Method, device and equipment for generating employment records based on entity extraction

Through the named entity recognition model, the introduction text data is marked and combined, and the preset format of the job record is generated, which solves the problem of insufficient accuracy and interpretability of the job record extraction in the existing technology, and achieves high precision and high interpretability Job record generation.

CN114048744BActive Publication Date: 2024-11-26YANCHENG TIANYANCHAWEI TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202111266091.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-10-28
Publication Date
2024-11-26
Estimated Expiration
2041-10-28

AI Technical Summary

Technical Problem

The existing naming entity identification combination relationship extraction technology has insufficient accuracy and interpretability in obtaining the extraction results of company names and company positions, and it is difficult to generate accurate and interpretable job records.

Method used

A named entity recognition model, including a two-way long and short-term memory network and a conditional random field network, data annotation of the introduction text data, obtain the company entity and company position entities, and generate job records through preset format combinations, and use training samples to model Training and verification to improve the accuracy and interpretability of the decimation results.

Benefits of technology

It realizes high-precision extraction of positions in the company and company in massive data, improves the interpretability and controllability of the generated job records, avoids the correlation of irrelevant information, and improves the recall and interpretability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114048744B_ABST
    Figure CN114048744B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of data processing technology, and its implementation method provides a method, device and equipment for generating employment records based on entity extraction. The method for generating employment records based on entity extraction includes: obtaining profile text data; using a named entity recognition model to annotate the profile text data to obtain the employment company entity and company position entity in the profile text data; wherein the named entity recognition model includes a bidirectional long short-term memory network and a conditional random field network, and the named entity recognition model is obtained by training with annotated training samples; combining the employment company entity and the company position entity to generate an employment record in a preset format. The implementation method provided by the present invention can improve the interpretability of the generated employment records.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method for generating employment records based on entity extraction, a device for generating employment records based on entity extraction, a device for generating employment records based on entity extraction, and a corresponding storage medium. Background Art

[0002] With the development of deep learning technology, especially in the field of natural language processing (NLP), people's efficiency in processing unstructured data such as text has been greatly improved. Named entity recognition (NER) (also known as entity recognition, entity chunking, and entity extraction) is a subtask in natural language processing (NLP) that aims to locate and classify named entities in text into predefined categories, such as place names, people, organizations, locations, time expressions, quantities, monetary values, percentages, etc. The profiles of company executives generally include the person's previous companies and company positions. Using the named entity recognition technology in deep learning, the company name and company position are annotated to obtain a data set to train the model. The trained model can accurately extract the company name and company position. Subsequently, through structured combination, a relatively accurate person's company and the corresponding company position in the company can be obtained. The existing named entity recognition combined with relationship extraction technology can extract company names and corresponding company positions, but the results obtained by this extraction method are generally accurate and have low interpretability. Summary of the invention

[0003] The purpose of the embodiments of the present invention is to provide a method, device and equipment for generating employment records based on entity extraction.

[0004] In order to achieve the above object, the first aspect of the present invention provides a method for generating employment records based on entity extraction, the method comprising:

[0005] Acquire profile text data; use a named entity recognition model to annotate the profile text data to obtain the company entity and company position entity in the profile text data; wherein the named entity recognition model includes a bidirectional long short-term memory network and a conditional random field network, and the named entity recognition model is obtained by training with annotated training samples; combine the company entity and the company position entity to generate a job record in a preset format.

[0006] Preferably, the bidirectional long short-term memory network in the named entity recognition model is replaced by a pre-trained language representation model network.

[0007] Preferably, the named entity recognition model is obtained by training with labeled training samples, including: collecting training samples of profile text data; labeling the employed companies and company positions in the training samples; converting the labeled training samples into an adaptive data format for the named entity recognition model, and dividing the labeled training samples into a training data set, a verification data set and a test data set; using the training data set to train the named entity recognition model, using the verification data set to verify the trained named entity recognition model; and using the test data set to evaluate the trained named entity recognition model.

[0008] Preferably, the employed company entity and the company position entity are combined to generate an employed record in a preset format, including: obtaining the order of the employed company entity and the company position entity in the profile text data; traversing the employed company entity and the company position entity in the order, caching the employed company entity, and combining the company position entity with the cached employed company entity into an employed record in a preset format; the cached employed company entity is cleared from the cache when it is combined at least once and the next employed company entity enters the cache; the employed record in a preset format includes a employed company entity and a company position entity.

[0009] Preferably, before combining the company position entity with the cached company entity into a job record in a preset format, the method further includes: determining that the company entity does not exist in the cache, and combining the company position entity with a null company entity into a job record in a preset format.

