index of product association patent portfolios is generated according to the complementary index that product is associated with patent portfolios sample
Range;
If the network correlation density for judging the patent subset is greater than pre-set density threshold value and patent
The complementary index of collection is located within the scope of preset complementary index, generates patent portfolios according to the patent subset and specifically wraps
It includes:
If judging, the network correlation density of the patent subset is greater than pre-set density threshold value and the patent subset
Complementary index is located within the scope of the complementary index of preset technology correlation patent portfolios, generates skill according to the patent subset
Art is associated with patent portfolios;
If judging, the network correlation density of the patent subset is greater than pre-set density threshold value and the patent subset
Complementary index is located within the scope of the complementary index of preset product association patent portfolios, is generated and is produced according to the patent subset
Product are associated with patent portfolios.
Optionally, pass through formulaNetwork correlation density is calculated, wherein D be patent subset or
The network correlation density of patent portfolios sample, RijFor the text in patent subset or patent portfolios sample between patent i and patent j
Similarity, N are patent total number included by patent subset or patent portfolios sample.
Optionally, the complementary index is generated according to intercrossing index, polymerism index and otherness index;
The intercrossing index passes through formulaIt is calculated, wherein RS is patent
The intercrossing index of collection or patent portfolios sample, piFor patent classification i the patent subset or patent portfolios sample it is all specially
Probability distribution value in sharp classification, pjIt is patent classification j in the patent subset or all patent classifications of patent portfolios sample
Probability distribution value, dijFor the distance value between patent classifications different in patent subset or patent portfolios sample, α and β are metering ginseng
Number;
The polymerism index passes through formulaIt is calculated, wherein CC is patent subset or patent portfolios sample
This polymerism index, 0≤CC≤1, NΔFor the quantity of ternary closure in patent subset or patent portfolios sample, N3For patent
The quantity of triple is connected in collection or patent portfolios sample;
The otherness index passes through formulaIt is calculated, wherein S is patent subset or patent
The otherness index of combined sample, the patent total number that N includes by patent subset or patent portfolios sample, RijFor patent subset
Or drawn intensity altogether between patent i and patent j in patent portfolios sample,
CijTo be drawn the frequency, C altogether between patent i and patent j in patent subset or patent portfolios sampleiAlways drawn frequency for patent i
It is secondary, CjFor the Total cited number of patent j.
To achieve the above object, described pre- the present invention also provides a kind of patent portfolios generating platform, including prediction module
Surveying module includes:
First computing unit, it is similar for calculating text in patent set to be grouped between each patent and other patents
Degree calculates the network correlation density and complementary index of each patent subset;
Judging unit, for judging whether the network correlation density of the patent subset is greater than pre-set density threshold value and institute
Whether the complementary index for stating patent subset is located within the scope of preset complementary index;
First generation unit, for the patent collection to be grouped to be grouped and generated more according to the text similarity
A patent subset, each patent subset include multiple patents, when the network correlation density for judging the patent subset is greater than
When the complementary index of pre-set density threshold value and the patent subset is located within the scope of preset complementary index, according to described special
Sharp subset generates patent portfolios, and the patent portfolios are that technology correlation patent portfolios or product are associated with patent portfolios.
Optionally, first computing unit is specifically used for using each in the generation of LDA model and patent set to be grouped
The corresponding theme feature vector of patent generates patent to be grouped according to the theme feature vector sum cosine similarity algorithm
Text similarity in set between each patent and other patents.
Optionally, which further includes training module, and the training module includes:
Acquiring unit, for obtaining multiple patent portfolios samples, the patent portfolios sample includes technology correlation patent group
It closes sample and is associated with patent portfolios sample with product;
Second computing unit, network correlation density and complementarity for calculating each patent portfolios sample refer to
Mark;
Second generation unit is generated according to the network correlation density of multiple patent portfolios samples and complementary index
Density threshold and complementary index range, the complementary index range include the complementary index model of technology correlation patent portfolios
Enclose the complementary index range that patent portfolios are associated with product.
Optionally, second generation unit is specifically used for the network correlation density to multiple patent portfolios samples
It is averaged or is minimized to generate density threshold, according to the complementary index generation technique of technology correlation patent portfolios sample
It is associated with the complementary index range of patent portfolios, product association is generated according to the complementary index that product is associated with patent portfolios sample
The complementary index range of patent portfolios;
First generation unit is specifically used for presetting when the network correlation density for judging the patent subset is greater than
The complementary index of density threshold and the patent subset is located at the complementary index range of preset technology correlation patent portfolios
When interior, patent portfolios are associated with according to the patent subset generation technique, when the network correlation for judging the patent subset is close
Degree is greater than pre-set density threshold value and the complementary index of the patent subset is located at the complementation that preset product is associated with patent portfolios
Property indication range in when, according to the patent subset generate product be associated with patent portfolios.
