CN108764363A - A kind of patented technology Life Cycle Analysis - Google Patents
A kind of patented technology Life Cycle Analysis Download PDFInfo
- Publication number
- CN108764363A CN108764363A CN201810570771.2A CN201810570771A CN108764363A CN 108764363 A CN108764363 A CN 108764363A CN 201810570771 A CN201810570771 A CN 201810570771A CN 108764363 A CN108764363 A CN 108764363A
- Authority
- CN
- China
- Prior art keywords
- principal component
- life cycle
- analysis
- data
- technology life
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention belongs to patent analysis technical fields, relate more specifically to a kind of patented technology Life Cycle Analysis being combined based on constituent analysis and K arest neighbors.The sorting technique is in such a way that feature extraction and sorting algorithm are combined, devise the grader based on principal component analysis and nearest neighbor algorithm, to multiple index analysis of technology life cycle, effective judgment technology degree of innovation, to integrated forecasting technology future developing trend.The redundancy and the complex nature of the problem of index have been lacked in this method letter significantly, to improve the accuracy and speed of patented technology life cycle analysis.
Description
Technical field
The invention belongs to patent analysis technical fields, relate more specifically to one kind and are mutually tied with K arest neighbors based on constituent analysis
The patented technology Life Cycle Analysis of conjunction.
Background technology
Include many recessive technical information in patent information, reveals out by carrying out mining analysis to it, can analyze
The development grain of technology, science determines the direction of interprise's intensive management and effectively promotes patent exploitation efficiency, has and preferably may be used
Letter property and accuracy.Using life cycle residing for patent information judgment technology and forecast analysis is carried out, is the hot spot of current research.
Technology life cycle is the use for describing a technology, is developed from base application science derivative, generally by skill
Art life cycle is divided into four-stage:Budding period, growth stage, maturity period and decline phase.The patent information in each stage can
Show different features, therefore the development trend that can master a skill by the stage residing for technology life cycle makes science and determines
Plan.Currently used analysis method technology life cycle has s curve methods, diagrammatic representation technology life cycle, relative indicatrix method, opposite
Five kinds of methods of growth rate method and TCT, this five kinds of methods are all based on the analysis of single index, such as the Shen of amount of the application for patent or patentee
Please quantity change with time trend, in fact, in fact influence the index of patented technology life cycle not only the two indexs,
The also such as many indexs of Patent classificating number, inventor's quantity, scientific and technological references numbers etc., they are to patented technology Life Cycle
The influence of phase also can not be ignored, and have more reasonability to the research of multi objective.
The usury pellet of Southwest Jiaotong University in 2011 has made the patented technology life cycle research of multi objective, is used in text
Nearest neighbor classifier analytical technology life cycle information, it is relatively simple, but arest neighbors sorting algorithm belongs to inertia algorithm,
Memory overhead is big, and calculation amount is larger when classifying to test sample, it may appear that sample imbalance problem.
Invention content
Place, the present invention propose a kind of patented technology life cycle sorting technique in view of the shortcomings of the prior art,
The sorting technique devises in such a way that feature extraction and sorting algorithm are combined based on principal component analysis and nearest neighbor algorithm
Grader, to multiple index analysis of technology life cycle, effective judgment technology degree of innovation, to integrated forecasting technology not
Carry out development trend.
To achieve the above object, the present invention adopts the following technical scheme that:
A kind of patented technology Life Cycle Analysis, it is characterised in that:
S1, patent file data are collected;
S2, n patent index is chosen, all patent files is analyzed according to n patent index, obtained following original
Data X, m are the time,
S3, smoothing processing, First Year and m annual datas remain unchanged, the data after moving average method calculatingDefinition
It is as follows:
S4, normalized, by the data difference divided by its maximum value in each year after smoothing processing, normalized
Data set definition afterwardsFor:
S5, data after normalization are calculatedCovariance matrix:S=(sij)n*n, wherein
S6, the eigenvalue λ for seeking covariance matrix S1≥λ2≥…λn>0 and corresponding standardization orthogonal eigenvectors V:
Feature vector V=(a1,a2,……,an),
S7, variance contribution ratio t is calculated using characteristic valuei:Each characteristic value divided by characteristic value summation,
I.e.:
S8, cumulative proportion in ANOVA G (r) is calculated:The sum of first variance contribution ratio, the first two variance contribution ratio ... ...,
The sum of all variance contribution ratios,
I.e.:
S9, principal component number r is chosen, and calculates each principal component scores in r principal component:The corresponding feature of principal component to
Amount is multiplied by standardized data battle array,
Fi=a1iY1+a2i Y2+…+aniYn, i=1,2 ..., r.
The technical program further optimizes, and further includes step S10, according to principal component scores, utilizes K nearest neighbor classifiers
Classify to principal component scores, determines the stage residing for patent life cycle.
The technical program further optimizes, and described further includes step S10, and the principal component scores of selection are drawn line chart,
And according to the four stage signature analysis line charts of technology life cycle.
The technical program further optimizes, and the step S8 chooses principal component counting method and is:Accumulate contribution rate G (r)
When >=85%, the minimum value of r.
It is different from the prior art, above-mentioned technical proposal has the following advantages that:
1. the use of Principal Component Analysis can be a small number of overall targets multi objective dimensionality reduction, patent life cycle point be improved
The speed of analysis.
2. the sorting technique classifies to test sample data using K nearest neighbor classifiers, patent Life Cycle is improved
The precision of phase analysis.
Description of the drawings
Fig. 1 is traditional patent multi objective development trend line chart;
Fig. 2 is nano biological sensor principal component scores figure;
Fig. 3 is thin film transistor liquid crystal display technology principal component scores figure;
Fig. 4 is triniscope display technology principal component scores figure.
Specific implementation mode
For the technology contents of technical solution, construction feature, the objects and the effects are described in detail, below in conjunction with specific reality
It applies example and attached drawing is coordinated to be explained in detail.
A present invention preferably embodiment, a kind of patented technology statement periodic classification method, it includes the following steps:
S1, patent file data are collected,
Collect data:The embodiment is the analysis and research to patent data life cycle, it is therefore desirable to from suitable patent
Patent data is extracted in database.Patent database has free and two classes of business.But due to the language of free patent database and
As a result format is various, therefore cannot extract target variable well, so the achievement data of this article is mainly derived from business patent
Database --- Derwent Innovations Index database (abbreviation DII).
It is that research object carries out patented technology with nano biological sensor (Nano-biosensor serves, abbreviation NBS)
The research of life cycle, period are 1985-2008.In addition instruction of two similar prior techniques as K- Nearest Neighbor Classifiers is chosen
Practice collection:Thin film transistor liquid crystal display technology (Thin FilmTransistor Liquid Crystal Display, referred to as
TFT-LCD):The budding period of 1978-1990, the growth stage of 1991-2007;Triniscope display technology (Cathode
RayTube, abbreviation CRT):The maturity period of 1973-2000, the decline phase of 2001-2008.
S2, n patent index is chosen, all patent files is analyzed according to n patent index, obtained following original
Data X, m are the time,
As shown in fig.1, being traditional patent multi objective development trend line chart.Technology lifes of the Fig. 1 based on patent document
It orders in cycle analysis model study paper (usury is red, Southwest Jiaotong University's master thesis, 2011) using patent tradition
The processing of multi-target analysis method.
Traditional patented technology life cycle is mostly referred to the research of patent applicant and number of applications two indices
Marking the less extraction of data influences the objectivity of experimental result, therefore this article influences the factor in patent period by investigation, from DII
In be extracted 13 patent indexes for patented technology life cycle research.13 indexs are respectively:Amount of the application for patent
(indicator 1), priority patent amount (indicator 2), house journal's power people's quantity (indicator 3), machine
Structure patentee quantity (indicator 4), inventor's quantity (indicator 5), science citation quantity (indicator
6), Patent Citation quantity (indicator 7), IPC quantity (indicator 8), IPC top5 quantity (indicator 9),
IPC top10 quantity (indicator 10), MC (indicator 11), MC top5 quantity (indicator 12), MC
Top10 quantity (indicator 13).
The present invention is using nano biological sensor data as embodiment one, and n is the time in the embodiment, and m is patent index, because
This, n=24, m=13 obtain raw data matrix X, as follows:
S3, smoothing processing, First Year and the n-th annual data remain unchanged, the data after moving average method calculatingDefinition is such as
Under:
Data matrix is obtained after raw data matrix X smoothing processingsIt is as follows:
S4, normalized, by the data difference divided by its maximum value in each year after smoothing processing, normalized
Data set definition afterwardsFor:
Data matrixData matrix is obtained after normalizedIt is as follows:
S5, data after normalization are calculatedCovariance matrix:S=(sij)n*n, wherein
Calculate data matrixCovariance matrix S, be worth as follows:
S6, the eigenvalue λ for seeking covariance matrix S1≥λ2≥…λn>0 and corresponding standardization orthogonal eigenvectors V:
Feature vector V=(a1,a2,……,an),
Characteristic value [the λ of embodiment covariance matrix S1,λ2..., λ13]=[1.424067108,0.043200726,
0.003943523,0.000556867,0.000126844,0.000104319,0.000031284,0.000014929,
0.000010130,0.000004721,0.000003165,0.000001953,0.000000626]
The value that the embodiment standardizes orthogonal eigenvectors V is as follows:
S7, variance contribution ratio t is calculated using characteristic valuei:Each characteristic value divided by characteristic value summation, i.e.,:
The embodiment [λ1,λ2..., λ13] corresponding variance contribution ratio be [0.967393390788461,
0.0293469992819856,0.0026789034687487,0.000378289555526965,
0.00008616738865728,0.0000708655354652318,0.0000212519998825547,
0.0000101413639684064,0.00000688135532971816,0.00000320704875106824,
0.00000215026524039108,0.00000132644715443356,0.00000042550082805501]。
S8, cumulative proportion in ANOVA G (r) is calculated:The sum of first variance contribution ratio, the first two variance contribution ratio ... ...,
The sum of all variance contribution ratios, i.e.,:
The embodiment calculates cumulative proportion in ANOVA:0.967393390788,0.996740390070,
0.999419293539,0.999797583095,0.999883750483,0.999954616019,0.999975868019,
0.999986009383,0.999992890738,0.999996097787,0.999998248052,0.999999574,1.
S9, principal component number r is chosen, and calculates each principal component scores in r principal component:The corresponding feature of principal component to
Amount is multiplied by standardized data battle array,
Fi=a1iY1+a2iY2+…+aniYn, i=1,2 ..., r.
Principal component, which is meant, represents original multiple indexs with a small number of indexs of synthesization, plays the role of dimensionality reduction, due to
Principal component scores are the materializations of principal component, and thus principal component scores are exactly the specific evaluation score of principal component.
Choose principal component number r:It is general to be chosen (accumulation contribution rate >=85%) using accumulation contribution rate.
Since first principal component contribution rate is 0.967393390788, noticeably greater than 0.85, therefore only need to extract in the example
First main at r=1.
Embodiment two
The embodiment is with thin film transistor liquid crystal display technology, and n is the time in embodiment, and m is patent index, therefore, n=
30, m=13, raw data matrix X is obtained, it is as follows:
Data matrix is obtained after raw data matrix X smoothing processingsIt is as follows:
Data matrixData matrix is obtained after normalizedIt is as follows:
Calculate data matrixCovariance matrix S, be worth as follows:
Characteristic value [the λ of two covariance matrix S of the embodiment1,λ2..., λ13]=[0.0000000102,
0.0000000406,0.0000048089,0.0000087108,0.0000109586,0.0000439284,
0.0001075078,0.0002337249,0.0010010296,0.0017922160,0.0048910841,
0.0441366666,1.1860855354]
The value of the standardization orthogonal eigenvectors V of the embodiment two is as follows:
Two [λ of the embodiment1,λ2..., λ13] corresponding variance contribution ratio be [0.9578212047,0.0356424844,
0.0039497860,0.0014473007,0.0008083796,0.0001887441,0.0000868177,
0.0000354743,0.0000088496,0.0000070344,0.0000038834,0.0000000328,
0.0000000083]
The embodiment two calculates cumulative proportion in ANOVA:0.957821205,0.993463689,0.997413475,
0.998860776,0.999669155,0.9998579,0.999944717,0.999980192,0.999989041,
0.999996076,0.999999959,0.999999992,1.
Since first principal component contribution rate is 0.9578212047, noticeably greater than 0.85, therefore the need to be only extracted in the example
One main at r=1.
Embodiment three
The embodiment is with triniscope display technology, and n is the time in embodiment, and m is patent index, therefore, n=36, m
=13, raw data matrix X is obtained, it is as follows:
Time | Index 1 | Index 2 | Index 3 | Index 4 | Index 5 | Index 6 | Index 7 | Index 8 | Index 9 | Index 10 | Index 11 | Index 12 | Index 13 |
1973 | 17 | 292 | 94 | 12 | 21 | 14 | 162 | 61 | 259 | 270 | 62 | 9 | 10 |
1974 | 22 | 285 | 84 | 8 | 36 | 12 | 158 | 60 | 247 | 251 | 73 | 12 | 12 |
1975 | 58 | 332 | 101 | 11 | 142 | 17 | 171 | 71 | 260 | 275 | 76 | 13 | 16 |
1976 | 67 | 421 | 121 | 14 | 317 | 14 | 189 | 76 | 320 | 345 | 155 | 26 | 32 |
1977 | 59 | 435 | 114 | 15 | 474 | 23 | 196 | 75 | 316 | 342 | 146 | 32 | 47 |
1978 | 88 | 468 | 125 | 22 | 630 | 33 | 215 | 92 | 323 | 377 | 235 | 31 | 80 |
1979 | 138 | 361 | 112 | 22 | 551 | 49 | 212 | 88 | 291 | 312 | 247 | 21 | 88 |
1980 | 233 | 589 | 138 | 14 | 608 | 41 | 219 | 93 | 530 | 550 | 285 | 26 | 152 |
1981 | 481 | 821 | 148 | 6 | 839 | 54 | 213 | 100 | 748 | 768 | 366 | 240 | 441 |
1982 | 1033 | 1048 | 179 | 11 | 1122 | 84 | 288 | 109 | 955 | 985 | 548 | 510 | 802 |
1983 | 1011 | 974 | 183 | 14 | 636 | 61 | 266 | 116 | 893 | 913 | 419 | 93 | 541 |
1984 | 1009 | 1013 | 202 | 9 | 581 | 57 | 252 | 125 | 942 | 962 | 490 | 83 | 324 |
1985 | 1264 | 1254 | 241 | 7 | 651 | 65 | 329 | 126 | 1149 | 1188 | 575 | 161 | 389 |
1986 | 1444 | 1386 | 271 | 12 | 751 | 83 | 353 | 120 | 1265 | 1294 | 673 | 221 | 401 |
1987 | 1332 | 1258 | 250 | 10 | 761 | 89 | 333 | 127 | 1136 | 1180 | 636 | 211 | 412 |
1988 | 1346 | 1263 | 283 | 8 | 809 | 66 | 325 | 125 | 1165 | 1199 | 643 | 266 | 490 |
1989 | 1269 | 1158 | 255 | 9 | 882 | 78 | 346 | 106 | 1080 | 1104 | 664 | 299 | 490 |
1990 | 1277 | 1194 | 265 | 10 | 878 | 70 | 351 | 120 | 1096 | 1118 | 831 | 623 | 761 |
1991 | 1327 | 1193 | 236 | 15 | 702 | 52 | 328 | 121 | 1133 | 1153 | 818 | 725 | 803 |
1992 | 1424 | 1259 | 246 | 8 | 767 | 82 | 373 | 125 | 1201 | 1216 | 807 | 813 | 874 |
1993 | 1310 | 1162 | 219 | 5 | 809 | 73 | 362 | 118 | 1090 | 1110 | 752 | 814 | 867 |
1994 | 1418 | 1281 | 304 | 8 | 961 | 85 | 432 | 146 | 1131 | 1163 | 947 | 879 | 940 |