[0010] Preferably, obtaining the order of the company entity and the company position entity in the profile text data includes: storing the company entity and the company position entity in the profile text data in sequence into a queue or array according to the data annotation; and determining the order of the company entity and the company position entity in the profile text data according to the storage positions of the company entity and the company position entity in the queue or array.

[0011] Preferably, the generation method further comprises: marking employment records in a preset format where the employment company entity is a null value.

[0012] In the second aspect of the present invention, there is also provided a device for generating employment records based on entity extraction, comprising: a data acquisition module for acquiring profile text data; an entity extraction module for annotating the profile text data using a named entity recognition model to obtain employment company entities and company position entities in the profile text data; wherein the named entity recognition model comprises a bidirectional long short-term memory network and a conditional random field network, and is obtained by training with annotated training samples; and a record generation module for combining the employment company entity and the company position entity to generate an employment record in a preset format.

[0013] In the third aspect of the present invention, there is also provided a device for generating employment records based on entity extraction, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the aforementioned method for generating employment records based on entity extraction when executing the computer program.

[0014] In a fourth aspect of the present invention, a computer-readable storage medium is provided, wherein instructions are stored in the storage medium, and when the instructions are executed on a computer, the computer executes the aforementioned method for generating employment records based on entity extraction.

[0015] A fifth aspect of the present invention provides a computer program product, comprising a computer program, which, when executed by a processor, implements the aforementioned method for generating employment records based on entity extraction.

[0016] The above technical solution has the following beneficial effects: the embodiment provided by the present invention is applied to an enterprise-level database, and the named entity recognition model trained with these massive data has a high accuracy in extracting the company and company position. By using the rules in the embodiment to perform the nearest match on the extracted company and company position, the interpretability of the finally generated employment record is ensured.

[0017] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings:

[0019] Figure 1 The following schematically shows an implementation diagram of a method for generating employment records based on entity extraction according to an embodiment of the present invention;

[0020] Figure 2A schematic diagram of an implementation of using training samples to train a named entity recognition model according to an embodiment of the present invention is shown;

[0021] Figure 3 The structure diagram of the device for generating employment records based on entity extraction according to an embodiment of the present invention is schematically shown. DETAILED DESCRIPTION

[0022] The specific implementation of the embodiment of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the embodiment of the present invention, and is not used to limit the embodiment of the present invention.

[0023] Figure 1 The following is a schematic diagram showing an implementation of a method for generating employment records based on entity extraction according to an embodiment of the present invention. Figure 1 As shown, this embodiment provides a method for generating employment records based on entity extraction, including:

[0024] S01, obtaining introduction text data;

[0025] The acquisition here includes but is not limited to direct input of text data, importing from a database, reading from a file, etc. In the database, there are multiple company names and company job titles stored, and these data can all become the data source for acquiring the profile text data in this step.

[0026] S02, using a named entity recognition model to annotate the profile text data, and obtain the company entity and company position entity in the profile text data; wherein the named entity recognition model includes a bidirectional long short-term memory network and a conditional random field network, and is obtained by training with annotated training samples;

[0027] The named entity recognition model composed of a bidirectional long short-term memory network and a conditional random field network is used for recognition. It is not only applicable to general types of named entities, but also can recognize specific named entities with high recognition accuracy.

[0028] Among them, it is obtained after training with labeled training samples, which roughly includes steps such as data collection, data labeling, sample construction, model training, and model verification.

[0029] S03: Combine the employment company entity and the company position entity to generate an employment record in a preset format.

[0030] In this embodiment, in order to make the generated employment records output in a standardized manner or to facilitate the next step of data processing, the format of the employment records can be preset, and the format includes column names, connectors, the format of company position entities and the format of employment company entities.

[0031] This step makes the employment records generated by the employment company entity and the company position entity more interpretable through a certain combination method, which is more conducive to subsequent data processing.

[0032] Through the above implementation, employment records can be quickly and accurately extracted from massive profile text data. Compared with the employment record generation methods in the prior art, it has better interpretability. Compared with the combination of named entity recognition and relationship extraction, the combination of named entity recognition and rules can greatly ensure the controllability of the extraction results, and prevent the problem that the company name and company position are obviously unrelated but are associated together, but there is no reasonable explanation for the occurrence of this problem. At the same time, this solution has a higher recall rate than relationship extraction.