The invention has the following advantages:
Patent portfolios generation method provided by the invention treats grouping patent collection according to text similarity and is grouped and gives birth to
At multiple patent subsets, when the network correlation density for judging patent subset is greater than the mutual of pre-set density threshold value and patent subset
When benefit property index is located within the scope of preset complementary index, patent portfolios are generated according to the patent subset, it can be with inhuman
The mode of work excavates the profound incidence relation between patented technology and generates patent portfolios, to promote the identification of patent portfolios
Efficiency and the transfer of the patent of enterprise patent and conversion ratio.
Referring to following description and accompanying drawings, only certain exemplary embodiments of this invention is disclosed in detail, specifies original of the invention
Reason can be in a manner of adopted.It should be understood that embodiments of the present invention are not so limited in range.In appended power
In the range of the spirit and terms that benefit requires, embodiments of the present invention include many changes, modifications and are equal.
The feature for describing and/or showing for a kind of embodiment can be in a manner of same or similar one or more
It uses in a other embodiment, is combined with the feature in other embodiment, or the feature in substitution other embodiment.
It should be emphasized that term "comprises/comprising" refers to the presence of feature, one integral piece, step or component when using herein, but simultaneously
It is not excluded for the presence or additional of one or more other features, one integral piece, step or component.
Detailed description of the invention
Fig. 1 is a kind of flow diagram for patent portfolios generation method that the embodiment of the present invention one provides;
Fig. 2 is a kind of flow diagram of patent portfolios generation method provided by Embodiment 2 of the present invention;
Fig. 3 is a kind of example flow schematic diagram of patent portfolios generation method provided by Embodiment 2 of the present invention;
Fig. 4 is a kind of structural schematic diagram for patent portfolios generating platform that the embodiment of the present invention three provides;
Fig. 5 is a kind of structural schematic diagram for patent portfolios generating platform that the embodiment of the present invention four provides.
Specific embodiment
To make those skilled in the art more fully understand technical solution of the present invention, with reference to the accompanying drawing in the present invention
Technical solution carry out clear, complete description, it is clear that described embodiment is a part of the embodiments of the present invention, without
It is whole embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art are not making creative work
Under the premise of every other embodiment obtained, shall fall within the protection scope of the present invention.
Patent portfolios described in the present embodiment refer to the set of the related patents under same obligee control, patent
Combined important feature is Technology Correlation.Patent portfolios include: the technology correlation patent portfolios based on technology similitude, technology
Association patent portfolios can form patent barrier, the degree of similarity in technology correlation patent portfolios between each patented technology compared with
Height, it is complementary lower;Combined product based on identical product different technologies is associated with patent portfolios, and product is associated in patent portfolios
The complementarity of each patented technology is higher.In the patented technology transfer and conversion of enterprise patent, both the above patent portfolios have
There is important value, by will have the multiple technologies scheme of similar technique or complementary technology to carry out being packed into patent portfolios,
On the one hand patent storage can be vitalized, on the other hand can be improved the overall value and transfer conversion ratio of patented technology.
Technology correlation patent portfolios are associated with the technology complementation of patent portfolios with product there are larger difference, and technology correlation is special
Each patent in benefit combination is unfolded around similar technology, shows stronger collection on patent classification or patent citation relationship
Middle trend and inheritance, thus in technology correlation patent portfolios each patent complementarity it is smaller;Product is associated with each in patent portfolios
Complementarity between patent is relatively large, and each patent in product association patent portfolios is possible to be distributed in different patent classifications
Or the citation relations (such as quotation coupling) between each patented technology are not close.
Embodiment one
Fig. 1 is a kind of flow diagram for patent portfolios generation method that the embodiment of the present invention one provides, as shown in Figure 1,
This method comprises:
Step 101 calculates text similarity in patent set to be grouped between each patent and other patents.
Step 102 is grouped the patent collection to be grouped according to the text similarity and generates multiple patent
Collection, each patent subset includes multiple patents.
Step 103, the network correlation density and complementary index for calculating each patent subset.
Step 104, judge the network correlation density of the patent subset whether be greater than pre-set density threshold value and it is described specially
Whether the complementary index of sharp subset is located within the scope of preset complementary index, if so, executing step 105;If it is not, process knot
Beam.
Step 105, according to the patent subset generate patent portfolios, the patent portfolios be technology correlation patent portfolios or
Product is associated with patent portfolios.
Patent portfolios generation method provided in this embodiment is treated grouping patent collection according to text similarity and is grouped simultaneously
Multiple patent subsets are generated, when the network correlation density for judging patent subset is greater than pre-set density threshold value and patent subset
When complementary index is located within the scope of preset complementary index, patent portfolios are generated according to the patent subset, it can be with non-
Artificial mode excavates the profound incidence relation between patented technology and generates patent portfolios, to promote the knowledge of patent portfolios
Other efficiency and the transfer of the patent of enterprise patent and conversion ratio.
Embodiment two
Fig. 2 is a kind of flow diagram of patent portfolios generation method provided by Embodiment 2 of the present invention, as shown in Fig. 2,
Method includes the following steps:
Step 201 obtains multiple patent portfolios samples, and the patent portfolios sample includes technology correlation patent portfolios sample
Patent portfolios sample is associated with product.
