CN112287109A - Information matching method and device, electronic equipment and computer readable storage medium - Google Patents
Information matching method and device, electronic equipment and computer readable storage medium Download PDFInfo
- Publication number
- CN112287109A CN112287109A CN202011181705.XA CN202011181705A CN112287109A CN 112287109 A CN112287109 A CN 112287109A CN 202011181705 A CN202011181705 A CN 202011181705A CN 112287109 A CN112287109 A CN 112287109A
- Authority
- CN
- China
- Prior art keywords
- information
- candidate
- vector
- experience
- representation
- 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 39
- 239000013598 vector Substances 0.000 claims abstract description 140
- 230000011218 segmentation Effects 0.000 claims description 6
- 238000006243 chemical reaction Methods 0.000 claims description 5
- 230000001131 transforming effect Effects 0.000 claims 1
- 230000010365 information processing Effects 0.000 abstract description 2
- 230000000875 corresponding effect Effects 0.000 description 48
- 238000011161 development Methods 0.000 description 17
- 230000018109 developmental process Effects 0.000 description 17
- 238000010586 diagram Methods 0.000 description 8
- 230000006870 function Effects 0.000 description 8
- 238000005516 engineering process Methods 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000012423 maintenance Methods 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 238000004590 computer program Methods 0.000 description 2
- 238000005034 decoration Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 230000003203 everyday effect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 235000019800 disodium phosphate Nutrition 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000037361 pathway Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 230000033772 system development Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
- G06F16/353—Clustering; Classification into predefined classes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/10—Office automation; Time management
- G06Q10/105—Human resources
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Business, Economics & Management (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Entrepreneurship & Innovation (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Economics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Databases & Information Systems (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the application provides an information matching method and device, electronic equipment and a computer readable storage medium, and relates to the field of information processing. The method comprises the following steps: acquiring at least one experience information in the candidate information, wherein the experience information comprises work experience or education experience information; converting at least one piece of experience information to obtain at least one corresponding candidate vector; determining the attention weight of each experience information, and acquiring candidate representation according to the candidate vector of each experience information and the corresponding attention weight; and matching the candidate representation with the position representation of the position to be matched to obtain the matching degree between the candidate information and the position to be matched. The information matching method provided by the application can improve the matching accuracy between the candidate representation and the position representation of the position to be matched, so that the recommended result is more accurate.
Description
Technical Field
The present application relates to the field of information processing technologies, and in particular, to an information matching method, an information matching apparatus, an electronic device, and a computer-readable storage medium.
Background
Enterprises receive massive job application applications every day, and each application usually contains resumes of job seekers, namely candidate information. And after the application information reaches the enterprise, the enterprise is required to automatically correspond to the application for examination and verification so as to determine whether to submit the application information to a user department for interview evaluation.
At present, when whether the candidate information is matched with the job position of job hunting is determined according to the candidate information, each section of education experience and each section of work experience of the candidate are treated indiscriminately, so that a certain more important section of experience is possibly covered by other information, the matching accuracy between the candidate information and the job position to be matched is not high, and the final recommendation result is not accurate.
Disclosure of Invention
The aim of the present application is to solve at least one of the above mentioned technical drawbacks, in particular the technical drawback of low matching accuracy.
In a first aspect, an information matching method is provided, where the method includes:
acquiring at least one experience information in the candidate information, wherein the experience information comprises work experience or education experience information;
converting at least one piece of experience information to obtain at least one corresponding candidate vector;
determining the attention weight of each experience information, and acquiring candidate representation according to the candidate vector of each experience information and the corresponding attention weight;
and matching the candidate representation with the position representation of the position to be matched to obtain the matching degree between the candidate information and the position to be matched.
In one possible implementation manner, the obtaining at least one piece of experience information in the candidate information includes:
acquiring candidate information, and performing word segmentation on the candidate information to obtain a plurality of words;
extracting keywords from the plurality of words, and dividing the keywords based on the types corresponding to the keywords to form at least one experience message.
In one possible implementation, converting the at least one experience information into a corresponding at least one candidate vector includes:
and inputting at least one experience information into a preset model structure to obtain at least one candidate person vector.
In one possible implementation, determining the attention weight of each experience information includes:
acquiring a position vector to be matched, and splicing at least one candidate vector with the position vector to obtain a spliced vector;
and determining attention weights respectively corresponding to each experience information based on the splicing vectors.
In one possible implementation manner, determining the attention weight corresponding to each experience information respectively based on the stitching vector includes:
multiplying the spliced vector by a preset attention moment array to obtain an attention vector;
and normalizing the attention vector to obtain the attention weight corresponding to each experience information.
In one possible implementation, the attention weight is used to represent the degree of attention experienced by each segment; obtaining a candidate representation according to the candidate vector of each experience information and the corresponding attention weight, comprising:
and acquiring the weighted sum of the candidate vectors of each experience information based on the attention weight of each experience information to obtain a candidate representation.
