Resume search method, device, electronic device and computer storage medium

Abstract

The embodiments of the present application provide a resume search method, device, electronic device and computer storage medium, which relate to the field of information search. The method includes: obtaining a search statement, performing word segmentation processing on the search statement, and obtaining a word segmentation result; determining the search intent, entities in the search statement and the relationship between entities according to the word segmentation result; and determining the resume according to the search intent, entities in the search statement and the relationship between entities. The embodiments of the present application analyze the search statement input by the user to determine the search intent, entities and the relationship between entities, so as to more directly and accurately identify the user's intent, and use knowledge graph matching to obtain more complete and comprehensive data to assist users in making better decisions.

Description

Technical Field

[0001] The present application relates to the field of information search technology, and in particular to a resume search method, device, electronic device and computer storage medium. Background Art

[0002] In today's Internet age, people are looking for jobs through the Internet, and companies are also looking for the talents they need on the Internet, generally by searching from talent databases. However, as more and more data accumulates in corporate talent databases, companies face great challenges in screening out suitable ones.

[0003] The current human resources search method is mainly keyword matching search. An index is built for all candidates in the talent database according to different fields. When a user searches, he selects the corresponding index and enters the corresponding keyword to search.

[0004] This type of search method requires regular maintenance of the dictionary, which has high maintenance costs and consumes a lot of resources. In addition, the search results are relatively simple and often cannot meet the needs of enterprises. Summary of the invention

[0005] The embodiments of the present application provide a resume search method, device, electronic device, and computer storage medium that overcome the above-mentioned problems or at least partially solve the above-mentioned problems.

[0006] In a first aspect, a resume search method is provided, the method comprising:

[0007] Obtain search statements, perform word segmentation on the search statements, and obtain word segmentation results;

[0008] Determine the search intent, entities in the search sentence, and the relationships between entities based on the word segmentation results;

[0009] Search results are obtained from the talent pool and pre-built knowledge graph based on the search intent, entities in the search statement, and the relationships between entities. The talent pool is used to store resumes, and the knowledge graph is used to store entities in the resumes and the relationships between entities.

[0010] In a possible implementation, the search intent, entities, and relationships between entities are determined based on the word segmentation results, including:

[0011] Input the word segmentation results into a pre-trained intent recognition neural network model to obtain the search intent output by the intent recognition neural network model;

[0012] Input the word segmentation results into a pre-trained entity recognition neural network model to obtain entities in the search sentence output by the entity recognition neural network model;

[0013] The entities are input into a preset semantic relationship model to obtain the relationship between the entities output by the preset semantic relationship model.

[0014] In another possible implementation, search results are obtained from a talent pool and a pre-built knowledge graph according to search intent, entities in a search sentence, and relationships between entities, including:

[0015] Search the preset talent pool for resumes containing entities in the search statement and select them as the first target resumes;

[0016] Determine a first target entity in a first target resume, the first target entity and the entity satisfying the relationship between entities;

[0017] According to the search intent, an entity with the same semantics as the first target entity is searched in the preset knowledge graph as the second target entity;

[0018] The resume containing the second target entity is searched in the talent pool as the second target resume, and the second target resume is used as the search result.

[0019] In another possible implementation, searching the talent pool for a resume containing the second target entity as the second target resume, taking the second target resume as the search result, and then further including the step of sorting the search results include:

[0020] The score of the second target resume is determined based on the time difference between the last update of the second target resume and the current time, and the similarity between the search statement and the second target resume;

[0021] Sort the second target resumes in descending order of scores, and return the sorted search results;

[0022] The score is negatively correlated with the time difference between the last update of the second target resume and the current moment, and positively correlated with the similarity between the search statement and the second target resume.

[0023] In another possible implementation, searching the talent pool for a resume containing the second target entity as the second target resume, taking the second target resume as the search result, and then further including the step of expanding the search result, including:

[0024] Searching the knowledge graph for attribute information of entities contained in the second target resume;

[0025] The attribute information is added to the second target resume to obtain an expanded search result; the attribute information of an entity refers to the characteristics of the entity.

[0026] In a second aspect, a resume search device is provided, the device comprising:

[0027] The acquisition module is used to acquire the search statement, perform word segmentation on the search statement, and obtain the word segmentation result;

[0028] The analysis module is used to determine the search intent, entities in the search statement, and the relationship between entities based on the word segmentation results;

[0029] The search module is used to obtain search results from the talent pool and the pre-built knowledge graph based on the search intent, entities in the search statement, and the relationships between entities. The talent pool is used to store resumes, and the knowledge graph is used to store entities in the resumes and the relationships between entities.