Preferably, each step in the present embodiment is generated by patent portfolios generating platform.
Patent portfolios sample is the patent portfolios approved through art technology expert, includes multiple special in patent portfolios sample
Benefit.Such as: patent portfolios sample can for several patent PP1, PP2s of the A enterprise in technical field T ... PPn.
Step 202, the network correlation density and complementary index for calculating each patent portfolios sample.
The network correlation density of patent portfolios sample is used to indicate the similarity in patent portfolios sample between patent.Patent
Each patent in combined sample is interrelated, and certain similarity is presented in patent text content.Such as: in patent portfolios sample
The technology contents of multiple patents, effect content and purposes content is similar or identical or partial monopoly in include secondary indexing content
(the USE field of such as Derwent database).
Specifically, using LDA topic model, (Latent Dirichlet Allocation, document subject matter generate mould first
Type) theme feature vector corresponding with patent each in patent portfolios sample is generated, further according to institute in patent portfolios sample
There is the theme feature vector sum cosine similarity algorithm of patent to generate in patent portfolios sample between each patent and other patents
Text similarity.Finally, passing through formulaCalculate the network correlation density of patent portfolios sample, wherein D is
The network correlation density of patent portfolios sample, RijFor the text similarity in the patent portfolios sample between patent i and patent j, N
For patent total number included by the patent portfolios sample.Theme feature vector is for constructing text relevant network, text phase
Text relevant network is based on like degree and network correlation density and Social Network Theory generates, and network correlation density indicates special
In sharp combined sample between patent the sum of text similarity and the sum of maximum text similarity ratio.
Optionally, it is right respectively with each patent in patent portfolios sample to be generated using the Gensim Open-Source Tools packet of Python
The theme feature vector answered: from 8 themes and use Gibbs sampling algorithm are extracted in each patent in patent portfolios sample
Whole patents carry out parameter Estimation, theme-Feature Words that entire patent portfolios sample is solved after 1500 iteration are general
Rate distribution situation corresponds to theme feature vector to ultimately generate each patent.Patent is expressed as to the mode of theme feature vector,
It can guarantee the accuracy rate that patent portfolios generate while realizing patent substantially dimensionality reduction.
The present embodiment chooses network correlation density to characterize the technology correlation in patent portfolios sample between patent, abides by
Principle of whole, principle of applicability and simplification principle are followed, network correlation density can more intuitively reflect patent group
Close the technology correlative character in sample between each patent.
The complementary index of patent portfolios sample is used to indicate the technological disparity degree in patent portfolios sample between patent.Patent
The complementary index of combined sample is generated according to intercrossing index, polymerism index and otherness index.Intercrossing index, polymerization
Property index and total alanysis method is respectively adopted in otherness index, Social Network Analysis Method & and citation analyzing generate.
The intercrossing index passes through formulaIt is calculated, wherein RS is patent group
Close the intercrossing index of sample, piFor probability distribution value of the patent classification i in all patent classifications of the patent portfolios sample,
pjFor probability distribution value of the patent classification j in all patent classifications of the patent portfolios sample, dijFor in patent portfolios sample
Distance value between different patent classifications, dijIt is generated using cosine similarity algorithm, α and β are measuring parameter.In intercrossing index
Each parameter is generated especially by the text relevant network that patent portfolios sample constructs.
Specifically, the patent in patent portfolios sample may have multiple Patent classificating numbers, and different Patent classificating numbers is
Corresponding different patent classification, all patent classifications of patent portfolios sample be all patents included by it have it is all not
With the total number of Patent classificating number.Intercrossing index is a kind of comprehensive measurement index, and it can be used to measure patent portfolios
Distance and otherness in sample between the characteristic distributions of each patent, patent classification.In the present embodiment, optionally, by patent portfolios
All patents in sample are expressed as the form of categorization vector to count and generate piAnd pj, the patent classification statistical form that specifically generates
Lattice can be as shown in following table one:
Table one
Certain patent portfolios sample | C1 | C2 | … | CN |
Z1 | 1 | 0 | … | 0 |
Z2 | 1 | 1 | … | 0 |
… | … | … | … | 1 |
ZM | 0 | 1 | … | 0 |
Wherein, wherein CiFor i-th of patent classification in patent portfolios sample, ZjIt is special for the jth part in patent portfolios sample
Benefit, 1 indicates that the patent belongs to a certain patent classification, and 0 indicates that the patent is not belonging to a certain patent classification.
Polymerism index passes through formulaIt is calculated, wherein CC is that the polymerism of patent portfolios sample refers to
Mark, 0≤CC≤1, NΔFor the quantity of ternary closure in patent portfolios sample, N3For the number for being connected to triple in patent portfolios sample
Amount.Specifically, the ternary closure in patent portfolios sample refers in text relevant network constructed by the patent portfolios sample
Ternary closure, each ternary closure can be considered three different connection triples.Polymerism index is related by metering text
The quantity of ternary closure describes the polymerism feature of each patent in patent portfolios sample in property network.N in polymerism indexΔ's
Factor can guarantee 0≤CC≤1 for 3.
Otherness index passes through formulaIt is calculated, wherein S is the difference of patent portfolios sample
Property index, the patent total number that N includes by patent portfolios sample, RijFor in patent portfolios sample between patent i and patent j
Drawn intensity altogether, CijFor being total between patent i and patent j in patent portfolios sample
Drawn the frequency, CiFor the Total cited number of patent i, CjFor the Total cited number of patent j.It is bigger to be drawn the frequency altogether, between expression patent
Citation relations it is closer.Otherness index is the index for having opposite attribute with similarity, and otherness index is bigger to indicate special
Similarity in sharp combined sample between patent is bigger and otherness is smaller, in the smaller expression patent portfolios sample of otherness index
Similarity between patent is smaller and otherness is bigger.
Step 203 generates density according to the network correlation density and complementary index of multiple patent portfolios samples
Threshold value and complementary index range, the complementary index range include technology correlation patent portfolios complementary index range and
The complementary index range of product association patent portfolios.
Specifically, step 203 includes: that the network correlation density of multiple patent portfolios samples is averaged or is taken
Minimum value is to generate density threshold;Patent portfolios are associated with according to the complementary index generation technique of technology correlation patent portfolios sample
Complementary index range;The mutual of product association patent portfolios is generated according to the complementary index that product is associated with patent portfolios sample
Benefit property indication range.
Step 201- step 203 generates model for realizing the training to patent portfolios sample, to establish patent portfolios.This
In embodiment, patent portfolios generate model and are understood that go out patent group according to density threshold and complementary index horizon prediction
It closes.Optionally, patent portfolios generate model and can also be tested, namely test density threshold and complementary index range
And be adjusted according to test result, to obtain optimal density threshold and complementary index range.Such as: patent portfolios generate mould
The test sample of type can be several specially for other except in above-mentioned patent portfolios sample patent in addition to of the A enterprise in technical field T
Benefit.
Density threshold and complementary index range for generating patent portfolios from patent set to be grouped, can according to
The actual conditions of grouping patent are set.Complementary index range includes the complementary index range of technology correlation patent portfolios
The complementary index range of patent portfolios is associated with product.After complementary index range is according to multiple patent portfolios sample weightings
Intercrossing index, polymerism index and otherness index generate, and the present embodiment determines each finger by the way of coefficient of variation weighting
Target weight, the size of weight can be used for indicating that corresponding index is associated with patent portfolios with product in differentiation technology correlation patent portfolios
When influence power.Weighting is taken to the intercrossing index, polymerism index and otherness index of technology correlation patent portfolios sample
The complementary index range for determining technology correlation patent portfolios afterwards, to product association patent portfolios sample intercrossing index,
Polymerism index determines that product is associated with the complementary index range of patent portfolios after taking weighting with otherness index.
Optionally, the weight generating mode of intercrossing index, polymerism index and otherness index includes: to pass through public affairs first
FormulaThe coefficient of variation v of each index is calculatedi, wherein SiFor the standard deviation of i-th of index,Exist for i-th of index
Average value in all patent portfolios samples.Pass through formula againTo the coefficient of variation of index be normalized with
Generate the weight w of each indexi。
Step 204 calculates text similarity in patent set to be grouped between each patent and other patents.
It include multiple patents in patent set to be grouped, this method of the present embodiment is specifically used for from patent set to be grouped
The one or more patent portfolios of middle generation.Such as: patent collection to be grouped is combined into A enterprise and removes above-mentioned patent group in technical field T
Close other all patents in sample outside patent.
Step 204 is distinguished specifically includes the following steps: being generated using LDA model with patent each in patent set to be grouped
Corresponding theme feature vector.It is generated according to the theme feature vector sum cosine similarity algorithm every in patent set to be grouped
Text similarity between a patent and other patents.
The specific method process of step 204 can be found in the text similarity in step 202 in patent combined sample between patent
Generating mode description, details are not described herein again.
Step 205 is grouped the patent collection to be grouped according to the text similarity and generates multiple patent
Collection, each patent subset includes multiple patents.
Specifically, the close multiple patents division of text similarity is generated as a patent subset.In patent subset
Multiple patents belong to same patent subclass.
Step 206, the network correlation density and complementary index for calculating each patent subset.
The network correlation density of patent subset and complementary index generating mode can be found in step 202 in step 206
The network correlation density and complementary index generating mode of patent portfolios sample describe, and details are not described herein again.
Step 207 judges whether the network correlation density of the patent subset is greater than pre-set density threshold value, if so, holding
Row step 208;If it is not, process terminates.
The network correlation density of patent subset is greater than pre-set density threshold value, indicates the similar of each patent in the patent subset
It spends larger, can be used as the candidate of patent portfolios.
Step 208 judges whether the complementary index of the patent subset is located at preset technology correlation patent portfolios
Within the scope of complementary index, if so, executing step 209;If it is not, executing step 210.
Step 209 is associated with patent portfolios according to the patent subset generation technique, and process terminates.
Step 210 judges whether the complementary index of the patent subset is located at preset product association patent portfolios
Within the scope of complementary index, if so, executing step 211;Process terminates.
Step 211 generates product association patent portfolios according to the patent subset, and process terminates.
The much complementary indexes in technology correlation patent portfolios of complementary index of product association patent portfolios.
The patent portfolios generation method of lower the present embodiment is illustrated in detail below:
In patent database retrieval obtain A mechanism the field T all 5452 patents magnificent, it is special through the technical field
Family carries out artificial interpretation to the above patent content, obtains 14 patent portfolios of the enterprise as patent described in the present embodiment
Combined sample.The network correlation density and complementary index for calculating separately 14 patent portfolios samples, according to 14 patent groups
The network correlation density and complementary index for closing sample obtain: density threshold 0.6, the complementarity of technology correlation patent portfolios
Indication range is 0.15 < z≤0.07, and the complementary index range that product is associated with patent portfolios is 0.07≤z < 0.15.Fig. 3 is this
The example flow schematic diagram for a kind of patent portfolios generation method that inventive embodiments two provide, as shown in figure 3, according to the density threshold
Value and complementary index range have finally predicted other 11 patent portfolios with combination potentiality of the enterprise.It is therein certain
A patent portfolios include 167 patents, are related to etching, LCD, liquid crystal display, installation pedestal, photoresist layer, substrate work
The contents such as skill, evaporator, plasma-deposited, the technical theme of the patent portfolios can be summarized as etch process.
The patent portfolios generation method of the offer of the present embodiment is treated grouping patent collection according to text similarity and is grouped
And multiple patent subsets are generated, when the network correlation density for judging patent subset is greater than pre-set density threshold value and patent subset
Complementary index when being located within the scope of preset complementary index, patent portfolios are generated according to the patent subset, can be with
Unartificial mode and computer assisted mode excavate the profound incidence relation between patented technology and generate patent portfolios,
To promote patent transfer and the conversion ratio of the patent portfolios recognition efficiency and enterprise patent of enterprise patent, while can be effective
Save human cost.
It should be noted that although describing the operation of the method for the present invention in the accompanying drawings with particular order, this is not required that
Or hint must execute these operations in this particular order, or have to carry out operation shown in whole and be just able to achieve the phase
The result of prestige.Additionally or alternatively, it is convenient to omit multiple steps are merged into a step and executed by certain steps, and/or will
One step is decomposed into execution of multiple steps.
Embodiment three
Fig. 4 is a kind of structural schematic diagram for patent portfolios generating platform that the embodiment of the present invention three provides, as shown in figure 4,
The platform includes prediction module 1, and the prediction module 1 includes the first computing unit 11, judging unit 12 and the first generation unit
13。
The text that first computing unit 11 is used to calculate in patent set to be grouped between each patent and other patents is similar
Degree calculates the network correlation density and complementary index of each patent subset.Judging unit 12 is described special for judging
Whether the network correlation density of sharp subset is greater than pre-set density threshold value and whether the complementary index of the patent subset is located at
Within the scope of preset complementary index.First generation unit 13 is used for according to the text similarity to the patent collection to be grouped
Multiple patent subsets are grouped and generate, each patent subset includes multiple patents, when the net for judging the patent subset
Network correlation density is greater than pre-set density threshold value and the complementary index of the patent subset is located at preset complementary index model
When enclosing interior, patent portfolios are generated according to the patent subset, the patent portfolios are that technology correlation patent portfolios or product are associated with
Patent portfolios.
The patent portfolios generating platform that the present embodiment three provides can be used for realizing the patent portfolios that above-described embodiment one provides
Generation method.
Patent portfolios generating platform provided in this embodiment, the first generation unit of prediction module are used for similar according to text
Degree treats grouping patent collection and is grouped and generates multiple patent subsets, when the network correlation density for judging patent subset is big
When the complementary index of pre-set density threshold value and patent subset is located within the scope of preset complementary index, according to the patent
Subset generates patent portfolios, and the profound incidence relation between patented technology can be excavated in unartificial mode and generates patent
Combination, to promote patent transfer and the conversion ratio of the recognition efficiency and enterprise patent of patent portfolios.
Example IV
Fig. 5 is a kind of structural schematic diagram for patent portfolios generating platform that the embodiment of the present invention four provides, as shown in figure 5,
The platform includes prediction module 1, and the prediction module 1 includes the first computing unit 11, judging unit 12 and the first generation unit
13。
The text that first computing unit 11 is used to calculate in patent set to be grouped between each patent and other patents is similar
Degree calculates the network correlation density and complementary index of each patent subset.Judging unit 12 is described special for judging
Whether the network correlation density of sharp subset is greater than pre-set density threshold value and whether the complementary index of the patent subset is located at
Within the scope of preset complementary index.First generation unit 13 is used for according to the text similarity to the patent collection to be grouped
Multiple patent subsets are grouped and generate, each patent subset includes multiple patents, when the net for judging the patent subset
Network correlation density is greater than pre-set density threshold value and the complementary index of the patent subset is located at preset complementary index model
When enclosing interior, patent portfolios are generated according to the patent subset, the patent portfolios are that technology correlation patent portfolios or product are associated with
Patent portfolios.