In one possible implementation manner, matching the candidate representation with the position representation of the position to be matched to obtain a matching degree between the candidate information and the position to be matched includes:
and calculating the cosine similarity between the candidate representation and the position representation, and taking the obtained cosine similarity as the matching degree between the candidate information and the position to be matched.
In a second aspect, an information matching apparatus is provided, the apparatus comprising:
the acquisition module is used for acquiring at least one experience message in the candidate information; the experience information includes work experience or educational experience information;
the conversion module is used for converting the at least one piece of experience information to obtain at least one corresponding candidate person vector;
the determining module is used for determining the attention weight of each experience information and acquiring candidate representation according to the candidate vector of each experience information and the corresponding attention weight;
and the matching module is used for matching the candidate representation with the position representation of the position to be matched to obtain the matching degree between the candidate information and the position to be matched.
In a possible implementation manner, the obtaining module, when obtaining at least one experience information in the candidate information, is specifically configured to:
acquiring candidate information, and performing word segmentation on the candidate information to obtain a plurality of words;
extracting keywords from the plurality of words, and dividing the keywords based on the types corresponding to the keywords to form at least one experience message.
In a possible implementation manner, when the transformation module transforms the at least one piece of experience information to obtain the corresponding at least one candidate vector, the transformation module is specifically configured to:
and transmitting the grouped data with the formats of all types into a preset model structure to obtain at least one candidate vector.
In one possible implementation, the determining module, when determining the attention weight of each experience information, is specifically configured to:
acquiring a position vector to be matched, and splicing at least one candidate vector with the position vector to obtain a spliced vector;
and determining attention weights respectively corresponding to each experience information based on the splicing vectors.
In one possible implementation manner, when determining the attention weight corresponding to each piece of experience information based on the stitching vector, the determining module is specifically configured to:
multiplying the spliced vector by a preset attention moment array to obtain an attention vector;
and normalizing the attention vector to obtain the attention weight corresponding to each experience information.
In one possible implementation manner, the obtaining module obtains the candidate representation according to the candidate vector of each experience information and the corresponding attention weight when the attention weight represents the attention degree experienced by each segment, and is specifically configured to:
and acquiring the weighted sum of the candidate vectors of each experience information based on the attention weight of each experience information to obtain a candidate representation.
In a possible implementation manner, the matching module is specifically configured to, when matching the candidate representation with the position representation of the position to be matched to obtain a matching degree between the candidate information and the position to be matched:
and calculating the cosine similarity between the candidate representation and the position representation, and taking the obtained cosine similarity as the matching degree between the candidate information and the position to be matched.
In a third aspect, an electronic device is provided, which includes:
the processor, the memory and the communication interface complete mutual communication through the communication bus;
the memory is used for storing at least one executable instruction, and the executable instruction enables the processor to execute the operation corresponding to the information matching method as shown in the first aspect of the application.
In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored, which when executed by a processor implements the information matching method shown in the first aspect of the present application.
The beneficial effect that technical scheme that this application provided brought is:
by acquiring at least one experience information in the candidate information and distinguishing the at least one experience information in the candidate information by using the attention weight, more effective candidate representation can be obtained, so that the matching accuracy between the candidate representation and the position representation of the position to be matched is improved, and the recommended result is more accurate.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings used in the description of the embodiments of the present application will be briefly described below.
Fig. 1 is a schematic flowchart of an information matching method according to an embodiment of the present disclosure;
fig. 2 is a schematic diagram of an example of a candidate vector undergoing information transformation in an information matching method according to an embodiment of the present application;
fig. 3 is a schematic diagram illustrating example job information to be matched according to an information matching method provided in an embodiment of the present application;
fig. 4 is a schematic diagram of an example of a stitching vector of an information matching method provided in an embodiment of the present application;
fig. 5 is a schematic diagram of an attention weight calculation method in an information matching method according to an embodiment of the present application;
fig. 6 is a schematic diagram of an example of a comparison result between candidate information and job information to be matched in an information matching method according to an embodiment of the present application;
fig. 7 is a schematic diagram of an example of a final result of an information matching method provided in an embodiment of the present application;
fig. 8 is a schematic structural diagram of an information matching apparatus according to an embodiment of the present disclosure;
fig. 9 is a schematic structural diagram of an information matching electronic device according to an embodiment of the present application.
Detailed Description
Reference will now be made in detail to the embodiments of the present application, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are exemplary only for the purpose of explaining the present application and are not to be construed as limiting the present invention.