[0030] Furthermore, the analysis module includes:

[0031] The intent analysis module is used to input the word segmentation results into a pre-trained intent recognition neural network model to obtain the search intent output by the intent recognition neural network model;

[0032] An entity analysis module is used to input the word segmentation results into a pre-trained entity recognition neural network model to obtain entities in the search sentence output by the entity recognition neural network model;

[0033] The relationship analysis module is used to input entities into a preset semantic relationship model to obtain the relationship between entities output by the preset semantic relationship model.

[0034] Furthermore, the search module includes:

[0035] A first resume search module searches for resumes containing entities in the search statement in a preset talent pool and uses the resumes as first target resumes;

[0036] A first entity search module determines a first target entity in a first target resume, and the first target entity and the entity satisfy an entity-to-entity relationship;

[0037] A second entity search module, used to search for an entity with the same semantics as the first target entity in a preset knowledge graph according to the search intent as a second target entity;

[0038] The second resume search module is used to search the resume containing the second target entity in the talent pool as the second target resume, and use the second target resume as the search result.

[0039] Furthermore, the second resume search module includes:

[0040] A ranking module, wherein the score of the second target resume is determined based on the time difference between the last update of the second target resume and the current time, and the similarity between the search statement and the second target resume;

[0041] Sort the second target resumes in descending order of scores, and return the sorted search results;

[0042] The score is negatively correlated with the time difference between the last update of the second target resume and the current moment, and positively correlated with the similarity between the search statement and the second target resume.

[0043] Furthermore, the second resume search module also includes:

[0044] An expansion module, used to search the knowledge graph for attribute information of entities contained in the second target resume;

[0045] Add the attribute information to the second target resume and return the expanded search results; the attribute information of an entity refers to the characteristics of the entity.

[0046] In a third aspect, an embodiment of the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the method provided in the first aspect are implemented.

[0047] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method provided in the first aspect.

[0048] The resume search method, device, electronic device and storage medium provided in the embodiments of the present application can more directly and accurately identify the user's intention by analyzing the search statement input by the user, determining the search intent, entities and the relationship between entities, and can obtain more complete and comprehensive data by using knowledge graph matching to assist users in making better decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in describing the embodiments of the present application are briefly introduced below.

[0050] Figure 1 A network architecture diagram of a resume search system provided in an embodiment of the present application;

[0051] Figure 2 A flow chart of a resume search method provided in an embodiment of the present application;

[0052] Figure 3 A schematic diagram of an intent recognition neural network model provided in an embodiment of the present application;

[0053] Figure 4 A schematic diagram of an entity recognition neural network model provided in an embodiment of the present application;

[0054] Figure 5 A flowchart for obtaining search results provided in an embodiment of the present application;

[0055] Figure 6 A schematic diagram of a resume search process provided by an embodiment of the present application;

[0056] Figure 7 A schematic diagram of the structure of a resume search device provided in an embodiment of the present application;

[0057] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0058] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application, and cannot be interpreted as limiting the present invention.

[0059] It will be understood by those skilled in the art that, unless expressly stated, the singular forms "one", "an" and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present application refers to the presence of features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof. It should be understood that when we refer to an element as being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be an intermediate element. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The term "and / or" used herein includes all or any unit and all combinations of one or more associated listed items.

[0060] The resume searching method provided in this application is intended to solve the above technical problems in the prior art.

[0061] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below in conjunction with the accompanying drawings.

[0062] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0063] First, this application can be applied in various resume search scenarios, such as when the HR (human resources) of an enterprise searches for job seekers who meet the requirements of the enterprise. A resume is a brief introduction sent by a job seeker to a recruiting unit, which contains various basic information, such as name, gender, age, ethnicity, place of origin, political status, education, contact information, etc., as well as self-evaluation, work experience, study experience, honors and achievements, job-seeking desires, understanding of the job, etc. Resumes are generally stored in a talent pool. When you need to search for resumes, you can search for the corresponding resume in the talent pool.

[0064] It should be understood that the resume search method provided in this application can be applied to any computer device or system with resume search function. Specifically, the resume search system can be run on Figure 1 In the network architecture diagram shown, Figure 1 A network architecture diagram of the resume search system provided in an embodiment of the present application, wherein the server 11 receives the search statement uploaded by the terminal 12, performs corresponding processing and search, and returns the obtained resume to the terminal 12.

[0065] In order to more clearly illustrate the technical solution of the present application, the present application will be explained in the resume search scenario in the following. The existing resume search method is generally to search directly in the talent pool. An index will be established in the existing talent pool, such as an inverted index, and keyword matching will be performed in the talent pool. The corresponding resume is searched according to the corresponding index. This search method is relatively simple, and the resumes searched are relatively simple and small in number. The maintenance cost of the talent pool is very high, and the resources consumed are high, which cannot fully meet the needs of the enterprise. In addition, when searching for resumes, the search statement entered by the user sometimes cannot fully match the corresponding index, and the resume that the user wants to search cannot be fully understood, that is, the user's intention cannot be clearly determined, and the user cannot be assisted in making decisions.

[0066] The present application embodiment provides a resume search method, such as Figure 2 As shown, Figure 2 A flow chart of a resume search method provided in an embodiment of the present application, the method comprising:

[0067] S101: Obtain a search statement, perform word segmentation on the search statement, and obtain a word segmentation result.

[0068] The search sentence is entered by the user, which can be entered on a terminal such as a computer or a mobile phone. The search sentence can be text and numbers, such as "Zhang San's graduation school"; "The mobile phone number is 12345678". Word segmentation refers to a simple segmentation of the search sentence, and the result is the word segmentation result. For example, the word segmentation of "Zhang San's graduation school" will result in "Zhang San" and "graduation school".

[0069] After obtaining the search statement, the embodiment of the present application performs word segmentation on the search statement. It should be understood that the search statement entered by the user may not be standard and meet the search requirements. The search statement entered by the user needs to be processed in a certain way before the subsequent search steps can be performed.

[0070] S102: Determine the search intent, entities in the search sentence, and the relationship between entities based on the word segmentation results.

[0071] Search intent refers to the content that the user wants to search for, which is expressed by the search sentence entered by the user in the user's mind. It can be obtained using the pre-trained intent recognition neural network model. For example, if the search sentence entered by the user is "Zhang San's alumni", the search intent is people who graduated from the same school as Zhang San. Entity refers to an objectively existing thing, for example, "Zhang San" is an entity. The relationship between entities refers to the connection between entities, for example, the relationship between the two entities "Zhang San" and "Peking University" may be graduation.

[0072] After obtaining the word segmentation results, the embodiment of the present application can determine the search intent, entities in the search statement, and the relationship between entities based on the search results. This step can be obtained through a pre-trained neural network model. For example, the search intent can be obtained using a pre-trained intent recognition neural network model; the entities in the search statement can be obtained using a pre-trained entity recognition neural network model; and the relationship between entities can be obtained using a preset semantic relationship model.

[0073] S103. Obtain search results from a talent pool and a pre-built knowledge graph based on search intent, entities in the search statement, and relationships between entities, wherein the talent pool is used to store resumes, and the knowledge graph is used to store entities in the resumes and relationships between entities.

[0074] After obtaining the search intent, entities in the search statement, and the relationships between entities, the embodiment of the present application can search in a pre-established talent pool and combine with the knowledge graph to determine the resume that the user wants to obtain. The specific search method is described in the embodiments later in this application.

[0075] The resume search method, device, electronic device and storage medium provided in the embodiments of the present application can more directly and accurately identify the user's intention by analyzing the search statement input by the user, determining the search intent, entities and the relationship between entities, and can obtain more complete and comprehensive data by using knowledge graph matching to assist users in making better decisions.

[0076] The embodiment of the present application also provides a possible implementation method, which determines the search intent, entities, and the relationship between entities according to the word segmentation results, including:

[0077] S1021. Input the word segmentation results into a pre-trained intent recognition neural network model to obtain the search intent output by the intent recognition neural network model.

[0078] The embodiment of the present application obtains the search intent by machine learning. Specifically, the word segmentation result is input into a pre-trained intent recognition neural network model to obtain the user's search intent output by the intent recognition neural network model. The SoftMax (flexible maximum) activation function is used in the output layer of the intent recognition neural network model. The SoftMax activation function can map the outputs of multiple neural networks to probabilities in the (0,1) interval, and select the result with the maximum probability as the search intent. The formula is as follows:

[0079]

[0080] Among them, f(Z i ) is the i-th sample search intent Z i The probability of is the i-th sample search intent Z i The exponent of , e is a mathematical constant, which is the base of natural logarithms, also known as Euler's number, and n is the total number of sample search intents.