1995 | 1555 | 1363 | 312 | 9 | 1036 | 76 | 451 | 152 | 1187 | 1234 | 1051 | 939 | 1008 |
1996 | 2061 | 1851 | 332 | 6 | 1245 | 77 | 540 | 154 | 1699 | 1754 | 1080 | 1284 | 1349 |
1997 | 1727 | 1476 | 285 | 8 | 1288 | 65 | 438 | 151 | 1365 | 1405 | 973 | 1088 | 1123 |
1998 | 1649 | 1439 | 283 | 2 | 1235 | 77 | 410 | 149 | 1327 | 1369 | 901 | 1064 | 1099 |
1999 | 1731 | 1536 | 264 | 4 | 1331 | 65 | 435 | 119 | 1441 | 1475 | 997 | 1068 | 1153 |
2000 | 1917 | 1673 | 301 | 9 | 1671 | 65 | 501 | 127 | 1576 | 1598 | 1016 | 1059 | 1186 |
2001 | 2177 | 1918 | 317 | 17 | 1739 | 85 | 544 | 147 | 1746 | 1802 | 1218 | 1514 | 1587 |
2002 | 1606 | 1382 | 318 | 11 | 1419 | 41 | 413 | 144 | 1241 | 1286 | 1224 | 1118 | 1144 |
2003 | 1142 | 906 | 275 | 16 | 1127 | 43 | 291 | 123 | 800 | 840 | 1074 | 713 | 730 |
2004 | 851 | 651 | 218 | 20 | 1138 | 42 | 227 | 100 | 544 | 597 | 1007 | 455 | 471 |
2005 | 722 | 564 | 225 | 27 | 1072 | 36 | 182 | 103 | 439 | 496 | 993 | 328 | 335 |
2006 | 783 | 677 | 271 | 26 | 1400 | 31 | 165 | 123 | 473 | 589 | 1477 | 463 | 471 |
2007 | 686 | 498 | 277 | 27 | 1265 | 32 | 132 | 115 | 314 | 412 | 1480 | 379 | 380 |
2008 | 455 | 305 | 178 | 12 | 746 | 12 | 66 | 98 | 190 | 244 | 988 | 216 | 219 |
Data matrix is obtained after raw data matrix X smoothing processingsIt is as follows:
Data matrixData matrix is obtained after normalizedIt is as follows:
Calculate data matrixCovariance matrix S, be worth as follows:
Characteristic value [the λ of three covariance matrix S of the embodiment1, λ2..., λ13]=[0.000002781,0.000022738,
0.000096266,0.000177450,0.000291382,0.001175813,0.001776267,0.005921460,
0.006516496,0.013468914,0.036388953,0.113305315,0.756048397]
The value of the standardization orthogonal eigenvectors V of the embodiment three is as follows:
Three [λ of the embodiment1, λ2..., λ13] corresponding variance contribution ratio be [0.808441700,0.121157246,
0.038910666,0.014402295,0.006968082,0.006331810,0.001899360,0.001257295,
0.000311574,0.000189747,0.000102938,0.000024313,0.000002974]
The embodiment three calculates cumulative proportion in ANOVA:0.8084417,0.929598945,0.968509612,
0.982911907,0.989879988,0.996211798,0.998111159,0.999368454,0.999680028,
0.999869775,0.999972713,0.999997026,1.
Since the sum of first and second variance contribution ratios are 0.929598945, noticeably greater than 0.85, therefore need to carry in the example
Take two it is main at r=2.
The variance contribution ratio of first characteristic value is calculated according to characteristic value:The first of thin film transistor liquid crystal display technology
The contribution rate of characteristic value is about 95.78%, and the contribution rate of the First Eigenvalue of triniscope display technology is about:
80.864%, the contribution rate of the First Eigenvalue of nano biological sensor is about:96.74%.
In order to more completely extract the information of original index, generally most when choosing contribution rate of accumulative total and reaching 85% or more
Few principal component number.When triniscope display technology chooses a principal component, contribution rate is less than 85%, so needing again
Second principal component, is chosen, contribution rate is accumulated at this time and meets the requirements to 92.98%.
In order to construct K- Nearest Neighbor Classifiers, training set and test set will choose same amount of principal component number, so
Two principal components are extracted to thin film transistor liquid crystal display technology and nano biological sensor.
It is each to calculate separately thin film transistor liquid crystal display technology, triniscope display technology and nano biological sensor
From the principal component scores of two principal components, as shown in the following table 1 to table 2.
1 thin film transistor liquid crystal display technology first principal component score 1 of table and Second principal component, score 2
2 triniscope display technology first principal component score 1 of table and Second principal component, score 2
3 first principal component score 1 of table and Second principal component, score 2
Thin film transistor liquid crystal display technology, triniscope display technology and nanometer are drawn out according to upper table 1 to table 3
The line chart of biosensor.As shown in Figures 2 to 4, respectively nano biological sensor principal component scores figure, thin film transistor (TFT)
LCD technology principal component scores figure and triniscope display technology principal component scores figure.Score 1 and score 2 in every width figure
Respectively first principal component score and Second principal component, score.
66 instructions are constructed with two principal component scores of thin film transistor liquid crystal display technology and triniscope display technology
Practice collection data, according to thin film transistor liquid crystal display technology (TFT-LCD):1978-1990 budding periods, the 1991-2007 year's harvest
Long-term and triniscope display technology (CRT):1973-2000 maturity periods, the existing proof system of 2001-2008 decline phases
Make point 66 training set data labels, as shown in table 4.Wherein 1 it is the budding period, 2 be the growth stage, 3 be the maturity period and 4 is decline
Phase.
4 66 training set data labels of table
Nano biological sensor is tested after building K- nearest neighbor classifiers according to the above training set data and training set label
24 score datas, show that the label of 24 years (1985-2008) is as shown in table 5 below.
5 24 test set labels of table
1885 | 1986 | 1987 | 1988 | 1989 | 1990 |
1 | 1 | 1 | 1 | 1 | 1 |
1991 | 1992 | 1993 | 1994 | 1995 | 1996 |
1 | 1 | 1 | 1 | 1 | 1 |
1997 | 1998 | 1999 | 2000 | 2001 | 2002 |
2 | 2 | 2 | 2 | 2 | 2 |
2003 | 2004 | 2005 | 2006 | 2007 | 2008 |
2 | 2 | 2 | 2 | 2 | 2 |
Technology life cycle analytical model research one of the usury pellet of the result and Southwest Jiaotong University based on patent document
The result of study of text is completely the same.Illustrating can be rationally anti-by the method that principal component analysis extraction feature and K arest neighbors are combined
Reflect actual conditions.In addition, the redundancy and the complex nature of the problem of index have been lacked in this method letter significantly, to improve patented technology
The accuracy and speed of life cycle analysis.
It should be noted that herein, relational terms such as first and second and the like are used merely to a reality
Body or operation are distinguished with another entity or operation, are deposited without necessarily requiring or implying between these entities or operation
In any actual relationship or order or sequence.Moreover, the terms "include", "comprise" or its any other variant are intended to
Non-exclusive inclusion, so that process, method, article or terminal device including a series of elements include not only those
Element, but also include other elements that are not explicitly listed, or further include for this process, method, article or end
The intrinsic element of end equipment.In the absence of more restrictions, being limited by sentence " including ... " or " including ... "
Element, it is not excluded that there is also other elements in process, method, article or the terminal device including the element.This
Outside, herein, " being more than ", " being less than ", " being more than " etc. are interpreted as not including this number;" more than ", " following ", " within " etc. understandings
It includes this number to be.
Although the various embodiments described above are described, once a person skilled in the art knows basic wounds
The property made concept, then additional changes and modifications can be made to these embodiments, so example the above is only the implementation of the present invention,
It is not intended to limit the scope of patent protection of the present invention, it is every to utilize equivalent structure made by description of the invention and accompanying drawing content
Or equivalent process transformation, it is applied directly or indirectly in other relevant technical fields, the patent for being similarly included in the present invention
Within protection domain.
Claims (4)
1. a kind of patented technology Life Cycle Analysis, it is characterised in that:
S1, patent file data are collected;
S2, n patent index is chosen, all patent files is analyzed according to n patent index, obtain following initial data
X, m are the time,
S3, smoothing processing, First Year and m annual datas remain unchanged, the data after moving average method calculatingIt is defined as follows:
S4, normalized, respectively divided by its maximum value in each year by the data after smoothing processing, after normalized
Data set definitionFor:
S5, data after normalization are calculatedCovariance matrix:S=(sij)n*n, wherein
S6, the eigenvalue λ for seeking covariance matrix S1≥λ2≥…λn>0 and corresponding standardization orthogonal eigenvectors V:
Feature vector V=(a1,a2,……,an),
S7, variance contribution ratio t is calculated using characteristic valuei:Each characteristic value divided by characteristic value summation,
I.e.:
S8, cumulative proportion in ANOVA G (r) is calculated:The sum of first variance contribution ratio, the first two variance contribution ratio ... ... own
The sum of variance contribution ratio,
I.e.:
S9, principal component number r is chosen, and calculates each principal component scores in r principal component:The corresponding feature vector of principal component multiplies
With standardized data battle array,
Fi=a1iY1+a2i Y2+…+aniYn, i=1,2 ..., r.
2. a kind of patented technology Life Cycle Analysis as described in claim 1, it is characterised in that:
Further include step S10, according to principal component scores, is classified to principal component scores using K nearest neighbor classifiers, determine that patent is given birth to
Order the stage residing for the period.
3. a kind of patented technology Life Cycle Analysis as described in claim 1, it is characterised in that:
Described further includes step S10, the principal component scores of selection is drawn line chart, and according to four stages of technology life cycle spy
Sign analysis line chart.
4. a kind of patented technology Life Cycle Analysis as described in claim 1, it is characterised in that:
The step S8 chooses principal component counting method:When accumulating contribution rate G (r) >=85%, the minimum value of r.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810570771.2A CN108764363A (en) | 2018-06-05 | 2018-06-05 | A kind of patented technology Life Cycle Analysis |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810570771.2A CN108764363A (en) | 2018-06-05 | 2018-06-05 | A kind of patented technology Life Cycle Analysis |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108764363A true CN108764363A (en) | 2018-11-06 |
Family
ID=63999992
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810570771.2A Pending CN108764363A (en) | 2018-06-05 | 2018-06-05 | A kind of patented technology Life Cycle Analysis |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108764363A (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070214004A1 (en) * | 2006-01-16 | 2007-09-13 | Hon Hai Precision Industry Co., Ltd. | System and method for managing the lifecycle of a patent case |
US20130086043A1 (en) * | 2011-10-03 | 2013-04-04 | Steven W. Lundberg | System and method for tracking patent ownership change |
CN104160417A (en) * | 2012-02-24 | 2014-11-19 | 艾提普发展有限责任公司 | Patent life cycle management system |
CN105677907A (en) * | 2016-02-16 | 2016-06-15 | 大连理工大学 | Patent technology evolution analysis method and system |
CN105741086A (en) * | 2016-02-05 | 2016-07-06 | 广州市国科禾路信息技术有限公司 | Full-life-circle patent reminding system and realization method therefor |
CN107644290A (en) * | 2017-09-05 | 2018-01-30 | 河北工程大学 | A kind of patented technology Life Cycle Analysis |
-
2018
- 2018-06-05 CN CN201810570771.2A patent/CN108764363A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070214004A1 (en) * | 2006-01-16 | 2007-09-13 | Hon Hai Precision Industry Co., Ltd. | System and method for managing the lifecycle of a patent case |
US20130086043A1 (en) * | 2011-10-03 | 2013-04-04 | Steven W. Lundberg | System and method for tracking patent ownership change |
CN104160417A (en) * | 2012-02-24 | 2014-11-19 | 艾提普发展有限责任公司 | Patent life cycle management system |
CN105741086A (en) * | 2016-02-05 | 2016-07-06 | 广州市国科禾路信息技术有限公司 | Full-life-circle patent reminding system and realization method therefor |
CN105677907A (en) * | 2016-02-16 | 2016-06-15 | 大连理工大学 | Patent technology evolution analysis method and system |
CN107644290A (en) * | 2017-09-05 | 2018-01-30 | 河北工程大学 | A kind of patented technology Life Cycle Analysis |
Non-Patent Citations (3)
Title |
---|
耿文龙 等: "基于主成分分析的我国区域专利保护水平评价研究", 《科技进步与对策》 * |
范宇 等: "基于LDA模型的专利信息聚类技术", 《计算机应用》 * |
高利丹: "基于专利文献的技术生命周期分析模式研究", 《中国优秀硕士学位论文全文数据库 经济与管理科学辑》 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Sathyadevan et al. | Comparative analysis of decision tree algorithms: ID3, C4. 5 and random forest | |
Niu et al. | An innovative integrated model using the singular spectrum analysis and nonlinear multi-layer perceptron network optimized by hybrid intelligent algorithm for short-term load forecasting | |
CN102521656A (en) | Integrated transfer learning method for classification of unbalance samples | |
Bastos et al. | Clustering financial time series with variance ratio statistics | |
CN111680225B (en) | WeChat financial message analysis method and system based on machine learning | |
CN107895322A (en) | A kind of fund combination product is chosen and dynamic monitors the method and system of adjustment | |
CN116644184B (en) | Human resource information management system based on data clustering | |
CN108595414A (en) | Heavy metal-polluted soil enterprise pollution source discrimination based on source remittance space variable reasoning | |
Jiang et al. | Hydrological characteristic-based methodology for dividing flood seasons: an empirical analysis from China | |
CN106204267A (en) | A kind of based on improving k means and the customer segmentation system of neural network clustering | |
Wang et al. | A nonparametric mean estimator for judgment poststratified data | |
Farimani et al. | Leveraging latent economic concepts and sentiments in the news for market prediction | |
CN104599062A (en) | Classification based value evaluation method and system for agricultural scientific and technological achievements | |
CN108764363A (en) | A kind of patented technology Life Cycle Analysis | |
Cova et al. | Iberian universities: a characterisation from ESI rankings | |
CN109033125B (en) | Time sequence data domination set information extraction method | |
CN113505863B (en) | Image multistage classification method and system based on cascade mean vector comprehensive scoring | |
Vinutha et al. | EPCA—enhanced principal component analysis for medical data dimensionality reduction | |
Ramakrishna Sajja et al. | Student Performance Monitoring System Using Decision Tree Classifier | |
Eckardt et al. | Marked spatial point processes: current state and extensions to point processes on linear networks | |
Savov et al. | Towards understanding the evolution of the WWW conference | |
Shaharudin et al. | Choice of cumulative percentage in principal component analysis for regionalization of peninsular Malaysia based on the rainfall amount | |
Nykänen et al. | Maternal and child health care quality assessment: An improved approach using K-means clustering | |
Doppala et al. | A Hybrid Intelligent Approach for Early Identification of Different Diseases in Plants | |
Bhadola et al. | Evolution and dynamics of the currency network |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181106 |
|
RJ01 | Rejection of invention patent application after publication |