As used herein, the singular forms "a", "an", "the" and "the" include plural referents unless the context clearly dictates otherwise. It will be further understood that the terms "comprises" and/or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof. It will be understood that when an element is referred to as being "connected" or "coupled" to another element, it can be directly connected or coupled to the other element or intervening elements may also be present. Further, "connected" or "coupled" as used herein may include wirelessly connected or wirelessly coupled. As used herein, the term "and/or" includes all or any element and all combinations of one or more of the associated listed items.
To make the objects, technical solutions and advantages of the present application more clear, embodiments of the present application will be described in further detail below with reference to the accompanying drawings.
Enterprises receive massive job application applications every day, and each application usually contains resumes of job seekers, namely candidate information. And after the application information reaches the enterprise, the enterprise is required to automatically correspond to the application for examination and verification so as to determine whether to submit the application information to a user department for interview evaluation.
At present, when whether the candidate information is matched with the job position of job hunting is determined according to the candidate information, each section of education experience and each section of work experience of the candidate are treated indiscriminately, so that a certain more important section of experience is possibly covered by other information, the matching accuracy between the candidate information and the job position to be matched is not high, and the final recommendation result is not accurate.
The application provides an information matching method, an information matching device, an electronic device and a computer-readable storage medium, which aim to solve the above technical problems in the prior art.
The following describes the technical solutions of the present application and how to solve the above technical problems with specific embodiments. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. Embodiments of the present application will be described below with reference to the accompanying drawings.
The embodiment of the application provides an information matching method, and as shown in fig. 1, the method includes:
step S101, at least one piece of experience information in the candidate information is obtained.
The candidate information may be resume information of the candidate, and may include height, weight, gender, hobbies, and the like of the candidate.
The experience information comprises work experience information or education experience information, and each experience information can be set with a corresponding time period for distinguishing.
Specifically, the candidate information may be obtained from a designated database, or web site. A candidate may contain many types of information, but some information is highly relevant to the position and some information is not highly relevant to the position.
Step S102, converting the at least one piece of experience information to obtain at least one corresponding candidate vector.
Wherein the story information includes work story information and educational story information. Of course, when some other information in the candidate information needs to be referred to, for example, age information and height information, etc., may be added according to the requirements of the job.
Specifically, the experience information may be selected according to a time sequence, for example, may be selected from near to far, and is converted through the network structure, and the experience information is input into the network structure, that is, a vector corresponding to each segment of experience information, that is, a corresponding at least one candidate vector may be output.
Step S103, determining the attention weight of each experience information, and acquiring candidate representation according to the candidate vector of each experience information and the corresponding attention weight.
The attention weight of each experience information plays a role in distinguishing and marking the importance degree of each piece of experience information.
Specifically, the attention weight of each experience information is positively correlated with the importance degree of each experience information, and after the attention weight is obtained through calculation, the attention weight of each experience information is combined with the candidate vector to obtain the candidate representation.
For example, the corresponding candidate vectors may be fused based on the attention weights to obtain candidate representations, and a process of specifically obtaining the candidate representations will be described in detail below.
In one example, the step of obtaining the attention weight may include:
(1) obtaining a candidate vector;
(2) acquiring corresponding position information to be matched from the positions to be matched, similarly, inputting the position information into a grid structure, namely outputting a vector corresponding to each position information, and sequencing and numbering the output vectors to obtain at least one position vector;
(3) transversely splicing the candidate person vector and the position vector to obtain a complete spliced vector, wherein the spliced vector can be sequentially arranged with candidate person vectors and position vectors formed by experience information and position information from left to right;
(4) multiplying the spliced vector by a preset attention moment array to obtain an attention vector; wherein, the preset attention matrix can adopt a 64-dimensional flat vector mode;
(5) and normalizing the attention vector to obtain the attention weight.
And step S104, matching the candidate representation with the position representation of the position to be matched to obtain the matching degree between the candidate information and the position to be matched.
And matching the candidate representation and the position representation, and judging whether each candidate is suitable for the corresponding position according to the matching result.
Specifically, the cosine similarity of the vector of the candidate representation and the position representation is calculated, the obtained result is used as the matching degree of the candidate representation and the position representation, whether the candidate representation and the position representation are matched or not is judged according to the result, and if the matching degree meets the requirement, the candidate can be identified as 'qualified'; if the match does not meet the requirement, the candidate may be identified as "unqualified".
In this embodiment, by acquiring at least one experience information in the candidate information and distinguishing the at least one experience information in the candidate information by using the attention weight, a more effective candidate representation can be obtained, so that the matching accuracy between the candidate representation and the position representation of the position to be matched is improved, and the recommendation result is more accurate.
A possible implementation manner is provided in the embodiment of the present application, and the acquiring of at least one piece of experience information in the candidate information in step S101 may include:
(1) acquiring candidate information, and performing word segmentation on the candidate information to obtain a plurality of words;
(2) a plurality of keywords are extracted from the plurality of words, and at least one experience information is generated based on the plurality of keywords.