Further, first computing unit 11 is specifically used for using in the generation of LDA model and patent set to be grouped
The corresponding theme feature vector of each patent, generates according to the theme feature vector sum cosine similarity algorithm wait be grouped
Text similarity in patent set between each patent and other patents.
Further, which further includes training module 2, and the training module 2 includes: that acquiring unit 21, second calculates
Unit 22 and the second generation unit 23.
For acquiring unit 21 for obtaining multiple patent portfolios samples, the patent portfolios sample includes technology correlation patent group
It closes sample and is associated with patent portfolios sample with product.Second computing unit 22 is used to calculate the network of each patent portfolios sample
Correlation density and complementary index.Second generation unit 23 is according to the network correlation density of multiple patent portfolios samples
Density threshold is generated with complementary index and complementary index range, the complementary index range include technology correlation patent group
The complementary index range of conjunction is associated with the complementary index range of patent portfolios with product.
Further, second generation unit 23 is specifically used for the network correlation to multiple patent portfolios samples
Density is averaged or is minimized to generate density threshold, is generated according to the complementary index of technology correlation patent portfolios sample
The complementary index range of technology correlation patent portfolios generates product according to the complementary index that product is associated with patent portfolios sample
It is associated with the complementary index range of patent portfolios.First generation unit 13, which is specifically used for working as, judges the patent subset
It is special that the complementary index that network correlation density is greater than pre-set density threshold value and the patent subset is located at preset technology correlation
When within the scope of the complementary index of benefit combination, patent portfolios are associated with according to the patent subset generation technique, it is described when judging
The network correlation density of patent subset is located at preset greater than the complementary index of pre-set density threshold value and the patent subset
When product is associated within the scope of the complementary index of patent portfolios, product is generated according to the patent subset and is associated with patent portfolios.
The patent portfolios generating platform that the present embodiment four provides can be used for realizing the patent portfolios that above-described embodiment two provides
Generation method.
Patent portfolios generating platform provided in this embodiment, the first generation unit of prediction module are used for similar according to text
Degree treats grouping patent collection and is grouped and generates multiple patent subsets, when the network correlation density for judging patent subset is big
When the complementary index of pre-set density threshold value and patent subset is located within the scope of preset complementary index, according to the patent
Subset generates patent portfolios, and the profound incidence relation between patented technology can be excavated in unartificial mode and generates patent
Combination, to promote patent transfer and the conversion ratio of the recognition efficiency and enterprise patent of patent portfolios.
It should be understood by those skilled in the art that, the embodiment of the present invention can provide as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the present invention
Apply the form of example.Moreover, it wherein includes the computer of computer usable program code that the present invention, which can be used in one or more,
The computer program implemented in usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) produces
The form of product.
The present invention be referring to according to the method for the embodiment of the present invention, the process of equipment (system) and computer program product
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
Specific embodiment is applied in the present invention, and principle and implementation of the present invention are described, above embodiments
Explanation be merely used to help understand method and its core concept of the invention;At the same time, for those skilled in the art,
According to the thought of the present invention, there will be changes in the specific implementation manner and application range, and the content of the present specification should not manage
Solution is limitation of the present invention.For those skilled in the art, spirit and substance of the present invention are not being departed from
In the case where, various changes and modifications can be made therein, these variations and modifications are also considered as protection scope of the present invention.
Claims (10)
1. a kind of patent portfolios generation method characterized by comprising
Calculate the text similarity in patent set to be grouped between each patent and other patents;
Multiple patent subsets, each patent are grouped and generated to the patent collection to be grouped according to the text similarity
Collection includes multiple patents;
Calculate the network correlation density and complementary index of each patent subset;
If judging, the network correlation density of the patent subset is greater than the complementation of pre-set density threshold value and the patent subset
Property index be located within the scope of preset complementary index, patent portfolios are generated according to the patent subset, the patent portfolios are
Technology correlation patent portfolios or product are associated with patent portfolios.
2. patent portfolios generation method according to claim 1, which is characterized in that described to calculate in patent set to be grouped
Text similarity between each patent and other patents specifically includes:
Theme feature vector corresponding with patent each in patent set to be grouped is generated using LDA model;
It is special with other that each patent in patent set to be grouped is generated according to the theme feature vector sum cosine similarity algorithm
Text similarity between benefit.
3. patent portfolios generation method according to claim 1, which is characterized in that calculate each patent described
Include: before the network correlation density and complementary index of collection
Multiple patent portfolios samples are obtained, the patent portfolios sample includes that technology correlation patent portfolios sample is associated with specially with product
Sharp combined sample;
Calculate the network correlation density and complementary index of each patent portfolios sample;
Density threshold and complementarity are generated according to the network correlation density of multiple patent portfolios samples and complementary index
Indication range, the complementary index range include that the complementary index range of technology correlation patent portfolios is associated with patent with product
Combined complementary index range.