[0081] It should be understood that before executing step S1021, the intent recognition neural network model can also be pre-trained, which can be obtained specifically by the following method: First, a certain number of sample search sentences are collected, and word segmentation processing is performed to obtain the word segmentation results of each sample search sentence, and then the search intent represented by the word segmentation results of each sample search sentence is determined. The search intent may include name, date of birth, mobile phone number, school, company, position, skills, etc. It can be searching for resumes by name, or searching for resumes by mobile phone number, etc. Randomly, based on the word segmentation results of the sample search sentences and the search intent represented by the word segmentation results of the sample search sentences, training is performed to obtain the intent recognition neural network model. Among them, the initial model can be a single neural network model or a combination of multiple neural network models.

[0082] Figure 3A schematic diagram of an intent recognition neural network model provided for an embodiment of the present application, wherein w0, w1, w2, w3, and w4 are words in the word segmentation results. The intent recognition neural network model of the embodiment of the present application can adopt a CNN+MLP neural network model, including a CNN (Convolutional Neural Networks) convolutional neural network and an MLP (Multilayer Perceptron) fully connected neural network. The embodiment of the present application does not make any specific limitation on the model structure of the intent recognition neural network model.

[0083] The training methods for the intent recognition neural network model include:

[0084] S201, initialize CNN layer and MLP layer;

[0085] S202, using the word segmentation results of a certain number of sample search sentences as training samples, using the search intent represented by the word segmentation results of the sample search sentences as sample labels, inputting the training samples and the sample labels into the CNN layer, and obtaining the feature vectors of the training samples output by the CNN layer;

[0086] S203, inputting the feature vector of the training sample into the MLP layer to obtain the search intent prediction result of the training sample output by the MLP layer;

[0087] S204. Calculate the deviation between the prediction result and the sample label corresponding to the training sample, and adjust the parameters of the CNN layer and the MLP layer by reverse feedback until the deviation reaches the convergence condition, thereby obtaining a trained intent recognition neural network model.

[0088] S1022. Input the word segmentation results into a pre-trained entity recognition neural network model to obtain entities in the search sentence output by the entity recognition neural network model.

[0089] The embodiment of the present application obtains entities through machine learning. Specifically, the word segmentation results are input into a pre-trained entity recognition neural network model to obtain the entities output by the entity recognition neural network model.

[0090] It should be understood that before executing step S1022, the entity recognition neural network model can also be pre-trained, which can be obtained by the following method: first, a certain number of sample search sentences are collected, and word segmentation processing is performed to obtain the word segmentation results of each sample search sentence, and then the entities contained in the word segmentation results of each sample search sentence are determined. For example, if the word segmentation results are "Zhang San" and "Peking University", the name entity Zhang San and the school entity Peking University can be obtained. Randomly, training is performed based on the word segmentation results of the sample search sentences and the entities contained in the word segmentation results of the sample search sentences to obtain the entity recognition neural network model. Among them, the initial model can be a single neural network model or a combination of multiple neural network models.

[0091] Figure 4 A schematic diagram of an entity recognition neural network model provided for an embodiment of the present application, wherein w0, w1, w2, w3, and w4 are words in the word segmentation results. The entity recognition neural network model of the embodiment of the present application can adopt a BiLSTM+CRF neural network model, including a BiLSTM (Bi-directional Long Short-Term Memory) neural network and a CRF (conditional random field) neural network. The embodiment of the present application does not make any specific limitation on the model structure of the entity recognition neural network model.

[0092] The training methods for the entity recognition neural network model include:

[0093] S301, initialize the BiLSTM layer and the CRF layer;

[0094] S302, using the word segmentation results of a certain number of sample search sentences as training samples, using the entities contained in the word segmentation results of the sample search sentences as sample labels, inputting the training samples and the sample labels into the BiLSTM layer, and obtaining the classification probability of the training samples output by the BiLSTM layer;

[0095] S303, inputting the classification probability of the training sample into the CRF layer to obtain the entity prediction result of the training sample output by the CRF layer;

[0096] S304, calculating the deviation between the prediction result and the sample label corresponding to the training sample, and adjusting the parameters of the BiLSTM layer and the CRF layer by reverse feedback until the deviation reaches the convergence condition, thereby obtaining a trained entity recognition neural network model.

[0097] S1023: Input the entity into a preset semantic relationship model to obtain the relationship between the entities output by the preset semantic relationship model.

[0098] By inputting the entities into the preset semantic relationship model, the relationship between the entities output by the preset semantic relationship model can be obtained. The open source LTP (Language Technology Platform) model is used. Various semantic relationships are defined in the LTP model. After the entities are input into the LTP model, the relationship between the entities can be determined using the defined semantic relationship. For example, the relationship between the name entity Zhang San and the school entity can be obtained as graduation. Specifically, Table 1 is a semantic role relationship definition table, as shown in Table 1:

[0099] Relationship Type Description Example ARG0 Agent [Zhang San ARG0] graduated from Peking University ARG1 Patient Zhang San graduated from [Peking University ARG1] FEAT Modifier Zhang San [graduated from FEAT] Peking University

[0100] Table 1: Semantic role relationship definition table

[0101] The embodiments of the present application utilize pre-trained intent recognition neural network models, entity recognition neural network models, and preset semantic relationship models, which can accurately identify the user's search intent expressed by the search statement, the entities contained, and the relationships between entities, and can perform more precise searches based on user needs.

[0102] The present application also provides a possible implementation method, such as Figure 5 As shown, Figure 5 A schematic diagram of a process for obtaining search results provided in an embodiment of the present application, obtaining search results from a talent pool and a pre-built knowledge graph according to search intent, entities, and relationships between entities, includes:

[0103] S1031. Search for a resume containing the entity in a preset talent pool and use it as the first target resume.

[0104] The talent pool has been established in advance, and resumes are stored in the talent pool. According to the entity obtained in the above steps, the resume containing the entity can be searched in the talent pool, and the resume searched at this time is used as the first target resume. For example: if the search statement entered by the user is "Zhang San's alumni", the entity Zhang San can be obtained through the above steps, and the first target resume, that is, the resume of the person named Zhang San, can be obtained by searching in the talent pool.

[0105] S1032. Determine a first target entity in a first target resume, where the first target entity and the entity satisfy an entity-to-entity relationship.

[0106] After determining the first target resume, the entity having the entity relationship with the entity can be determined as the first target entity based on the entity relationship obtained in the above steps. For example, the search sentence input by the user is "Zhang San's alumni", and the above-mentioned relationship between the entities can be obtained as graduation, so the first target entity can be determined as "Peking University".

[0107] S1033. Search for an entity with the same semantics as the first target entity in the preset knowledge graph according to the search intent as the second target entity.

[0108] After obtaining the first entity, you can search in the preset knowledge graph according to the user's search intent to obtain an entity with the same actual meaning as the first target entity as the second target entity. It should be understood that the knowledge graph is pre-established. The knowledge graph stores various entities, relationships between entities, and attributes of entities. The attributes of an entity refer to various characteristics of an entity, and the content written in a resume may not be all the attributes represented by the entity. The second target entity can be determined based on the attributes of the entity stored in the knowledge graph. For example, the nicknames of the school entity "Peking University" include "Yuanmingyuan Vocational and Technical College" and "Peking University".

[0109] S1034. Search the talent pool for a resume containing the second target entity as the second target resume, and use the second target resume as the search result.

[0110] After obtaining the second target entity, you can continue searching in the preset talent pool to obtain all resumes containing the second target entity as second target resumes. The second target resume is the resume that the user wants to search for. For example, search for all resumes containing "Peking University", "Peking University", and "Yuanmingyuan Vocational and Technical College" as second target resumes.

[0111] The embodiment of the present application uses the knowledge graph to perform a more extensive and accurate search based on the search intent, entities, and relationships between entities expressed by the search statement entered by the user, thereby meeting all the user's search needs.

[0112] The present application also provides a possible implementation method, such as Figure 6 As shown, Figure 6 A schematic diagram of a resume search process provided in an embodiment of the present application:

[0113] Among them, F represents the server, R represents the talent pool, and the server F searches for the resume containing the entity in the search statement in the talent pool R, obtains the resume containing the entity in the search statement in the talent pool R, and takes it as the first target resume; then the server F takes the entity in the first target resume that satisfies the entity relationship in the search statement as the first target entity.

[0114] T represents the knowledge graph. Server F searches in the knowledge graph T according to the user's search intention and obtains the entity in the knowledge graph T with the same semantics as the first target entity as the second target entity.

[0115] Z represents a terminal. The server F searches the talent pool R, obtains the resume of the second target entity in the talent pool R as a search result, and returns it to the terminal Z.

[0116] The embodiment of the present application solves the problem that accurate search is difficult due to a large number of resumes stored in the talent pool by organically combining the knowledge graph with the talent pool for search, making the search results extensive and accurate.

[0117] The embodiment of the present application also provides a possible implementation method, which searches for a resume containing a second target entity in the talent pool as the second target resume, takes the second target resume as the search result, and then further includes the step of sorting the search results, including:

[0118] The score of the second target resume is determined based on the time difference between the last update of the second target resume and the current time, and the similarity between the search statement and the second target resume;

[0119] Sort the second target resumes in descending order of scores, and return the sorted search results;

[0120] The score is negatively correlated with the time difference between the last update of the second target resume and the current moment, and positively correlated with the similarity between the search statement and the second target resume.

[0121] A large number of resumes are stored in the talent pool, and job seekers will constantly update the content of their resumes as time goes by and their experiences change. The time difference between the last update and the current search can be used as a criterion to allow users to select the resumes they need. The similarity between the user's search statement and the resume content can also be a condition closely related to the user's needs. The smaller the time difference between the last update and the current search, the greater the similarity between the user's search statement and the resume content, and the higher the score. The following algorithm formula can calculate the score of the resume under the above two conditions:

[0122] Formula (1):

[0123] score=bm25+0.5*e -0.05*t

[0124] Formula (2):

[0125]

[0126] Formula (1) is the calculation formula for the score, where bm25 is the similarity between the search statement and the resume, t is the difference between the last update time of the resume and the current search time, and e is a natural constant in mathematics, also known as the Euler number.

[0127] Formula (2) is the calculation formula for the similarity bm25 between the search statement and the resume, where r i is the number of relevant resumes containing participle i, R is the total number of relevant resumes containing participle i, n i is the total number of resumes, N is the total number of resumes, K, k1, k2 are constants, f i is the frequency of the participle i appearing in the resume, qf i is the frequency of word i appearing in the user's search sentence.

[0128] The embodiment of the present application sorts resumes in descending order by using two judgment conditions: the difference between the time when the resume was last updated and the time when the resume is currently searched, and the similarity between the search statement and the resume. This allows users to have a very intuitive perspective on the searched resumes and can better meet user needs.

[0129] The embodiment of the present application also provides a possible implementation method, which searches for a resume containing a second target entity in the talent pool as the second target resume, and uses the second target resume as the search result, and then further includes the step of expanding the search result, including:

[0130] Searching the knowledge graph for attribute information of entities contained in the second target resume;

[0131] Add attribute information to the second target resume; the attribute information of an entity refers to the characteristics of the entity.

[0132] After searching for a resume, users may need to determine whether the job seeker with the resume meets their needs based on the attribute information of one or more entities in the resume, or they can make a comparison to allow users to select the job seeker. The knowledge graph stores the attribute information of the entity, that is, the characteristics of the entity. The attribute information of the entity in the knowledge graph can be added to the resume to allow users to make a selection.

[0133] For example, if the resumes that the user wants to search for show that some people graduated from Peking University and some graduated from Tsinghua University, then the attribute information of the entity Peking University and the entity Tsinghua University stored in the knowledge graph can be added to the resumes for comparison, such as Peking University's strong professional directions, teaching resources, etc., and Tsinghua University's strong professional directions, teaching resources, etc.

[0134] The embodiment of the present application allows users to fully understand the attribute information of different entities in the resumes of different job seekers by adding the attribute information of entities in the knowledge graph to the resume, allowing users to make better decisions based on their needs and select job seekers who are more suitable for the user's company.

[0135] The present application embodiment provides a resume search device, such asFigure 7 As shown, Figure 7 A schematic diagram of a resume search device provided in an embodiment of the present application, the device may include: an acquisition module 101, an analysis module 102 and a search module 103, wherein:

[0136] The acquisition module 101 is used to acquire a search statement, perform word segmentation processing on the search statement, and obtain a word segmentation result;

[0137] An analysis module 102 is used to determine the search intent, entities in the search sentence, and the relationship between entities based on the word segmentation results;

[0138] The search module 103 is used to obtain search results from a talent pool and a pre-built knowledge graph based on the search intent, entities in the search statement, and the relationships between entities, wherein the talent pool is used to store resumes, and the knowledge graph is used to store entities in the resumes and the relationships between entities.

[0139] The resume search device provided in the embodiment of the present application specifically executes the above method embodiment process. For details, please refer to the content of the resume search method embodiment above, which will not be repeated here. The resume search device provided in the embodiment of the present application analyzes the search statement input by the user to determine the search intent, entities and the relationship between entities. It can more directly and accurately identify the user's intent, and use knowledge graph matching to obtain more complete and comprehensive data to assist users in making better decisions.

[0140] Furthermore, the analysis module 102 includes:

[0141] The intent analysis module is used to input the word segmentation results into a pre-trained intent recognition neural network model to obtain the search intent output by the intent recognition neural network model;

[0142] An entity analysis module is used to input the word segmentation results into a pre-trained entity recognition neural network model to obtain entities in the search sentence output by the entity recognition neural network model;

[0143] The relationship analysis module is used to input entities into a preset semantic relationship model to obtain the relationship between entities output by the preset semantic relationship model.

[0144] Furthermore, the search module 103 includes:

[0145] A first resume search module searches for resumes containing entities in the search statement in a preset talent pool and uses the resumes as first target resumes;

[0146] A first entity search module determines a first target entity in a first target resume, and the first target entity and the entity satisfy an entity-to-entity relationship;

[0147] A second entity search module, used to search for an entity with the same semantics as the first target entity in a preset knowledge graph according to the search intent as a second target entity;

[0148] The second resume search module is used to search the resume containing the second target entity in the talent pool as the second target resume, and use the second target resume as the search result.

[0149] Furthermore, the second resume search module includes:

[0150] A ranking module, wherein the score of the second target resume is determined based on the time difference between the last update of the second target resume and the current time, and the similarity between the search statement and the second target resume;

[0151] Sort the second target resumes in descending order of scores, and return the sorted search results;

[0152] The score is negatively correlated with the time difference between the last update of the second target resume and the current moment, and positively correlated with the similarity between the search statement and the second target resume.

[0153] Furthermore, the second resume search module also includes:

[0154] An expansion module, used to search the knowledge graph for attribute information of entities contained in the second target resume;

[0155] Add the attribute information to the second target resume and return the expanded search results; the attribute information of an entity refers to the characteristics of the entity.

[0156] An electronic device is provided in an embodiment of the present application, and the electronic device includes: a memory and a processor; at least one program, stored in the memory, for being executed by the processor, which can achieve, compared with the prior art: analyzing the search statement input by the user, determining the search intent, entities, and the relationship between entities, and can more directly and accurately identify the user's intent, and can obtain more complete and comprehensive data by using knowledge graph matching, thereby assisting users in making better decisions.

[0157] In an alternative embodiment, an electronic device is provided, such as Figure 8 As shown, Figure 8 The electronic device 4000 shown includes: a processor 4001 and a memory 4003. The processor 4001 and the memory 4003 are connected, such as through a bus 4002. Optionally, the electronic device 4000 may also include a transceiver 4004. It should be noted that in actual applications, the transceiver 4004 is not limited to one, and the structure of the electronic device 4000 does not constitute a limitation on the embodiments of the present application.

[0158] Processor 4001 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 devices, transistor logic devices, hardware components or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application. Processor 4001 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0159] The bus 4002 may include a path to transmit information between the above components. The bus 4002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus 4002 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 8 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0160] The memory 4003 may be a ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compressed optical disk, laser disk, optical disk, digital versatile disk, Blu-ray disk, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited to these.

[0161] The memory 4003 is used to store the application code for executing the solution of the present application, and the execution is controlled by the processor 4001. The processor 4001 is used to execute the application code stored in the memory 4003 to implement the contents shown in the above method embodiment.

[0162] The embodiment of the present application provides a computer-readable storage medium, which stores a computer program, and when the computer is run on a computer, the computer can execute the corresponding content in the aforementioned method embodiment. Compared with the prior art, by analyzing the search statement input by the user, determining the search intent, entities and the relationship between entities, the user's intent can be identified more directly and accurately, and more complete and comprehensive data can be obtained by using knowledge graph matching to assist users in making better decisions.

[0163] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0164] The above are only some implementation methods of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A resume search method, characterized in that: include: Obtaining a search statement, performing word segmentation processing on the search statement, and obtaining a word segmentation result; Determine the search intent, entities in the search sentence, and relationships between entities based on the word segmentation results; Obtaining search results from a talent pool and a pre-built knowledge graph according to the search intent, entities in the search statement, and relationships between entities, wherein the talent pool is used to store resumes, the knowledge graph is used to store entities in the resumes and relationships between entities, and the knowledge graph includes attributes of entities; Obtaining search results from a talent pool and a pre-built knowledge graph according to the search intent, entities in the search statement, and relationships between entities includes: Searching for a resume containing the entity in the search statement in a preset talent pool and taking it as the first target resume; Determine a first target entity in the first target resume, where the first target entity and the entity satisfy the relationship between the entities; According to the search intent and the attribute of the first target entity, searching for an entity with the same semantics as the first target entity in a preset knowledge graph as a second target entity; Searching the talent pool for a resume containing at least one of the first target entity and the second target entity as a second target resume, and using the second target resume as a search result; The determining the search intent, entities in the search sentence, and the relationship between entities according to the word segmentation result includes: Inputting the word segmentation result into a pre-trained intention recognition neural network model to obtain the search intention output by the intention recognition neural network model; Inputting the word segmentation result into a pre-trained entity recognition neural network model to obtain entities in the search sentence output by the entity recognition neural network model; Inputting the entities into a preset semantic relationship model to obtain the relationship between the entities output by the preset semantic relationship model; The step of searching the talent pool for a resume containing at least one of the first target entity and the second target entity as the second target resume and taking the second target resume as the search result, and then further including the step of sorting the search results, comprises: The score of the second target resume is determined based on the time difference between the last update of the second target resume and the current time, and the similarity between the search statement and the second target resume; Sorting the second target resumes in descending order of scores, and returning the sorted search results; The score is negatively correlated with the time difference between the last update of the second target resume and the current time, and is positively correlated with the similarity between the search statement and the second target resume.

2. The resume search method according to claim 1, characterized in that: The step of searching the talent pool for a resume containing at least one of the first target entity and the second target entity as the second target resume and using the second target resume as the search result, and then further including the step of expanding the search result, includes: Searching the knowledge graph for attribute information of entities contained in the second target resume; The attribute information is added to the second target resume to obtain an expanded search result; the attribute information of the entity refers to the characteristics of the entity.

3. A resume search device, characterized in that: include: An acquisition module is used to acquire a search statement, perform word segmentation processing on the search statement, and obtain a word segmentation result; An analysis module, used to determine the search intent, entities in the search statement, and the relationship between entities based on the word segmentation results; A search module, configured to obtain search results from a talent pool and a pre-built knowledge graph according to the search intent, entities in the search statement, and relationships between entities, wherein the knowledge graph includes attributes of entities; The search module includes: A first resume search module, used to search for resumes containing entities in the search statement in a preset talent pool and use them as first target resumes; A first entity search module, used to determine a first target entity in a first target resume, the first target entity and the entity satisfying an entity-to-entity relationship; A second entity search module, used to search for an entity with the same semantics as the first target entity in a preset knowledge graph according to the search intent and the attribute of the first target entity as a second target entity; A second resume search module, used to search the talent pool for a resume containing at least one of the first target entity and the second target entity as a second target resume, and use the second target resume as a search result; The analysis module comprises: An intention analysis module, used for inputting the word segmentation result into a pre-trained intention recognition neural network model to obtain the search intention output by the intention recognition neural network model; An entity analysis module, which is used to input the word segmentation results into a pre-trained entity recognition neural network model to obtain entities in the search sentence output by the entity recognition neural network model; A relationship analysis module, wherein the entity inputs a preset semantic relationship model and obtains the relationship between the entities output by the preset semantic relationship model; The second resume search module includes: A sorting module is used to sort the second target resume in descending order of score and return the sorted search results; wherein the score of the second target resume is determined based on the time difference between the last update of the second target resume and the current moment, and the similarity between the search statement and the second target resume; the score is negatively correlated with the time difference between the last update of the second target resume and the current moment, and is positively correlated with the similarity between the search statement and the second target resume.

4. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the resume searching method according to any one of claims 1 to 2 are implemented.

5. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions enable the computer to execute the steps of the resume searching method according to any one of claims 1 to 2.

Images