Specifically, generating at least one experience information based on the plurality of keywords may include:
a. determining the type and the time period corresponding to each keyword respectively;
b. based on the determined type and time period, the plurality of keywords are divided, thereby obtaining at least one experience information.
A possible implementation manner is provided in the embodiment of the present application, where the converting of the at least one piece of experience information to obtain the corresponding at least one candidate vector in step S102 may include:
and inputting at least one experience information into a preset model structure to obtain at least one candidate person vector.
Specifically, the preset model structure may be a neural network, for example, a word2vec model, where the word2vec model is a model for generating word vectors and may convert texts into corresponding vectors.
Specifically, at least one piece of experience information may be sequentially input into a preset model in a time sequence from near to far, the model may perform vectorization processing on the information, and at least one candidate human vector represented by each piece of experience information may be output.
As shown in fig. 2, in one example, the process of obtaining candidate vectors may be as follows: keywords summarized aiming at the work experience 1 in the aspects of 'e-commerce', 'background system', 'high concurrency' can be converted into a corresponding candidate vector, namely a vector 2 shown in the figure; the keywords summarized for the work experience 2 in "forum", "maintenance", "development" can be converted into a corresponding candidate vector, i.e. vector 4 shown in the figure.
A possible implementation manner is provided in the embodiment of the present application, and the attention weight of each experience information determined in step S103 may include:
(1) and acquiring a position vector to be matched, and splicing at least one candidate vector with the position vector to obtain a spliced vector.
Specifically, the candidate vectors and the position vectors may be spliced by using an unsupervised embedded layer model to obtain a spliced vector.
As shown in fig. 3, in one example, the position vector may be obtained by converting the position name and position requirement, for example, the position title of "JAVA development engineer" may be converted into a corresponding position vector, i.e. vector 1 shown in the figure; the job requirements "computer", "JAVA" and "highly concurrent" can be translated into a corresponding job vector, vector 2 shown in the figure.
As shown in fig. 4, the candidate vector in fig. 2 and the position vector in fig. 3 may be concatenated to obtain a concatenated vector.
(2) And determining attention weights respectively corresponding to each experience information based on the splicing vectors.
Specifically, determining the attention weight corresponding to each experience information based on the stitching vector may include:
a. multiplying the spliced vector by a preset attention moment array to obtain an attention vector;
the preset attention matrix can adopt 64-dimensional flat vectors, so that elements passing through the model can be transversely arranged, and the flat extension of the vectors is still kept in the subsequent calculation;
b. and normalizing the attention vector to obtain the attention weight corresponding to each experience information.
Specifically, the obtained attention vector may be normalized and mapped to a probability space through a Softmax function, so as to obtain an attention weight corresponding to each experience information.
The Softmax function, also called normalized exponential function, is a logic function that "compresses" a multidimensional vector containing any real number into another multidimensional vector, so that each element ranges between 0 and 1, and the sum of all elements is 1.
As shown in fig. 5, the attention weight is obtained by multiplying the stitching vector by the attention moment matrix and then normalizing the result by the Softmax function.
The embodiment of the application provides a possible implementation manner, and the attention weight is used for representing the attention degree experienced by each section; obtaining a candidate representation according to the candidate vector of each experience information and the corresponding attention weight, comprising:
and acquiring the weighted sum of the candidate vectors of each experience information based on the attention weight of each experience information to obtain a candidate representation.
Wherein, the attention weight reflects the attention degree of each experience for representing at least one of the attention degree, the importance degree or the referential of each experience information.
As shown in fig. 6, the expression form of the attention weight in the figure is represented by a connection line of the group formed by each participle in the candidate representation and the position representation, the thickness of the connection line reflects the size of the weight, the thicker the line, the larger the weight, the thinner the line, the smaller the weight, the weight of the comparison result of the candidate information and the position information shown in fig. 5 is, the weight of the "JAVA development" is larger, and the weight of the "PHP development" is smaller.
The embodiment of the present application provides a possible implementation manner, which matches a candidate representation with a position representation of a position to be matched, obtains a matching degree between candidate information and the position to be matched, and may include:
and calculating the cosine similarity between the candidate representation and the position representation, and taking the obtained cosine similarity as the matching degree between the candidate information and the position to be matched.
The cosine similarity is to evaluate the similarity of two vectors by calculating the cosine value of the included angle of the two vectors. Drawing the vector into a vector space according to the coordinate value by the cosine similarity; the matching degree is calculated by adopting a cosine similarity calculation method, the calculated amount is small, and the matching efficiency is high.
In order to more clearly explain the information matching method of the present application, the following description will be further made with reference to specific examples.