4. patent portfolios generation method according to claim 3, which is characterized in that according to multiple patent portfolios samples
Network correlation density and complementary index generate density threshold and complementary index range and specifically include:
The network correlation density of multiple patent portfolios samples is averaged or is minimized to generate density threshold;
The complementary index range of patent portfolios is associated with according to the complementary index generation technique of technology correlation patent portfolios sample;
The complementary index range of product association patent portfolios is generated according to the complementary index that product is associated with patent portfolios sample;
If the network correlation density for judging the patent subset is greater than pre-set density threshold value and the patent subset
Complementary index is located within the scope of preset complementary index, generates patent portfolios according to the patent subset and specifically includes:
If judging, the network correlation density of the patent subset is greater than the complementation of pre-set density threshold value and the patent subset
Property index be located within the scope of the complementary index of preset technology correlation patent portfolios, according to the patent subset generation technique close
Join patent portfolios;
If judging, the network correlation density of the patent subset is greater than the complementation of pre-set density threshold value and the patent subset
Property index be located at preset product association patent portfolios complementary index within the scope of, according to the patent subset generate product close
Join patent portfolios.
5. patent portfolios generation method according to claim 1 to 4, which is characterized in that
Pass through formulaNetwork correlation density is calculated, wherein D is patent subset or patent portfolios sample
Network correlation density, RijFor the text similarity in patent subset or patent portfolios sample between patent i and patent j, N is special
Patent total number included by sharp subset or patent portfolios sample.
6. patent portfolios generation method according to claim 1 to 4, which is characterized in that the complementary index according to
Intercrossing index, polymerism index and otherness index generate;
The intercrossing index passes through formulaBe calculated, wherein RS be patent subset or
The intercrossing index of patent portfolios sample, piIt is patent classification i in the patent subset or all patent classes of patent portfolios sample
Probability distribution value in not, pjIt is general in the patent subset or all patent classifications of patent portfolios sample for patent classification j
Rate Distribution Value, dijFor the distance value between patent classifications different in patent subset or patent portfolios sample, α and β are measuring parameter;
The polymerism index passes through formulaIt is calculated, wherein CC is patent subset or patent portfolios sample
Polymerism index, 0≤CC≤1, NΔFor the quantity of ternary closure in patent subset or patent portfolios sample, N3For patent subset or
The quantity of triple is connected in patent portfolios sample;
The otherness index passes through formulaIt is calculated, wherein S is patent subset or patent portfolios
The otherness index of sample, the patent total number that N includes by patent subset or patent portfolios sample, RijFor patent subset or specially
Drawn intensity altogether between patent i and patent j in sharp combined sample,CijFor
Drawn the frequency, C altogether between patent i and patent j in patent subset or patent portfolios sampleiFor the Total cited number of patent i, Cj
For the Total cited number of patent j.
7. a kind of patent portfolios generating platform, which is characterized in that including prediction module, the prediction module includes:
First computing unit is counted for calculating the text similarity in patent set to be grouped between each patent and other patents
Calculate the network correlation density and complementary index of each patent subset;
Judging unit, for judge the network correlation density of the patent subset whether be greater than pre-set density threshold value and it is described specially
Whether the complementary index of sharp subset is located within the scope of preset complementary index;
First generation unit, it is multiple special for being grouped and being generated to the patent collection to be grouped according to the text similarity
Sharp subset, each patent subset include multiple patents, are preset when the network correlation density for judging the patent subset is greater than
When the complementary index of density threshold and the patent subset is located within the scope of preset complementary index, according to patent
Collection generates patent portfolios, and the patent portfolios are that technology correlation patent portfolios or product are associated with patent portfolios.
8. patent portfolios generating platform according to claim 7, which is characterized in that first computing unit is specifically used for
Theme feature vector corresponding with patent each in patent set to be grouped is generated using LDA model, according to the theme
Feature vector and cosine similarity algorithm generate the text similarity in patent set to be grouped between each patent and other patents.
9. patent portfolios generating platform according to claim 7, which is characterized in that the platform further includes training module, institute
Stating training module includes:
Acquiring unit, for obtaining multiple patent portfolios samples, the patent portfolios sample includes technology correlation patent portfolios sample
This is associated with patent portfolios sample with product;
Second computing unit, for calculating the network correlation density and complementary index of each patent portfolios sample;
Second generation unit generates density according to the network correlation density of multiple patent portfolios samples and complementary index
Threshold value and complementary index range, the complementary index range include technology correlation patent portfolios complementary index range and
The complementary index range of product association patent portfolios.
10. patent portfolios generating platform according to claim 9, which is characterized in that
Second generation unit be specifically used for the network correlation density of multiple patent portfolios samples is averaged or
It is minimized to generate density threshold, patent group is associated with according to the complementary index generation technique of technology correlation patent portfolios sample
The complementary index range of conjunction generates product association patent portfolios according to the complementary index that product is associated with patent portfolios sample
Complementary index range;
First generation unit is specifically used for being greater than pre-set density when the network correlation density for judging the patent subset
When the complementary index of threshold value and the patent subset is located within the scope of the complementary index of preset technology correlation patent portfolios,
Patent portfolios are associated with according to the patent subset generation technique, when the network correlation density for judging the patent subset is greater than
The complementary index of pre-set density threshold value and the patent subset is located at the complementary index of preset product association patent portfolios
When in range, product is generated according to the patent subset and is associated with patent portfolios.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910004140.9A CN109726401B (en) | 2019-01-03 | 2019-01-03 | Patent combination generation method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910004140.9A CN109726401B (en) | 2019-01-03 | 2019-01-03 | Patent combination generation method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109726401A true CN109726401A (en) | 2019-05-07 |
CN109726401B CN109726401B (en) | 2022-09-23 |
Family
ID=66298005
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910004140.9A Active CN109726401B (en) | 2019-01-03 | 2019-01-03 | Patent combination generation method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109726401B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2021017640A1 (en) * | 2019-07-26 | 2021-02-04 | 南京瑞祥信息技术有限公司 | Query method of technical digital assets |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101625680A (en) * | 2008-07-09 | 2010-01-13 | 东北大学 | Document retrieval method in patent field |
CN101894170A (en) * | 2010-08-13 | 2010-11-24 | 武汉大学 | Semantic relationship network-based cross-mode information retrieval method |
US20180232361A1 (en) * | 2017-02-15 | 2018-08-16 | Specifio, Inc. | Systems and methods for using machine learning and rules-based algorithms to create a patent specification based on human-provided patent claims such that the patent specification is created without human intervention |
-
2019
- 2019-01-03 CN CN201910004140.9A patent/CN109726401B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101625680A (en) * | 2008-07-09 | 2010-01-13 | 东北大学 | Document retrieval method in patent field |
CN101894170A (en) * | 2010-08-13 | 2010-11-24 | 武汉大学 | Semantic relationship network-based cross-mode information retrieval method |
US20180232361A1 (en) * | 2017-02-15 | 2018-08-16 | Specifio, Inc. | Systems and methods for using machine learning and rules-based algorithms to create a patent specification based on human-provided patent claims such that the patent specification is created without human intervention |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2021017640A1 (en) * | 2019-07-26 | 2021-02-04 | 南京瑞祥信息技术有限公司 | Query method of technical digital assets |
Also Published As
Publication number | Publication date |
---|---|
CN109726401B (en) | 2022-09-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Zhou et al. | Stochastic multicriteria decision‐making approach based on SMAA‐ELECTRE with extended gray numbers | |
Cowan et al. | Asymptotic formulae for likelihood-based tests of new physics | |
Chen | A comparative analysis of score functions for multiple criteria decision making in intuitionistic fuzzy settings | |
CN109117380A (en) | A kind of method for evaluating software quality, device, equipment and readable storage medium storing program for executing | |
Cohen et al. | Iterative plan construction for the workflow satisfiability problem | |
Huang et al. | Flood disaster loss comprehensive evaluation model based on optimization support vector machine | |
Nieto-Silleras et al. | Device-independent randomness generation from several Bell estimators | |
CN106600052A (en) | User attribute and social network detection system based on space-time locus | |
CN110058936A (en) | For determining the method, equipment and computer program product of the stock number of dedicated processes resource | |
Babich et al. | Diquark correlations in baryons on the lattice with overlap quarks | |
CN110059261A (en) | Content recommendation method and device | |
Chukhrova et al. | Nonparametric fuzzy hypothesis testing for quantiles applied to clinical characteristics of COVID‐19 | |
Sui et al. | Distantly supervised relation extraction in federated settings | |
Mouhib et al. | TSMAA‐TRI: A temporal multi‐criteria sorting approach under uncertainty | |
CN109726401A (en) | A kind of patent portfolios generation method and platform | |
CN105224801B (en) | A kind of multiple-factor reservoir reservoir inflow short-period forecast evaluation method | |
Otero‐de‐la‐Roza et al. | A fast and accurate algorithm for QTAIM integration in solids | |
CN109255389A (en) | A kind of equipment evaluation method, device, equipment and readable storage medium storing program for executing | |
Alebrahim et al. | Optimizing functional and quality requirements according to stakeholders’ goals | |
Hunter | Two dimensional uncertainty in persuadee modelling in argumentation | |
Sproston | Probabilistic timed automata with clock-dependent probabilities | |
Olver | Shifted GMRES for oscillatory integrals | |
Wang et al. | An enhancing many-objective evolutionary algorithm using chaotic mapping and solution ranking mechanism for large-scale optimization | |
Liu et al. | Personal Credit Evaluation Under the Big Data and Internet Background Based on Group Character | |
Li et al. | Framework information based Java software architecture recovery |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |