CN115936389A - Big data technology-based method for matching evaluation experts with evaluation materials - Google Patents
Big data technology-based method for matching evaluation experts with evaluation materials Download PDFInfo
- Publication number
- CN115936389A CN115936389A CN202211673975.1A CN202211673975A CN115936389A CN 115936389 A CN115936389 A CN 115936389A CN 202211673975 A CN202211673975 A CN 202211673975A CN 115936389 A CN115936389 A CN 115936389A
- Authority
- CN
- China
- Prior art keywords
- expert
- review
- experts
- big data
- matching
- 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
-
- 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
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a matching method of review experts and review materials based on big data technology, which comprises the steps of firstly establishing a big data computing system for thesis review, forming academic portraits of experts through a first-level subject, a second-level subject and research direction information related to big data summarization experts, then establishing an automatic review expert and review material matching evaluation model, carrying out step-by-step matching and respectively assigning on the first-level subject, the second-level subject and the research direction information of the review experts and the review materials, realizing quick and accurate matching of the review materials and the experts through the high and low matching values, and finally carrying out model training and big data supplement on expert information collected in the big data computing system so as to find the review experts most suitable for subdivision fields.
Description
Technical Field
The invention belongs to the technical field of high-grade education thesis review big data in internet education, and particularly relates to a matching method of review experts and review materials based on a big data technology.
Background
In the paper evaluation, evaluation and scoring are carried out on the papers from four dimensions of the selection of questions and the review, the innovation, the paper value, the scientific research capability, the basic knowledge and the paper normalization by evaluation experts, and expert comments are attached.
The traditional evaluation method is to manually select a thesis evaluation expert to carry out submission and the expert feeds back evaluation. The development of new crown epidemic situation and informatization has prompted the network review of the papers. At present, the artificial selection of the experts for thesis evaluation is urgent, the matching degree and efficiency of the experts are to be improved, and the influence of human factors is large, which is not favorable for the fairness and justice of the thesis evaluation, so that how to find the experts for thesis evaluation which are most suitable for the subdivision field and perform the intelligent matching between the thesis and the experts for evaluation is an urgent problem to be solved at present.
Disclosure of Invention
The technical problem to be solved by the invention is to overcome the existing defects, and the invention provides a method for matching an evaluation expert with evaluation materials based on a big data technology, so as to solve the problem of searching a thesis evaluation expert most suitable for a subdivision field to perform intelligent matching of the thesis and the evaluation expert, which is proposed in the background technology.
In order to achieve the purpose, the invention provides the following technical scheme: the intelligent matching of the thesis and the review experts is carried out, and the intelligent matching method comprises the following steps:
the method comprises the following steps: establishing a big data computing system of a thesis review expert;
step two: collecting and classifying expert information collected in a big data computing system, collecting first-level subject, second-level subject and research direction information related to an expert, and forming an academic portrait of the expert;
step three: matching the first-level subject, the second-level subject and the research direction information of the evaluation expert and the evaluation material step by step, respectively assigning values, and constructing an automatic matching evaluation model of the evaluation expert and the evaluation material by using a machine learning algorithm;
step four: training the model in the third step to obtain a trained model;
step five: sending a paper to be reviewed to the expert with the matching degree of more than 85% and the highest matching degree screened from the model;
step six: and (4) putting the expert information of the new field into a warehouse and inputting the expert information into the model, and performing big data supplement on the model so as to find the most suitable evaluation expert for subdividing the field.
Preferably, in the first step, the paper review expert big data computing system mainly computes the matching degree of the paper and the expert; the information of the thesis can be divided into three categories of primary subject, secondary subject and research direction.
Preferably, in the second step, the academic sketch of the expert is an expert accurate sketch based on tag sorting, and the academic expertise of the expert is described by accurate and effective primary subject, secondary subject and research direction tags.
Preferably, in the third step, when the research directions of the experts are classified, the research directions can be systematically classified into three cases, namely, the research direction in which the experts are most adept, the research direction in which the experts are better, and the research direction in which the experts are relatively adept.
Preferably, in the third step, the automated matching evaluation model of the review expert and the review material evaluates and analyzes the research direction which is the most adept by the expert and the research direction of the paper, and scores are given according to the matching degree of the research direction which is the most adept by the expert and the research direction of the paper.
Preferably, in the fourth step, training is performed on the expert and paper matching degree model, and according to the sample data, a group of values (ω, b) needs to be found, so that Loss takes a minimum value, and the derivative of the Loss function at the extreme point is 0, so that a gradient descent method is selected: for convenience of representation, the optimization process of two parameters is taken as an example: l = L (ω 5, ω 9)
The method comprises the following steps:
randomly selecting a set of initial values, for example: [ omega 5, omega 9] = [ -100.0,1-00.0]
Selecting the next point
[ ω 5', ω 9' ], such that: l (ω 5', ω 9') < L (ω 5, ω 9)
Repeating the above steps until the loss function hardly decreases
Among them, the selection of the next point, [ omega 5', omega 9' ] is crucial
(1) Ensuring that L is decreasing;
(2) The tendency to decline is as fast as possible.
Preferably, in the fourth step, in the training expert and paper matching degree model, the gradient calculation code is as follows:
gradient_w0＝(z1-y1)*x1[0]
print(‘gradient_w0{}'.format(gradient_w0))。
preferably, in the fifth step, the matching degree between the experts screened from the model and the paper needs to be more than 85%, and the expert with the highest matching degree reviews the paper.
Preferably, in the sixth step, when the data of the big data computing system is supplemented, the sources of the data include expert information filled by an expert himself or expert information searched through a network.
Compared with the prior art, the invention provides a method for matching review experts with review materials based on big data technology, which has the following beneficial effects:
the invention intelligently analyzes the matching degree of materials and experts through a big data technical system by setting a big data computer system, matches experts through three levels of a first-level subject, a second-level subject and a research direction, carries out statistical distribution on big data by accumulating the big data, carries out academic portrayal on the experts, and finds the experts with the highest matching degree by matching the most adept research direction, the most adept research direction and the relatively adept research direction of the experts through normal distribution.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention provides a technical scheme that: a big data technology-based method for matching review experts with review materials comprises the following steps:
the method comprises the following steps: establishing a big data computing system of a thesis review expert;
step two: collecting and classifying expert information collected in a big data computing system, collecting information of a first-level subject, a second-level subject and research direction related to an expert, and forming an academic portrait of the expert;
step three: matching the first-level subject, the second-level subject and the research direction information of the evaluation expert and the evaluation material step by step, respectively assigning values, and constructing an automatic matching evaluation model of the evaluation expert and the evaluation material by using a machine learning algorithm;
step four: training the model in the third step to obtain a trained model;
step five: sending a paper to be reviewed to the expert with the matching degree higher than 85% and the highest matching degree screened from the model;
step six: and (4) putting the new domain expert information into a warehouse and inputting the information into the model, and performing big data supplement on the model so as to find the most suitable evaluation expert for subdividing the domain.
Preferably, in the first step, the thesis review expert big data computing system mainly computes the matching degree of the thesis and the experts. The information of the thesis can be divided into three categories of primary subject, secondary subject and research direction.
Preferably, in the second step, the academic sketch of the expert is an expert accurate sketch based on tag sorting, and the academic expertise of the expert is described by accurate and effective primary subject, secondary subject and research direction tags.
Preferably, in the third step, when the research directions of the experts are classified, the research directions can be systematically classified into three cases, namely, the research direction in which the experts are most adept, the research direction in which the experts are better, and the research direction in which the experts are relatively adept.
Preferably, in the third step, the automated matching evaluation model of the review expert and the review material performs review analysis between the research direction which is the most adept by the expert and the research direction of the paper, and scores are performed according to the matching degree between the research direction which is the most adept by the expert and the research direction of the paper.
Preferably, in the fourth step, training is performed on the expert and paper matching degree model, and according to the sample data, a group of values (ω, b) needs to be found, so that Loss takes a minimum value, and the derivative of the Loss function at the extreme point is 0, so that a gradient descent method is selected: for convenience of representation, the optimization process of two parameters is taken as an example: l = L (ω 5, ω 9)
The method comprises the following steps:
randomly selecting a set of initial values, for example: [ omega 5, omega 9] = [ -100.0,1-00.0]
Selecting the next point
[ ω 5', ω 9' ], such that: l (ω 5', ω 9') < L (ω 5, ω 9)
Repeating the above steps until the loss function hardly decreases
Among them, the selection of the next point, [ omega 5', omega 9' ] is crucial
(1) Ensuring that L is decreasing;
(2) The tendency to decline is as fast as possible.
Preferably, in the fourth step, in the training expert and paper matching degree model, the gradient calculation code is as follows:
gradient_w0＝(z1-y1)*x1[0]
print(‘gradient_w0{}'.format(gradient_w0))。
preferably, in the fifth step, the matching degree between the experts screened from the model and the paper needs to be more than 85%, and the expert with the highest matching degree reviews the paper.
Preferably, in the sixth step, when the data of the big data computing system is supplemented, the sources of the data include expert information filled by an expert himself or expert information searched through a network.
The invention intelligently analyzes the matching degree of materials and experts through a big data computer system, matches the experts through three levels of a first-level subject, a second-level subject and a research direction, statistically distributes big data through accumulation of the big data, carries out academic portrayal on the experts, statistically analyzes the most adept research direction, the more adept research direction and the relatively adept research direction of the experts through normal distribution, only needs to match the first-level subject, the second-level subject, the research direction and the like of a paper with the experts after accurately analyzing the information of the experts through the big data, finds the expert with the highest matching degree, reviews the paper through the expert, and then carries out step-by-step matching according to the first-level subject, the second-level subject and the research direction through a matching algorithm of the big data, thereby improving the matching efficiency and the matching accuracy.
Examples
A big data technology-based method for matching review experts with review materials comprises the following steps:
the method comprises the following steps: establishing a big data computing system of a thesis review expert;
step two: collecting and classifying expert information collected in a big data computing system, collecting information of a first-level subject, a second-level subject and research direction related to an expert, and forming an academic portrait of the expert;
step three: matching the first-level subject, the second-level subject and the research direction information of the evaluation expert and the evaluation material step by step, respectively assigning values, and constructing an automatic matching evaluation model of the evaluation expert and the evaluation material by using a machine learning algorithm;
step four: training the model in the third step to obtain a trained model;
step five: sending a paper to be reviewed to the expert with the matching degree of more than 85% and the highest matching degree screened from the model;
step six: and (4) storing the new domain expert information in a warehouse and inputting the information into the model, and performing big data supplement on the model so as to find the most suitable review expert for subdividing the domain.
Preferably, in the first step, the thesis review expert big data computing system mainly computes the matching degree of the thesis and the experts. The information of the thesis can be divided into three categories of primary subject, secondary subject and research direction.
Preferably, in the second step, the academic sketch of the expert is an expert accurate sketch based on tag sorting, and the academic expertise of the expert is described by accurate and effective primary subject, secondary subject and research direction tags.
Preferably, in the third step, when the research directions of the experts are classified, the research directions can be systematically classified into three cases, namely, the research direction in which the experts are most adept, the research direction in which the experts are better, and the research direction in which the experts are relatively adept.
Preferably, in the third step, the automated matching evaluation model of the review expert and the review material performs review analysis between the research direction which is the most adept by the expert and the research direction of the paper, and scores are performed according to the matching degree between the research direction which is the most adept by the expert and the research direction of the paper.
Preferably, in the fourth step, training is performed on the expert and thesis matching degree model, and according to sample data, a set of (ω, b) values needs to be found, so that Loss takes a minimum value, and the derivative of the Loss function at an extreme point is 0, so that a gradient descent method is selected: for convenience of representation, the optimization process of two parameters is taken as an example: l = L (ω 5, ω 9)
The method comprises the following steps:
randomly selecting a set of initial values, for example: [ omega 5, omega 9] = [ -100.0,1-00.0]
Selecting the next point
[ ω 5', ω 9' ], such that: l (ω 5', ω 9') < L (ω 5, ω 9)
Repeating the above steps until the loss function hardly decreases any more
Among them, the selection of the next point, [ omega 5', omega 9' ] is crucial
(1) Ensuring that L is decreasing;
(2) The tendency to fall is as fast as possible.
Preferably, in the fourth step, in the training expert and paper matching degree model, the gradient calculation codes are as follows:
gradient_w0＝(z1-y1)*x1[0]
print(‘gradient_w0{}'.format(gradient_w0))。
preferably, in the fifth step, the matching degree between the experts screened from the model and the paper needs to be more than 85%, and the expert with the highest matching degree reviews the paper.
Preferably, in the sixth step, when the data of the big data computing system is supplemented, the sources of the data include expert information filled by an expert himself or expert information searched through a network.
The invention intelligently analyzes the matching degree of materials and experts through a big data technical system by setting a big data computer system, matches experts through three levels of a first-level subject, a second-level subject and a research direction, carries out academic portrait on the experts through the statistic distribution of big data through the accumulation of big data, and finds the expert with the highest matching degree by matching the information of the first-level subject, the second-level subject and the research direction of the thesis with the experts in the most adept research direction, the relatively adept research direction and the relatively adept research direction through the normal distribution.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (10)
1. A big data technology-based method for matching review experts with review materials is characterized by comprising the following steps: the method comprises the following steps:
the method comprises the following steps: establishing a big data computing system of a thesis review expert;
step two: collecting and classifying expert information collected in a big data computing system, collecting first-level subject, second-level subject and research direction information related to an expert, and forming an academic portrait of the expert;
step three: matching the first-level subject, the second-level subject and the research direction information of the evaluation expert and the evaluation material step by step, respectively assigning values, and constructing an automatic matching evaluation model of the evaluation expert and the evaluation material by using a machine learning algorithm;
step four: training the model in the third step to obtain a trained model;
step five: sending a paper to be reviewed to the expert with the matching degree of more than 85% and the highest matching degree screened from the model;
step six: and (4) putting the new domain expert information into a warehouse and inputting the information into the model, and performing big data supplement on the model so as to find the most suitable evaluation expert for subdividing the domain.
2. The big data technology-based matching method of the review experts and the review materials according to claim 1, wherein: in the first step, the big data computing system mainly computes the matching degree of the paper and the expert.
3. The big data technology-based matching method of the review experts and the review materials according to claim 1, wherein: in the second step, the academic portrait of the expert is the accurate portrait of the expert based on the label sorting, and the academic expertise of the expert is described by accurate and effective primary subject, secondary subject and research direction labels.
4. The big data technology-based matching method of the review experts and the review materials according to claim 1, wherein: in the third step, the evaluation expert and the evaluation material are automatically matched with the evaluation model, and when the research directions of the experts are classified, the three conditions can be systematically classified into the research direction which is the most adept by the expert, the research direction which is relatively adept by the expert and the research direction which is relatively adept by the expert.
5. The big data technology-based matching method of review experts and review materials according to claim 1, wherein: in the third step, the samples of expert information, thesis information and the like are extracted by feature vectors in the model of matching degree of the experts and the thesis.
6. The big data technology-based matching method of the review experts and the review materials according to claim 1, wherein: in the third step: and evaluating and analyzing the research direction which is the most adept of the expert and the research direction of the thesis in the matching degree model of the expert and the thesis, and scoring according to the matching degree of the research direction which is the most adept of the expert and the research direction of the thesis.
7. The big data technology-based matching method of the review experts and the review materials according to claim 1, wherein: in the fourth step, training is performed on the expert and thesis matching degree model, and according to sample data, a group of (ω, b) values needs to be found, so that Loss takes a minimum value, and the derivative of the Loss function at an extreme point is 0, so that a gradient descent method is selected: for convenience of representation, the optimization process of two parameters is taken as an example: l = L (ω 5, ω 9)
The method comprises the following steps:
randomly selecting a set of initial values, for example: [ omega 5, omega 9] = [ -100.0,1-00.0]
Selecting the next point
[ ω 5', ω 9' ], such that: l (ω 5', ω 9') < L (ω 5, ω 9)
Repeating the above steps until the loss function hardly decreases any more
Among them, the selection of the next point, [ omega 5', omega 9' ] is crucial
(1) Ensuring that L is decreasing;
(2) The tendency to decline is as fast as possible.
8. The big data technology-based matching method of the review experts and the review materials according to claim 1, wherein: in the fifth step, in the training expert and paper matching degree model, gradient calculation codes are as follows:
gradient_w0＝(z1-y1)*x1[0]
print(‘gradient_w0{}'.format(gradient_w0))。
9. the big data technology-based matching method of review experts and review materials according to claim 1, wherein: in the seventh step, the matching degree of the experts screened from the model and the paper needs to be more than 85%, and the expert with the highest matching degree reviews the paper.
10. The big data technology-based matching method of the review experts and the review materials according to claim 1, wherein: in the step eight, the scoring condition of the expert with the matching degree of more than 85% needs to be more than 90 points, namely the matching degree of the paper and the expert is considered to be too close, otherwise, the step is repeated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211673975.1A CN115936389A (en) | 2022-12-26 | 2022-12-26 | Big data technology-based method for matching evaluation experts with evaluation materials |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211673975.1A CN115936389A (en) | 2022-12-26 | 2022-12-26 | Big data technology-based method for matching evaluation experts with evaluation materials |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115936389A true CN115936389A (en) | 2023-04-07 |
Family
ID=86648880
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211673975.1A Pending CN115936389A (en) | 2022-12-26 | 2022-12-26 | Big data technology-based method for matching evaluation experts with evaluation materials |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115936389A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116842137A (en) * | 2023-07-18 | 2023-10-03 | 北京智信佳科技有限公司 | Method for submitting review opinion difference by audit expert |
-
2022
- 2022-12-26 CN CN202211673975.1A patent/CN115936389A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116842137A (en) * | 2023-07-18 | 2023-10-03 | 北京智信佳科技有限公司 | Method for submitting review opinion difference by audit expert |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107766929B (en) | Model analysis method and device | |
| CN108845988B (en) | Entity identification method, device, equipment and computer readable storage medium | |
| CN104966105A (en) | Robust machine error retrieving method and system | |
| CN109299271A (en) | Training sample generation, text data, public sentiment event category method and relevant device | |
| CN111614491A (en) | Power monitoring system oriented safety situation assessment index selection method and system | |
| CN108470022A (en) | A kind of intelligent work order quality detecting method based on operation management | |
| CN110263979A (en) | Method and device based on intensified learning model prediction sample label | |
| CN113434688B (en) | Data processing method and device for public opinion classification model training | |
| CN108304890A (en) | A kind of generation method and device of disaggregated model | |
| CN105469219A (en) | Method for processing power load data based on decision tree | |
| CN111415131A (en) | Big data talent resume analysis method based on natural language processing technology | |
| CN113407644A (en) | Enterprise industry secondary industry multi-label classifier based on deep learning algorithm | |
| CN115544348A (en) | Intelligent mass information searching system based on Internet big data | |
| CN110705283A (en) | Deep learning method and system based on matching of text laws and regulations and judicial interpretations | |
| CN115936389A (en) | Big data technology-based method for matching evaluation experts with evaluation materials | |
| CN111008215B (en) | Expert recommendation method combining label construction and community relation avoidance | |
| CN110175657A (en) | A kind of image multi-tag labeling method, device, equipment and readable storage medium storing program for executing | |
| CN115269958A (en) | Internet reliability data information acquisition and analysis system | |
| CN112785156B (en) | Industrial collar and sleeve identification method based on clustering and comprehensive evaluation | |
| CN116150455B (en) | Heterogeneous data analysis method | |
| CN109599096A (en) | A kind of data screening method and device | |
| CN114912460A (en) | Method and equipment for identifying transformer fault through refined fitting based on text mining | |
| CN110414819B (en) | Work order scoring method | |
| CN113191569A (en) | Enterprise management method and system based on big data | |
| CN112182211A (en) | Text classification method and device |
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 |