In one example, the information matching method of the present application may include:
(1) acquiring candidate information and acquiring position information of positions to be matched;
the job information description content to be matched may be as follows:
job title: JAVA development engineer
Job requirements are as follows:
1. the computer-related professions or masters, 1 to 5 years of work experience, are familiar with JAVA technology stacks and spring technology stacks;
2. skilled use of redis, mysql, various message queues, large data platforms Hbase, Hive, and spark use experience preference;
3. familiarity with distributed, high-concurrency, high-load, high-consistency, high-availability and other system development;
4. clear thought, good at communication and thinking, independent analysis and problem solving, and independent carrying of complete project
The candidate information (which may be a candidate resume) content may be as follows:
name: wangli food
Sex: for male
Age: age 25
Intention to seek employment: back end development
Educational history information:
2012-09 to 2011-06 Beijing university of aerospace software engineering Master
Mathematics this department of university of Tianjin Zhi Koch 2008-09-2012-06 Tianjin
The work experience information:
PHP development of 2011-07 to 2014-06 Beijing century interconnection technology Co Ltd
summary: is responsible for the maintenance and development of forums
JAVA development of 2014-07 to 2020-05 Beijing east science and technology Limited
summary: development supporting high-concurrency power generation business background system
The job information to be matched in the above example is JAVA development, and specifies computer-related specialties. The candidate information in the example includes two pieces of education story information and two pieces of work story information. The experiences of 'software engineering Master of Beijing aerospace university', 'JAVA development of Beijing east science and technology Limited company' and 'development supporting background system of high-concurrence power generator' of obvious candidates are highly related to the posts.
(2) Extracting candidate vectors may include:
specifically, the word segmentation is performed on the candidate resume, and the following format can be extracted:
"name":"",
"age":25,
forward: "back-end development",
"education":
{ "name": Beijing university of aerospace "," degree ": researcher", "specialty": software engineering "},
{ "name": Tianjin university of science, degree ": this family, specialty": math "},
"experience":
{ "name": Beijing century Internet technology Co., Ltd., "title": PHP development "," time _ start ": 2011-07", "time _ end": 2014-06"," summary ": forum", "maintenance", "development" },
{ "name": Beijing east science and technology Co., Ltd., "title": JAVA development "," time _ start ": 2014-07", "time _ end": 2020-05"," summary ": E-commerce", "background system", "high concurrency" }
Secondly, selecting 3 sections of work experience information from the at least one experience information according to the sequence of time from near to far, sequentially inputting the 3 sections of work experience information into a preset model structure shown in fig. 2, selecting 2 sections of education experience information according to the sequence of time from near to far, and outputting at least one candidate vector according to the same operation;
(3) converting at least one piece of position information to be matched into a position vector;
(4) splicing at least one candidate vector with the position vector to obtain a spliced vector;
(5) and (3) obtaining an attention vector by multiplying the spliced vector by a preset 64-dimensional flat vector by adopting an attention weight calculation method, and carrying out normalized mapping on the obtained attention vector to a probability space through a Softmax function to obtain the attention weight corresponding to each experience information. Combining the attention weight of the experience information with the candidate vector to obtain a candidate representation;
(6) placing each feature represented by the candidate on the left side of the double tower model shown in fig. 7, and placing each feature represented by the position on the right side of the double tower to form a double tower structure;
wherein the attention layer is a representation of a process of undergoing information conversion into attention weights;
the embedding layer (embedding layer) refers to an embedding layer in which candidate information and position information are subjected to vectorization processing;
the job emb is a representation form of the job information after vectorization processing.
(7) And calculating cosine similarity between the candidate representation and the position representation at the two sides as matching degree, identifying the matching degree of the candidate, and judging whether the matching degree meets the standard according to the identified matching degree so as to screen whether the candidate is qualified.
According to the information matching method, the at least one experience information in the candidate information is obtained, and the at least one experience information in the candidate information is distinguished by the attention weight, so that more effective candidate representation can be obtained, the matching accuracy between the candidate representation and the position representation of the position to be matched is improved, and the recommendation result is more accurate.
In the embodiment of the present application, an information matching apparatus is provided, and as shown in fig. 8, the information matching apparatus 80 may include: an acquisition module 801, a conversion module 802, a determination module 803, and a matching module 804, wherein,
an obtaining module 801, configured to obtain at least one piece of experience information in the candidate information; the experience information includes work experience information or educational experience information;
a conversion module 802, configured to convert at least one piece of experience information to obtain at least one corresponding candidate vector;
a determining module 803, configured to determine an attention weight of each experience information according to the candidate vector of each experience information and the corresponding attention weight;
the matching module 804 is configured to match the candidate representation with the position representation of the position to be matched, so as to obtain a matching degree between the candidate information and the position to be matched.
In the embodiment of the present application, a possible implementation manner is provided, and when acquiring at least one piece of experience information in candidate information, the acquiring module 801 is specifically configured to:
acquiring candidate information, and performing word segmentation on the candidate information to obtain a plurality of words;
extracting keywords from the plurality of words, and dividing the keywords based on the types corresponding to the keywords to form at least one experience message.
In the embodiment of the present application, a possible implementation manner is provided, and when converting at least one piece of experience information to obtain at least one corresponding candidate vector, the converting module 802 is specifically configured to:
and transmitting the grouped data with the formats of all types into a preset model structure to obtain at least one candidate vector.
In the embodiment of the present application, a possible implementation manner is provided, and when determining the attention weight of each experience information, the determining module 803 is specifically configured to:
acquiring a position vector to be matched, and splicing at least one candidate vector with the position vector to obtain a spliced vector;
and determining attention weights respectively corresponding to each experience information based on the splicing vectors.
In the embodiment of the present application, a possible implementation manner is provided, and when determining the attention weight corresponding to each experience information based on the concatenation vector, the determining module 803 is specifically configured to:
multiplying the spliced vector by a preset attention moment array to obtain an attention vector;
and normalizing the attention vector to obtain the attention weight corresponding to each experience information.
In the embodiment of the present application, a possible implementation manner is provided, and when the attention weight reflects the attention degree of each experience, and the obtaining module 801 obtains a candidate representation according to a candidate vector of each experience information and a corresponding attention weight, specifically configured to:
and connecting the attention weight with the candidate vector through a full connection layer to obtain a candidate representation.
The embodiment of the present application provides a possible implementation manner, and the matching module 804 is specifically configured to, when matching the candidate representation with the position representation of the position to be matched to obtain the matching degree between the candidate information and the position to be matched:
and calculating the cosine similarity between the candidate representation and the position representation, and taking the obtained cosine similarity as the matching degree between the candidate information and the position to be matched.
The information matching apparatus of this embodiment can execute the information matching method shown in the above embodiments of this application, and the implementation principles thereof are similar, and are not described herein again.
An embodiment of the present application provides an electronic device, including: a memory and a processor; at least one program stored in the memory for execution by the processor may enable an improved accuracy of the matching between the candidate representation and the position representation of the position to be matched compared to the prior art, thereby making the result of the recommendation more accurate.
In an alternative embodiment, an electronic device is provided, as shown in fig. 9, an electronic device 9000 shown in fig. 9 comprising: a processor 9001 and a memory 9003. Among other things, the processor 9001 and memory 9003 are coupled, such as via a bus 9002. Optionally, the electronic device 9000 can also include a transceiver 9004. Note that the transceiver 9004 is not limited to one in practical use, and the structure of the electronic device 9000 is not limited to the embodiment of the present application.
The Processor 9001 may be a CPU (Central Processing Unit), a general purpose Processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other Programmable logic device, transistor logic, hardware component, or any combination thereof. Which may implement or perform the various illustrative logical blocks, modules, and circuits described in connection with the disclosure. The processor 9001 may also be a combination of computing functions, e.g., comprising one or more microprocessors, a combination of DSPs and microprocessors, or the like.
The bus 9002 may include a pathway to transfer information between the aforementioned components. The bus 9002 may be a PCI (Peripheral Component Interconnect) bus, an EISA (Extended Industry Standard Architecture) bus, or the like. The bus 9002 may be divided into an address bus, a data bus, a control bus, and the like. For ease of illustration, only one thick line is shown in FIG. 9, but this does not indicate only one bus or one type of bus.
The Memory 9003 may be a ROM (Read Only Memory) or other types of static storage devices that may store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that may store information and instructions, an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical Disc storage, optical Disc storage (including Compact Disc, laser Disc, optical Disc, digital versatile Disc, blu-ray Disc, etc.), a magnetic disk storage medium or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited to these.
The memory 9003 is used to store application code for performing aspects of the present application and is controlled by the processor 9001 for execution. The processor 9001 is configured to execute application program code stored in the memory 9003 to implement what is shown in the foregoing method embodiments.
Electronic devices include, but are not limited to, terminals such as desktop computers, notebook computers, PADs (tablets), etc. The present application provides a computer-readable storage medium, on which a computer program is stored, which, when running on a computer, enables the computer to execute the corresponding content in the foregoing method embodiments. Compared with the prior art, the accuracy of matching between the candidate representation and the position representation of the position to be matched is improved, so that the recommended result is more accurate.
It should be understood that, although the steps in the flowcharts of the figures are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and may be performed in other orders unless explicitly stated herein. Moreover, at least a portion of the steps in the flow chart of the figure may include multiple sub-steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed alternately or alternately with other steps or at least a portion of the sub-steps or stages of other steps.
The foregoing is only a partial embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.
Claims (10)
1. An information matching method, comprising:
acquiring at least one experience information in the candidate information, wherein the experience information comprises work experience or education experience information;
converting the at least one experience information to obtain at least one corresponding candidate vector;
determining the attention weight of each experience information, and acquiring candidate representation according to the candidate vector and the corresponding attention weight of each experience information;
and matching the candidate representation with the position representation of the position to be matched to obtain the matching degree between the candidate information and the position to be matched.
2. The information matching method according to claim 1, wherein acquiring at least one piece of history information in the candidate information includes:
acquiring candidate information, and performing word segmentation on the candidate information to obtain a plurality of words;
extracting keywords from the plurality of words, and dividing the keywords based on the types corresponding to the keywords to form the at least one experience information.
3. The information matching method of claim 1, wherein transforming the at least one experience information into a corresponding at least one candidate vector comprises:
and inputting the at least one experience information into a preset model structure to obtain at least one candidate human vector.
4. The information matching method of claim 1, wherein determining the attention weight of each experience information comprises:
acquiring a position vector to be matched, and splicing the at least one candidate vector with the position vector to obtain a spliced vector;
based on the splicing vector, attention weights corresponding to the experience information are determined.
5. The information matching method according to claim 4, wherein determining the attention weight corresponding to each experience information based on the stitching vector comprises:
multiplying the splicing vector by a preset attention moment array to obtain an attention vector;
and normalizing the attention vector to obtain the attention weight corresponding to each experience information.
6. The information matching method according to claim 1, wherein the attention weight is used to represent a degree of attention each segment undergoes; obtaining a candidate representation according to the candidate vector and the corresponding attention weight of each experience information, comprising:
and acquiring the weighted sum of the candidate vectors of each experience information based on the attention weight of each experience information to obtain a candidate representation.
7. The information matching method according to claim 1, wherein matching the candidate representation with a position representation of a position to be matched to obtain a matching degree between the candidate information and the position to be matched comprises:
and calculating the cosine similarity between the candidate representation and the position representation, and taking the obtained cosine similarity as the matching degree between the candidate information and the position to be matched.
8. An information matching apparatus, comprising:
the acquisition module is used for acquiring at least one experience message in the candidate information; the experience information comprises work experience information or educational experience information;
the conversion module is used for converting the at least one piece of experience information to obtain at least one corresponding candidate person vector;
a determining module, configured to determine an attention weight of each experience information according to the candidate vector of each experience information and the corresponding attention weight;
and the matching module is used for matching the candidate representation with the position representation of the position to be matched to obtain the matching degree between the candidate information and the position to be matched.
9. An electronic device, comprising:
one or more processors;
a memory;
one or more application programs, wherein the one or more application programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs configured to: performing the information matching method according to any of claims 1-7.
10. A computer-readable storage medium for storing computer instructions which, when executed on a computer, cause the computer to perform the information matching method of any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011181705.XA CN112287109A (en) | 2020-10-29 | 2020-10-29 | Information matching method and device, electronic equipment and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011181705.XA CN112287109A (en) | 2020-10-29 | 2020-10-29 | Information matching method and device, electronic equipment and computer readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112287109A true CN112287109A (en) | 2021-01-29 |
Family
ID=74353492
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011181705.XA Pending CN112287109A (en) | 2020-10-29 | 2020-10-29 | Information matching method and device, electronic equipment and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112287109A (en) |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120123956A1 (en) * | 2010-11-12 | 2012-05-17 | International Business Machines Corporation | Systems and methods for matching candidates with positions based on historical assignment data |
| CN106384230A (en) * | 2016-10-21 | 2017-02-08 | 北京搜前途科技有限公司 | Method of matching work experience in resume with recruitment job and method of matching resume with recruitment information |
| CN107590133A (en) * | 2017-10-24 | 2018-01-16 | 武汉理工大学 | The method and system that position vacant based on semanteme matches with job seeker resume |
| CN108920544A (en) * | 2018-06-13 | 2018-11-30 | 桂林电子科技大学 | A kind of personalized position recommended method of knowledge based map |
| CN109754233A (en) * | 2019-01-29 | 2019-05-14 | 上海嘉道信息技术有限公司 | A kind of method and system of intelligent recommendation job information |
| CN110059923A (en) * | 2019-03-13 | 2019-07-26 | 平安科技（深圳）有限公司 | Matching process, device, equipment and the storage medium of post portrait and biographic information |
| CN110059162A (en) * | 2019-04-28 | 2019-07-26 | 苏州创汇智信息技术有限公司 | A kind of matching process and device of job seeker resume and position vacant |
| CN110874714A (en) * | 2018-08-31 | 2020-03-10 | 阿里巴巴集团控股有限公司 | Data matching method and device |
| CN111078971A (en) * | 2019-11-19 | 2020-04-28 | 平安金融管理学院(中国·深圳) | Resume file screening method and device, terminal and storage medium |
| CN111078835A (en) * | 2019-12-05 | 2020-04-28 | 平安金融管理学院(中国·深圳) | Resume evaluation method and device, computer equipment and storage medium |
| CN111144723A (en) * | 2019-12-17 | 2020-05-12 | 埃摩森网络科技(上海)有限公司 | Method and system for recommending people's job matching and storage medium |
-
2020
- 2020-10-29 CN CN202011181705.XA patent/CN112287109A/en active Pending
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120123956A1 (en) * | 2010-11-12 | 2012-05-17 | International Business Machines Corporation | Systems and methods for matching candidates with positions based on historical assignment data |
| CN106384230A (en) * | 2016-10-21 | 2017-02-08 | 北京搜前途科技有限公司 | Method of matching work experience in resume with recruitment job and method of matching resume with recruitment information |
| CN107590133A (en) * | 2017-10-24 | 2018-01-16 | 武汉理工大学 | The method and system that position vacant based on semanteme matches with job seeker resume |
| CN108920544A (en) * | 2018-06-13 | 2018-11-30 | 桂林电子科技大学 | A kind of personalized position recommended method of knowledge based map |
| CN110874714A (en) * | 2018-08-31 | 2020-03-10 | 阿里巴巴集团控股有限公司 | Data matching method and device |
| CN109754233A (en) * | 2019-01-29 | 2019-05-14 | 上海嘉道信息技术有限公司 | A kind of method and system of intelligent recommendation job information |
| CN110059923A (en) * | 2019-03-13 | 2019-07-26 | 平安科技（深圳）有限公司 | Matching process, device, equipment and the storage medium of post portrait and biographic information |
| CN110059162A (en) * | 2019-04-28 | 2019-07-26 | 苏州创汇智信息技术有限公司 | A kind of matching process and device of job seeker resume and position vacant |
| CN111078971A (en) * | 2019-11-19 | 2020-04-28 | 平安金融管理学院(中国·深圳) | Resume file screening method and device, terminal and storage medium |
| CN111078835A (en) * | 2019-12-05 | 2020-04-28 | 平安金融管理学院(中国·深圳) | Resume evaluation method and device, computer equipment and storage medium |
| CN111144723A (en) * | 2019-12-17 | 2020-05-12 | 埃摩森网络科技(上海)有限公司 | Method and system for recommending people's job matching and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 刘臣等: "融合情感符号的自注意力BLSTM情感分析", 《软件导刊》, vol. 19, no. 3, pages 2 - 4 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108959246B (en) | Answer selection method and device based on improved attention mechanism and electronic equipment | |
| CN110298019A (en) | Name entity recognition method, device, equipment and computer readable storage medium | |
| CN112016295B (en) | Symptom data processing method, symptom data processing device, computer equipment and storage medium | |
| CN111382255A (en) | Method, apparatus, device and medium for question and answer processing | |
| WO2021027125A1 (en) | Sequence labeling method and apparatus, computer device and storage medium | |
| CN112883190A (en) | Text classification method and device, electronic equipment and storage medium | |
| CN114547267A (en) | Intelligent question-answering model generation method and device, computing equipment and storage medium | |
| CN112085091B (en) | Short text matching method, device, equipment and storage medium based on artificial intelligence | |
| CN113886550A (en) | Question-answer matching method, device, equipment and storage medium based on attention mechanism | |
| CN113722512A (en) | Text retrieval method, device and equipment based on language model and storage medium | |
| CN115221276A (en) | Chinese image-text retrieval model training method, device, equipment and medium based on CLIP | |
| CN116821373A (en) | Map-based prompt recommendation method, device, equipment and medium | |
| CN113722441B (en) | Similar text generation method, device, equipment and storage medium | |
| CN113342940A (en) | Text matching analysis method and device, electronic equipment and storage medium | |
| WO2018138205A1 (en) | Model search method and device based on semantic model framework | |
| CN112287109A (en) | Information matching method and device, electronic equipment and computer readable storage medium | |
| CN113656586B (en) | Emotion classification method, emotion classification device, electronic equipment and readable storage medium | |
| CN114139530A (en) | Synonym extraction method and device, electronic equipment and storage medium | |
| CN113705692A (en) | Emotion classification method and device based on artificial intelligence, electronic equipment and medium | |
| CN111859985A (en) | AI customer service model testing method, device, electronic equipment and storage medium | |
| CN113792539B (en) | Entity relationship classification method and device based on artificial intelligence, electronic equipment and medium | |
| CN116364223B (en) | Feature processing method, device, computer equipment and storage medium | |
| CN117056836B (en) | Program classification model training and program category identification method and device | |
| CN117874202A (en) | Intelligent question-answering method and system based on large model | |
| CN114116971A (en) | Model training method and device for generating similar texts and computer equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |