CN109614396A - A kind of method for cleaning of address data structure and standardization - Google Patents
A kind of method for cleaning of address data structure and standardization Download PDFInfo
- Publication number
- CN109614396A CN109614396A CN201811543929.3A CN201811543929A CN109614396A CN 109614396 A CN109614396 A CN 109614396A CN 201811543929 A CN201811543929 A CN 201811543929A CN 109614396 A CN109614396 A CN 109614396A
- Authority
- CN
- China
- Prior art keywords
- address
- cleaning
- classification
- standardization
- data structure
- 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
- 238000000034 method Methods 0.000 title claims abstract description 52
- 238000004140 cleaning Methods 0.000 title claims abstract description 23
- 230000005611 electricity Effects 0.000 claims abstract description 40
- 230000008569 process Effects 0.000 claims abstract description 8
- 238000011156 evaluation Methods 0.000 claims abstract description 5
- 238000013517 stratification Methods 0.000 claims abstract description 4
- 238000004364 calculation method Methods 0.000 claims description 10
- 238000004458 analytical method Methods 0.000 claims description 5
- 230000004888 barrier function Effects 0.000 abstract description 4
- 238000012423 maintenance Methods 0.000 abstract description 4
- 230000004044 response Effects 0.000 abstract description 4
- 230000000694 effects Effects 0.000 abstract description 3
- 239000012535 impurity Substances 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000010276 construction Methods 0.000 description 3
- 238000003066 decision tree Methods 0.000 description 2
- 239000000047 product Substances 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 239000013589 supplement Substances 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 238000013107 unsupervised machine learning method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to the method for cleaning of a kind of address data structure and standardization, comprising the following steps: S1: original address text initialization process；S2 original address stratification parsing；S3: the address date of hierarchical parsing is matched with base address dictionary library；S4: being judged according to whether matching degree meets the requirements, and the address date for the hierarchical parsing that matching degree is met the requirements is added to base address dictionary library as clearing achievements, is unsatisfactory for desired data and is returned to S2 into next cleaning circulation；S5: comprehensive assessment is carried out to clearing achievements using the algorithm of similarity and compliance evaluation, to confirm the accuracy and validity of achievement；The present invention can effectively improve the integrality and accuracy rate of customer electricity address；Accuracy that user's report barrier address judges can be improved, the response speed that improves emergency maintenance, the zone user to be affected by power failure send prompting message, grasp zonal power load demand etc. and all having played very important effect.
Description
Technical field
The present invention relates to address data structureizations and standardization field, more particularly, to a kind of address data structure
With the method for cleaning of standardization.
Background technique
In today that urban construction is maked rapid progress, numerous streets, community are planned again and construction, this phenomenon cause to supply
More and more customer electricity address dates and real address are inconsistent in electric enterprise marketing system.In addition to this, it is gone through due to some
History reason causes existing customer electricity address date to there is phenomena such as a large amount of mistake, poikilonymy, imperfect information, such as will
Table number is as address, cell, Lou Dong etc. without standard appellation etc.；Simultaneously as the customer electricity address date of storage is not knot
The data of structure, there are the regular inconsistent or even same cell difference development periods, difference that the customer address of different community is filled in
Fill in rule all inconsistence problems in address between Lou Dong；These problems have seriously affected customer service work, emergency check man
The quality of work also produces serious influence to all kinds of supporting system for analysis and decision making construction carried out based on address date.
Summary of the invention
The present invention is to overcome customer electricity address described in the above-mentioned prior art imperfect, inaccurate, user's report barrier address
Accuracy of judgement degree is not high, and emergency maintenance response speed is slow, cannot be that the zone user transmission prompting message etc. being affected by power failure lacks
It falls into, the method for cleaning of a kind of address data structure and standardization is provided.
It the described method comprises the following steps:
S1: power supply enterprise's storage customer electricity original address data are obtained, and carry out initialization process；
S2: stratification parsing is carried out to power supply enterprise's storage customer electricity original address data after initialization；
S3: the address date of hierarchical parsing is matched with base address dictionary library；
S4: according to the matching degree between the address date of hierarchical parsing and base address dictionary library whether meet the requirements into
Row judgement；
The address date for the hierarchical parsing that matching degree is met the requirements is added to base address dictionary as clearing achievements
Library；
Matching degree is unsatisfactory for desired data return S2 and parse again in next cleaning circulation, until some cleaning
Until circulation cannot obtain the address date for meeting matching degree requirement again.
S5: the algorithm of building similarity and compliance evaluation, and comprehensive assessment is carried out to clearing achievements.
This method is the treatment process repeatedly recycled, and the achievement cleared up every time can all be used to supplement with modified basis
Location dictionary library, it is whole until completing then with the treatment process for participating in next round by supplement and modified base address dictionary library
A scale removal process.
All customer electricity addresses can be carried out the processing of structuring and standardization by the present invention, realize administrative region, street
It does, the name of cell unification, i.e., customer electricity address is uniformly processed and is stated are as follows: city+district+street+cell+Lou Dong+door
The form (not cell road+road form can be used) of the trade mark effectively increases customer electricity address integrality and quasi-
True rate；Improving accuracy, the response speed of raising emergency maintenance, the region to be affected by power failure that user's report barrier address judges
User sends prompting message, the zonal power load demand of grasp etc. and has played very important effect.
Preferably, the analytic method of the step S2 is to be improved to the participle based on text feature by traditional segmenting method
Method.
It is preferably based on the segmenting method of text feature are as follows: on the basis of understanding " segmenting method based on statistics ", into
The extension of row algorithm, except applying frequency (DF), increases information gain (IG), mutual information, X2Count (CHI), expectation intersects
Four kinds of methods of entropy (CE).
Preferably, in step S5 the algorithm of similarity and compliance evaluation pass through comprehensive clustering algorithm, k nearest neighbor algorithm,
CART classification tree regression algorithm constructs.
Preferably, information gain is occurs in electricity consumption address by counting some characteristic item or the number that does not occur is come in advance
The classification of electricity consumption address is surveyed, the calculation formula of information gain is as follows:
Wherein Pr(ci) indicate the probability that feature occurs in the sample, Pr(ci| t) indicate each in the case that feature occurs
The probability of classification is how many respectively.
Information gain G (t) reflects reduction of the feature t to classification confusion degree, that is, to the information content of classification in reality
By being sorted according to the information gain value of each feature in existing, and it is sub according to the feature that the threshold value of setting selects proper size
Collection.
Preferably, the association relationship of mutual information is completed to extract by calculating the correlation between feature t and classification c；It calculates
Formula are as follows:
Wherein: A is the number that t and c occurs simultaneously；B is that t occurs and number that c does not occur；C is c appearance and t does not have
The number of appearance；N is all electricity consumption number of addresses；If t and c are uncorrelated, I (t, c) value is 0；It is then right if there is m class
M value is had in each t, takes being averaged for they, so that it may obtain a linear order needed for Feature Selection；I average value is bigger
The probability that feature is selected is bigger.
Preferably, χ2The calculation formula of statistics can be expressed as:.
Wherein, t indicates that characteristic item, c indicate classification.
Preferably, it is expected that intersect closely related calculation formula as follows,
Wherein Pr(ci| t) and P (ci) the same information gain of meaning；If entry and electricity consumption address classes strong correlation, also
It is Pr(ci| t) greatly, and corresponding classification probability of occurrence is small, then illustrates that influence of the entry to classification is big, corresponding CE value is with regard to big, quilt
Choose the possibility as characteristic item bigger；It is expected that intersecting the closely related probability distribution for reflecting text categories and some spy occurring
Determine the distance between the probability distribution of text categories under conditions of word, the expectation of entry t intersects closely related bigger, is distributed to text categories
Influence it is also bigger.
Compared with prior art, the beneficial effect of technical solution of the present invention is: providing a kind of address date for power supply enterprise
The method for cleaning of structuring and standardization realizes that administrative region, neighbourhood committee, the name of cell are unified, effectively increases client's use
Electric address integrality and accuracy rate；In the accuracy for improving the judgement of user's report barrier address, the response speed for improving emergency maintenance, it is
The zone user that is affected by power failure send prompting message, grasp zonal power load demand etc. all played it is very heavy
The effect wanted.
Detailed description of the invention
Fig. 1 is the method for cleaning flow chart of the present embodiment address data structure and standardization.
Specific embodiment
The attached figures are only used for illustrative purposes and cannot be understood as limitating the patent；
In order to better illustrate this embodiment, the certain components of attached drawing have omission, zoom in or out, and do not represent actual product
Size；
To those skilled in the art, it is to be understood that certain known features and its explanation, which may be omitted, in attached drawing
's.
The following further describes the technical solution of the present invention with reference to the accompanying drawings and examples.
The present embodiment provides the method for cleaning of a kind of address data structure and standardization, as shown in Figure 1, the method packet
Include following steps:
S1: power supply enterprise's storage customer electricity original address data are obtained, and carry out initialization process；
S2: to " segmenting method based on statistics " understand on the basis of, carry out algorithm extension, applying frequency (DF) it
Outside, information gain (IG), mutual information, χ are increased2(CHI), expectation cross entropy (CE) four kinds of methods are counted, by traditional participle
Method, which is improved, becomes the segmenting method based on text feature, to realize to power supply enterprise's storage customer electricity original address data
Carry out stratification parsing；Specific analytic method specification:
DF (Document frequency): it can be expressed as electricity consumption address frequency herein；DF is indicated
The electricity consumption number of addresses of some characteristic item t.This method for measuring characteristic item significance level is based on such a hypothesis: DF is lesser
Influence of the characteristic item to classification results is smaller；This method preferentially takes the biggish characteristic item of DF, and the lesser characteristic item of DF will be by
It rejects.
Information gain (IG): IG occurs in electricity consumption address by counting some characteristic item or the number that does not occur is predicted
The calculation formula of the classification of electricity consumption address, IG is as follows:
Wherein Pr(ci) indicate the probability that feature occurs in the sample, Pr(ci| t) indicate each in the case that feature occurs
The probability of classification is how many respectively, and m indicates the number of classification.
Information gain G (t) reflects reduction of the feature t to classification confusion degree, that is, the information content to classification.In reality
By being sorted according to the information gain value of each feature in existing, and it is sub according to the feature that the threshold value of setting selects proper size
Collection.
Mutual information Ml (Mutual information): association relationship, it is related between classification c by calculating feature t
Property is completed to extract, calculation formula are as follows:
Wherein: A is the number that t and c occurs simultaneously；B is that t occurs and number that c does not occur；C is c appearance and t does not have
The number of appearance.N is all electricity consumption number of addresses；If t and c are uncorrelated, I (t, c) value is 0；If there is m classification, then
M value is had for each t, takes being averaged for they, so that it may obtain a linear order needed for Feature Selection；Big I average value
Feature a possibility that being selected it is big.
χ2Count (CHI): CHI method has thought substantially similar with Ml method, same by calculating feature t and classification c
Between degree of dependence complete to extract；If characteristic item t and classification c inverse correlation, just illustrate the electricity consumption address containing characteristic item t not
The probability for belonging to c wants larger, this is also to have very much directive significance for judging whether electricity consumption address is not belonging to classification；To overcome
This defect, CHI calculate the correlation of characteristic item t and classification c using formula；Calculation formula can be expressed as:.
It is expected that cross entropy (CE): expectation intersects closely related (CE) and is defined as follows,
Wherein Pr(ci| t) and P (ci) the same information gain of meaning；If entry and electricity consumption address classes strong correlation, also
It is Pr(ci| t) greatly, and corresponding classification probability of occurrence is small, then illustrates that influence of the entry to classification is big, corresponding CE value is with regard to big, just
It is likely to selected as characteristic item；It is expected that intersecting the closely related probability distribution for reflecting text categories and some specific word occurring
Under conditions of text categories the distance between probability distribution；The expectation of entry t intersects closely related bigger, to be distributed to text categories shadow
Sound is also bigger.
S3: the address date of hierarchical parsing is matched with base address dictionary library, the level that matching degree is met the requirements
Dissolve the address date of analysis as clearing achievements, and is added to base address dictionary library；
S4: the data that matching degree is unsatisfactory for requiring are put into next cleaning circulation and parse again, until some cleaning follows
Until ring cannot obtain the address date for meeting matching degree requirement again；
S5: comprehensive clustering algorithm, k nearest neighbor algorithm, CART classification tree regression algorithm, building similarity and compliance evaluation
Algorithm, and comprehensive assessment is carried out to clearing achievements；Specific method is described as follows:
Clustering algorithm: similar electricity consumption address similarity is larger under normal circumstances, and inhomogeneous electricity consumption address similarity
It is smaller.As a kind of unsupervised machine learning method, cluster is not due to needing training process, and does not need in advance to text
Mark classification by hand, therefore there is certain flexibility and higher automatic processing ability.
One electricity consumption address shows as one and is made of word, word and number, in terms of most famous information retrieval can be used
Electricity consumption address is expressed as weighted feature vector D=D (T1, W1 by vector space model (vector space model, VSM)；
T2, W2；…；Tn, Wn), then, the classification of sample to be divided is determined by calculating the method for electricity consumption address similarity.Work as electricity consumption
When address is represented as vector space model, the similarity of electricity consumption address can be come by the inner product between feature vector
It indicates.Most electricity consumption address can be regarded as and be made of several words in simple terms, each word be converted to weight with
Afterwards, each weight can regard the one-component in vector as, then an electricity consumption address can regard one in n-dimensional space as
Vector, here it is the origin of vector space model；The corresponding weight of word can be calculated by TF-IDF weighting technique.
CART post-class processing: being a kind of Decision-Tree Method, estimates letter using the gini index based on minimum range
Number, for determining the expansion shape of the decision tree generated by the Sub Data Set；In the method, key is to examine some address sample
The Geordie impurity level of the post-class processing of this collection；Geordie impurity level indicates that an address sample chosen at random is divided in the subsets
A possibility that wrong (such as a customer electricity address is assigned to a wrong cell)；Geordie impurity level is selected for this sample
In probability multiplied by it by the probability of misclassification.When all samples are all a classes in a node, Geordie impurity level is zero.
K nearest neighbor algorithm: its core concept is if big in the K in feature space most adjacent samples of a sample
Majority belongs to some classification, then the sample also belongs to this classification, and the characteristic with sample in this classification；It is close using K
Key factor of the adjacent algorithm when assessing the consistency of an address sample set is its distance function.It applies in the method
Minkowski Distance formula:
Wherein, xi、yiFor two-dimentional variable, p indicates variable element.
The terms describing the positional relationship in the drawings are only for illustration, should not be understood as the limitation to this patent；
Obviously, the above embodiment of the present invention be only to clearly illustrate example of the present invention, and not be pair
The restriction of embodiments of the present invention.For those of ordinary skill in the art, may be used also on the basis of the above description
To make other variations or changes in different ways.There is no necessity and possibility to exhaust all the enbodiments.It is all this
Made any modifications, equivalent replacements, and improvements etc., should be included in the claims in the present invention within the spirit and principle of invention
Protection scope within.
Claims (8)
1. the method for cleaning of a kind of address data structure and standardization, which is characterized in that the described method comprises the following steps:
S1: power supply enterprise's storage customer electricity original address data are obtained, and carry out initialization process；
S2: stratification parsing is carried out to power supply enterprise's storage customer electricity original address data after initialization；
S3: the address date of hierarchical parsing is matched with base address dictionary library；
S4: sentenced according to whether the matching degree between the address date of hierarchical parsing and base address dictionary library meets the requirements
It is disconnected；
The address date for the hierarchical parsing that matching degree is met the requirements is added to base address dictionary library as clearing achievements；
Matching degree is unsatisfactory for desired data return S2 and parse again in next cleaning circulation, until some cleaning recycles
Until the address date for meeting matching degree requirement cannot be obtained again；
S5: assessing similarity and consistency, and carries out comprehensive assessment to clearing achievements.
2. the method for cleaning of address data structure according to claim 1 and standardization, which is characterized in that the step
The analytic method of S2 is the segmenting method based on text feature.
3. the method for cleaning of address data structure according to claim 2 and standardization, which is characterized in that be based on text
The segmenting method of feature are as follows: on the basis of " segmenting method based on statistics ", algorithm extension is carried out, except applying frequency,
Information gain, mutual information, χ are increased simultaneously2Statistics, expectation four kinds of methods of cross entropy.
4. the method for cleaning of address data structure according to claim 1 and standardization, which is characterized in that in step S5
The algorithm of similarity and compliance evaluation is by comprehensive k nearest neighbor algorithm, comprehensive clustering algorithm, CART classification tree regression algorithm come structure
It builds.
5. the method for cleaning of address data structure according to claim 3 and standardization, which is characterized in that information gain
It is the number for occurring in electricity consumption address by counting some characteristic item or not occurring the classification of predicting electricity consumption address, information increases
The calculation formula of benefit is as follows:
Wherein Pr(ci) indicate the probability that feature occurs in the sample, Pr(ci| t) indicate each classification in the case that feature occurs
Probability is how many respectively, and m indicates the number of classification.
6. the method for cleaning of address data structure according to claim 3 and standardization, which is characterized in that mutual information
Association relationship is completed to extract by calculating the correlation between feature t and classification c；Calculation formula are as follows:
Wherein: A is the number that t and c occurs simultaneously；B is that t occurs and number that c does not occur；C is c appearance and t does not occur
Number；N is all electricity consumption number of addresses；If t and c are uncorrelated, I (t, c) value is 0.
7. the method for cleaning of address data structure according to claim 3 and standardization, which is characterized in that χ2Statistics
Calculation formula can be expressed as:
Wherein, t indicates characteristic item and c indicates classification, and A is the number that t and c occurs simultaneously；B is that t occurs and c does not occur time
Number；C is that c occurs and number that t does not occur；N is all electricity consumption number of addresses.
8. the method for cleaning of address data structure according to claim 3 and standardization, which is characterized in that expectation intersects
Closely related calculation formula is as follows,
Wherein Pr(ci) indicate the probability that feature occurs in the sample, Pr(ci| t) indicate each classification in the case that feature occurs
Probability is how many respectively, and m indicates the number of classification.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811543929.3A CN109614396A (en) | 2018-12-17 | 2018-12-17 | A kind of method for cleaning of address data structure and standardization |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811543929.3A CN109614396A (en) | 2018-12-17 | 2018-12-17 | A kind of method for cleaning of address data structure and standardization |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109614396A true CN109614396A (en) | 2019-04-12 |
Family
ID=66010485
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811543929.3A Pending CN109614396A (en) | 2018-12-17 | 2018-12-17 | A kind of method for cleaning of address data structure and standardization |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109614396A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113448923A (en) * | 2020-04-17 | 2021-09-28 | 北京新氧科技有限公司 | File generation method and device and terminal |
| CN115168548A (en) * | 2022-09-05 | 2022-10-11 | 吉奥时空信息技术股份有限公司 | Recall-sorting based address matching method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102169498A (en) * | 2011-04-14 | 2011-08-31 | 中国测绘科学研究院 | Address model constructing method and address matching method and system |
| CN106055650A (en) * | 2016-05-31 | 2016-10-26 | 深圳市永兴元科技有限公司 | Address standardization method and device |
| CN106709065A (en) * | 2017-01-19 | 2017-05-24 | 国家电网公司 | Standardization processing method and standardized processing device for address information |
| CN107329950A (en) * | 2017-06-13 | 2017-11-07 | 武汉工程大学 | It is a kind of based on the Chinese address segmenting method without dictionary |
| CN108228825A (en) * | 2018-01-02 | 2018-06-29 | 北京市燃气集团有限责任公司 | A kind of station address data cleaning method based on participle |
| CN109190997A (en) * | 2018-09-18 | 2019-01-11 | 广东电网有限责任公司 | Chinese address hierarchical analysis and standard processing method and system |
-
2018
- 2018-12-17 CN CN201811543929.3A patent/CN109614396A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102169498A (en) * | 2011-04-14 | 2011-08-31 | 中国测绘科学研究院 | Address model constructing method and address matching method and system |
| CN106055650A (en) * | 2016-05-31 | 2016-10-26 | 深圳市永兴元科技有限公司 | Address standardization method and device |
| CN106709065A (en) * | 2017-01-19 | 2017-05-24 | 国家电网公司 | Standardization processing method and standardized processing device for address information |
| CN107329950A (en) * | 2017-06-13 | 2017-11-07 | 武汉工程大学 | It is a kind of based on the Chinese address segmenting method without dictionary |
| CN108228825A (en) * | 2018-01-02 | 2018-06-29 | 北京市燃气集团有限责任公司 | A kind of station address data cleaning method based on participle |
| CN109190997A (en) * | 2018-09-18 | 2019-01-11 | 广东电网有限责任公司 | Chinese address hierarchical analysis and standard processing method and system |
Non-Patent Citations (2)
| Title |
|---|
| 宋才华等: "供电企业存量客户用电地址数据结构化与规范化的清理方法研究", 《微型电脑应用》 * |
| 高文雅: "中文文本分类中的关键技术研究", 《中国优秀硕士论文全文数据库信息科技辑》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113448923A (en) * | 2020-04-17 | 2021-09-28 | 北京新氧科技有限公司 | File generation method and device and terminal |
| CN113448923B (en) * | 2020-04-17 | 2023-09-12 | 北京新氧科技有限公司 | File generation method, device and terminal |
| CN115168548A (en) * | 2022-09-05 | 2022-10-11 | 吉奥时空信息技术股份有限公司 | Recall-sorting based address matching method |
| CN115168548B (en) * | 2022-09-05 | 2022-11-22 | 吉奥时空信息技术股份有限公司 | Recall-sorting based address matching method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104408095B (en) | One kind is based on improved KNN file classification methods | |
| CN104778186B (en) | Merchandise items are mounted to the method and system of standardized product unit | |
| CN108363810A (en) | Text classification method and device | |
| CN110532351A (en) | Recommend word methods of exhibiting, device, equipment and computer readable storage medium | |
| CN108388929A (en) | Client segmentation method and device based on cost-sensitive and semisupervised classification | |
| CN110990718A (en) | Social network model building module of company image improving system | |
| CN108763496A (en) | A kind of sound state data fusion client segmentation algorithm based on grid and density | |
| CN109614396A (en) | A kind of method for cleaning of address data structure and standardization | |
| Sarantitis et al. | A network analysis of the United Kingdom’s consumer price index | |
| CN109658156A (en) | A kind of material price measuring method, device, terminal device and storage medium | |
| KR101625124B1 (en) | The Technology Valuation Model Using Quantitative Patent Analysis | |
| CN113239266B (en) | Personalized recommendation method and system based on local matrix decomposition | |
| CN109190997A (en) | Chinese address hierarchical analysis and standard processing method and system | |
| CN104572623B (en) | A kind of efficient data analysis and summary method of online LDA models | |
| Lejeune et al. | Optimization for simulation: LAD accelerator | |
| Wang et al. | Computer supported data-driven decisions for service personalization: a variable-scale clustering method | |
| CN115630708A (en) | Model updating method and device, electronic equipment, storage medium and product | |
| CN115392351A (en) | Risk user identification method and device, electronic equipment and storage medium | |
| CN108629506A (en) | Modeling method, device, computer equipment and the storage medium of air control model | |
| Keskin et al. | Cohort fertility heterogeneity during the fertility decline period in Turkey | |
| CN113537461A (en) | Network key node discovery method and system based on SIR value learning | |
| Lee et al. | A comparison of the predictive powers of tenure choices between property ownership and renting | |
| SEWART et al. | Graphical models in credit scoring | |
| Schaidnagel et al. | DNA: an online algorithm for credit card fraud detection for games merchants | |
| CN111461199A (en) | Security attribute selection method based on distributed junk mail classified data |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190412 |