[0033] In one embodiment of the present application, the bidirectional long short-term memory network in the named entity recognition model is replaced by a BERT network. The named entity recognition model in the previous embodiment uses a combination of LSTM+CRF to extract company names and company position entities, which has the advantages of simplicity, high efficiency, and easy parallel training. If there are no restrictions on the timeliness of the extraction, the precision and recall of entity extraction can be improved by using a combination of BERT+CRF. BERT (Bidirectional Encoder Representations from Transformers) is a pre-trained language representation model. It emphasizes that it is no longer pre-trained by using the traditional unidirectional language model or the shallow splicing of two unidirectional language models as in the past, but a new masked language model (MLM) is used to generate deep bidirectional language representations.

[0034] Figure 2 The schematic diagram of the implementation of using training samples to train a named entity recognition model according to an embodiment of the present invention is shown schematically. Figure 2 In one embodiment of the present application, the named entity recognition model is obtained by training with labeled training samples, and includes the following steps:

[0035] S11: Collect data;

[0036] In the database, multiple company names and company positions are stored. These company names and company positions are stored as brief text data or data, and the above data is pulled for subsequent data annotation;

[0037] S12: labeling the collected data;

[0038] In this embodiment, S12 specifically includes: labeling the company names and company positions in the data of step S11 by using a regular matching method or a third-party tool.

[0039] Among them, the quality of the named entity recognition model trained by deep learning is largely determined by the quality of the labeled training samples. Since it has a customized and complete set of company names and company positions, when the company names and company positions are labeled using regular matching methods combined with flexible labeling strategies, the set of training samples will have very good quality.

[0040] S13: converting the format of the annotated data and dividing the converted data into functions;

[0041] Named entity recognition has different annotation methods such as BIO, BIOES, BMES, etc. The annotated data set is converted into a data format recognizable by the deep learning model (here, the named entity recognition model) through a sequence processing program;

[0042] In this embodiment, dividing the converted data into functions specifically includes: dividing the converted data into a training data set (training set), a validation data set (validation set), and a test data set (test set).

[0043] S14: Build or select a model to be trained;

[0044] In the field of named entity recognition in deep learning, the model composed of BiLSTM and CRF is widely used because of its simplicity, high efficiency, and easy parallel training. BiLSTM can learn the semantic information of text sequences, including forward and reverse, and CRF can learn the feature transfer characteristics of text. The combination of the two has a considerable effect on general named entity recognition tasks. This step can be located anywhere before S15, and this is just an example.

[0045] S15: Train the model according to the divided data.

[0046] In this embodiment, the divided data includes: training data, verification data and test data. The training data is used to train the parameters in the model, the verification data is used to select the hyperparameters in the model, and the test data is used to test the effect of the model in a real production environment.

[0047] In this embodiment, before the data is input into the model, the character data is encoded into Arabic numerals that can be directly used by the model through a text processing program. By observing the size of parameters such as loss and acc during the model training process, the training effect of the model can be understood.

[0048] Through the above implementation, a named entity recognition model with higher recognition effect can be obtained.

[0049] In one embodiment of the present application, the employed company entity and the company position entity are combined to generate an employment record, including: obtaining the order of the employed company entity and the company position entity in the profile text data; traversing the employed company entity and the company position entity in the order, caching the employed company entity, and combining the company position entity with the cached employed company entity into an employment record in a preset format; the cached employed company entity is cleared from the cache when it is combined at least once and the next employed company entity enters the cache; the employment record in a preset format includes a employed company entity and a company position entity.

[0050] The prior art does not provide standard rules for combining and generating employment records. This embodiment follows the natural language logic in the original profile text data and performs preset processing rules on the extracted employment company entities and company position entities, thereby making the generated employment records more interpretable.

[0051] This implementation method divides the possible order of appearance of the serving company entity and the company position entity into the following multiple situations, and explains their processing processes respectively. Situation 1: If no serving company entity appears before the company position entity, a serving record including a null value and the company position entity is generated; Situation 2: If a company position entity appears after a continuous serving company entity, the company position entity and the continuous serving company entity are combined respectively to generate a serving record equal to the number of serving company entities; Situation 3: If a continuous company position entity appears after a discontinuous serving company entity, the serving company entity and the continuous company position entity are combined respectively to generate a serving record equal to the number of company position entities; Situation 4: If a continuous company position entity appears after a continuous serving company entity, it can be regarded as a combination of Situation 2 and Situation 3, and each serving company entity and each company position entity are combined with each other to form multiple serving records.

[0052] For example: the order of the obtained company entities and company position entities is: (company position entity 1, company entity 1, company entity 2, company position entity 2, company position entity 3, company entity 3, company position entity 4), then the following multiple employment records are generated: (null value, company position entity 1), (company entity 1, company position entity 2), (company entity 2, company position entity 2), (company entity 1, company position entity 3), (company entity 2, company position entity 3) and (company entity 3, company position entity 4). Among them, the combination of company entity 1 and company entity 2, company position entity 2 and company position entity 3 can generate 4 employment records. Through the above implementation method, a large number of company entities and company position entities can be batch processed by software programs to quickly generate more explainable employment records.

[0053] The following are steps for generating employment records using the named entity recognition model and generation rules, which are described in continuation of the aforementioned step S15, including:

[0054] S16: Annotate the introduction text data according to the named entity recognition model;

[0055] In this embodiment, after the model training is completed, a named entity recognition model is obtained, and the data of the profile text data of the company executives is annotated by using the model, so as to obtain the company entity and company position entity therein;

[0056] S17: storing the obtained company entity and company position entity;

[0057] In this embodiment, after obtaining the company entity and the company position entity, the company entity and the company position entity are stored in a preset manner;

[0058] Specifically, the company entities are traversed in order, and when a company entity is encountered, it is cached or stored. If the company entity is followed by another company entity, the company entities that appear continuously are recorded until a company position entity is encountered; the company position entities are traversed in order, and if the company position entity is followed by a company position entity, the company position entities that appear continuously are recorded;

[0059] S18: Combine the recorded company entity and company position entity according to a preset format.

[0060] Combine the company entity and company position entity cached in S17 to ensure that each company entity and each company position entity produce a job record.

[0061] To better illustrate the execution results of S17 to S18, an example is given below:

[0062] The profile text data is as follows: Zhang San: male, born in 1963, Chinese nationality, no permanent residence abroad, EMBA from Fudan University, senior engineer. From September 1994 to February 2012, he served as a director of A Co., Ltd.; from February 1998 to November 2005, he served as vice chairman and vice president of B Group Co., Ltd.; from September 2003 to March 2010, he served as chairman of C Communications Co., Ltd.; from January 2005 to March 2011, he served as chairman of D Culture Communication Co., Ltd.; from May 2007 to March 2014, he served as chairman of E Technology Co., Ltd.; from October 2011 to date, he has served as executive director and manager of F Investment Management Co., Ltd., and from April 2014 to date, he has served as the company's vice chairman.

[0063] The results after being processed by the method provided by the present invention are:

[0064] Name: Zhang San

[0065] Date of Birth: 1963

[0066] {'company':”,'job':'Senior Engineer'}

[0067] {'company':'A Co., Ltd.','job':'Director'}

[0068] {'company':'B Group Co., Ltd.','job':'Vice Chairman'}

[0069] {'company':'B Group Co., Ltd.','job':'Vice President'}

[0070] {'company':'C Communications Co., Ltd.','job':'Chairman'}

[0071] {'company':'D Culture Communication Co., Ltd.','job':'Chairman'}

[0072] {'company':'E Technology Co., Ltd.','job':'Chairman'}

[0073] {'company':'F Investment Management Co., Ltd.','job':'Executive Director'}

[0074] {'company':'F Investment Management Co., Ltd.','job':'Manager'}

[0075] {'company':'Our company','job':'Vice Chairman'}.

[0076] In one embodiment of the present application, obtaining the order of the company entities and the company position entities in the profile text data includes: storing the company entities and the company position entities in the profile text data in order into a queue or array according to the data annotations; determining the order of the company entities and the company position entities according to the storage positions of the company entities and the company position entities in the queue or array. In the actual process of processing the company entities and the company position entities using software programs, data structures such as queues or arrays can be used for implementation. First, an empty queue or array is established, and then by traversing the profile text data with data annotations, the company entities and the company position entities are extracted in turn and stored in the queue or array, so as to obtain a queue or array that reflects the order of the company entities and the company position entities.

[0077] In one embodiment of the present application, the generation method further includes: marking the employment record whose employment company entity is null. When the generated employment record is (null value, company position entity), it needs to be marked to remind that the employment record may need attention or further processing.

[0078] Figure 3 The schematic diagram of the structure of the employment record generation device based on entity extraction according to an embodiment of the present invention is shown as follows: Figure 3 As shown. This embodiment provides a device for generating employment records based on entity extraction, including: a data acquisition module for acquiring profile text data; an entity extraction module for annotating the profile text data using a named entity recognition model to obtain the employment company entity and the company position entity in the profile text data; wherein the named entity recognition model includes a bidirectional long short-term memory network and a conditional random field network, and is obtained by training with annotated training samples; and a record generation module for combining the employment company entity and the company position entity to generate an employment record in a preset format.

[0079] The specific definition of each functional module in the above-mentioned employment record generation device based on entity extraction can refer to the definition of the employment record generation method based on entity extraction in the above text, and will not be repeated here. Each module in the above-mentioned device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0080] In some embodiments provided by the present invention, there is also provided a device for generating employment records based on entity extraction, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the aforementioned method for generating employment records based on entity extraction when executing the computer program. The processor herein has the functions of numerical calculation and logical operation, and at least has a central processing unit CPU with data processing capability, a random access memory RAM, a read-only memory ROM, a variety of I / O ports, and an interrupt system, etc. The processor includes a kernel, and the kernel retrieves the corresponding program unit from the memory. One or more kernels may be provided, and the aforementioned method may be implemented by adjusting kernel parameters. The memory may include a non-permanent memory in a computer-readable medium, a random access memory (RAM), and / or a non-volatile memory, such as a read-only memory (ROM) or a flash memory (flash RAM), and the memory includes at least one storage chip.

[0081] In some embodiments provided by the present invention, a computer-readable storage medium is also provided, wherein instructions are stored in the storage medium, and when the instructions are executed on a computer, the computer executes the aforementioned method for generating employment records based on entity extraction.

[0082] In one embodiment of the present invention, a computer program product is provided, including a computer program, which implements the above-mentioned method for generating employment records based on entity extraction when executed by a processor.

[0083] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0084] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1A device that provides the functions specified in a block or multiple blocks.

[0085] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0086] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0087] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0088] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0089] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0090] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0091] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A method for generating employment records based on entity extraction, characterized in that: The method comprises: Get the introduction text data; A named entity recognition model is used to annotate the profile text data to obtain the company entity and company position entity in the profile text data; wherein the named entity recognition model includes a bidirectional long short-term memory network and a conditional random field network, and the named entity recognition model is obtained by training with annotated training samples; The employment company entity and the company position entity are combined to generate an employment record in a preset format, including: Obtaining the order of the company entity and the company position entity in the profile text data; Traversing the employed company entity and the company position entity in the order, caching the employed company entity, and combining the company position entity with the cached employed company entity into an employed record in a preset format; the cached employed company entity is cleared from the cache when it is combined at least once and the next employed company entity enters the cache; The employment record in a preset format includes an employment company entity and a company position entity; It is determined that there is no employed company entity in the cache, and the company position entity and the employed company entity with a null value are combined into an employed record in a preset format.

2. The method according to claim 1, characterized in that The bidirectional long short-term memory network in the named entity recognition model is replaced by a pre-trained language representation model network.

3. The method according to claim 1, characterized in that The named entity recognition model is obtained by training with labeled training samples, including: Collect training samples of profile text data; Labeling the companies and company positions in the training samples; Converting the labeled training samples into an adaptation data format of the named entity recognition model, and dividing the labeled training samples into a training data set, a verification data set, and a test data set; The training data set is used to train the named entity recognition model, the verification data set is used to verify the trained named entity recognition model; and the test data set is used to evaluate the trained named entity recognition model.

4. The method according to claim 1, characterized in that: Obtaining the order of the company entity and the company position entity in the profile text data includes: According to the data annotation, the company entity and the company position entity in the profile text data are stored in a queue or an array in order; The order of the company entity and the company position entity in the profile text data is determined according to the storage positions of the company entity and the company position entity in the queue or array.

5. The method according to claim 1, characterized in that The method further includes: marking employment records in a preset format in which the employment company entity is a null value.

6. A device for generating employment records based on entity extraction, characterized in that: include: A data acquisition module, used to acquire introduction text data; An entity extraction module, used to annotate the profile text data using a named entity recognition model to obtain company entities and company position entities in the profile text data; wherein the named entity recognition model includes a bidirectional long short-term memory network and a conditional random field network, and the named entity recognition model is obtained by training with annotated training samples; and The record generation module is used to combine the employment company entity and the company position entity to generate an employment record in a preset format, including: Obtaining the order of the company entity and the company position entity in the profile text data; Traversing the employed company entity and the company position entity in the order, caching the employed company entity, and combining the company position entity with the cached employed company entity into an employed record in a preset format; the cached employed company entity is cleared from the cache when it is combined at least once and the next employed company entity enters the cache; The employment record in a preset format includes an employment company entity and a company position entity; It is determined that there is no employed company entity in the cache, and the company position entity and the employed company entity with a null value are combined into an employed record in a preset format.

7. A device for generating employment records based on entity extraction, 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 computer program, the method for generating employment records based on entity extraction according to any one of claims 1 to 5 is implemented.

8. A computer-readable storage medium, characterized in that: The storage medium stores instructions, which, when executed on a computer, enable the computer to execute the method for generating employment records based on entity extraction as described in any one of claims 1 to 5.