CN111444718A - Insurance product demand document processing method and device and electronic equipment - Google Patents
Insurance product demand document processing method and device and electronic equipment Download PDFInfo
- Publication number
- CN111444718A CN111444718A CN202010172659.0A CN202010172659A CN111444718A CN 111444718 A CN111444718 A CN 111444718A CN 202010172659 A CN202010172659 A CN 202010172659A CN 111444718 A CN111444718 A CN 111444718A
- Authority
- CN
- China
- Prior art keywords
- entity
- insurance product
- product requirement
- requirement document
- target
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/08—Insurance
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Technology Law (AREA)
- General Business, Economics & Management (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The application provides a method and a device for processing insurance product requirement documents and electronic equipment, wherein the method comprises the following steps: acquiring a target insurance product requirement document to be processed; inputting the target insurance product requirement document into a pre-trained entity rule recognition model to obtain each entity in the target insurance product requirement document and an entity type corresponding to each entity; carrying out structured storage on each obtained entity and the entity type corresponding to each entity; the entity rule identification model is obtained based on sample insurance product requirement document training, and in the process of training the entity rule identification model, all entities in the training set insurance product requirement document in the sample insurance product requirement document and entity types corresponding to all the entities are marked. Therefore, according to the technical scheme provided by the embodiment of the application, the configuration efficiency and the configuration accuracy of the rule information in the insurance product requirement document are improved, and the development cycle of the insurance product is further shortened.
Description
Technical Field
The present application relates to the field of insurance technologies, and in particular, to a method and an apparatus for processing insurance product requirement documents, and an electronic device.
Background
With the rapid development of economy in China, the living standard of people is continuously improved, the risk guarantee consciousness is increasingly strengthened, a good external environment is provided for the rapid development of the insurance industry, and accordingly, insurance products meeting the requirements of people are continuously released by various insurance companies.
When an insurance product is released, insurance product requirements documents for the insurance product are written by insurance product personnel, and the insurance product requirements documents generally comprise a plurality of rules, such as purchasing rules of the insurance product. Since the rules included in the insurance product requirement documents of each insurance product are usually inconsistent, for each insurance product requirement document of each insurance product, a technician is required to read the rules, extract the rule information of the rules, and configure the extracted rule information into the rule database.
Since technicians manually read the plurality of rules, extract the rule information of the plurality of rules, and configure the extracted rule information to the rule database, a large amount of time is consumed, thereby resulting in low efficiency and long period for developing insurance products.
Disclosure of Invention
In order to solve the technical problems described in the background art, the application shows a method, a device and an electronic device for processing insurance product requirement documents.
In a first aspect, the present application shows a method for processing an insurance product requirement document, the method comprising:
acquiring a target insurance product requirement document to be processed;
inputting the target insurance product requirement document into a pre-trained entity rule recognition model to obtain each entity in the target insurance product requirement document and an entity type corresponding to each entity;
carrying out structured storage on each obtained entity and the entity type corresponding to each entity;
the entity rule identification model is obtained based on sample insurance product requirement document training, and in the process of training the entity rule identification model, all entities in a training set insurance product requirement document in the sample insurance product requirement document and entity types corresponding to all the entities are marked;
the labeling of each entity in the training set insurance product requirement documents in the sample insurance product requirement documents and the entity type corresponding to each entity includes:
performing word segmentation on the document content of the training set insurance product requirement document in the sample insurance product requirement document to obtain a word segmentation result; determining target word segmentation belonging to the entity in the word segmentation result; and labeling the entity types of the target participles, wherein the target participles belonging to the synonyms are labeled as the same entity type.
Optionally, the process of training the entity rule identification model further includes:
acquiring a sample insurance product demand document;
dividing the sample insurance product requirement document into a training set insurance product requirement document and a test set insurance product requirement document, wherein the training set insurance product requirement document is used for training a preset neural network model, and the test set insurance product requirement document is used for testing the preset neural network model;
inputting a training set insurance product requirement document labeled with each entity and the entity type corresponding to each entity into a predetermined neural network model, and training the predetermined neural network model to obtain a trained predetermined neural network model;
and inputting the requirement document of the test set insurance product into the trained preset neural network model, and determining the trained preset neural network model as a trained entity rule recognition model when the accuracy of the entity and the entity type output from the trained preset neural network model is greater than the preset accuracy.
Optionally, the insurance product requirement document includes: a purchase rule for an insurance product and/or an underwriting rule for an insurance product.
Optionally, when the insurance product requirement document includes an insurance rule of an insurance product, the target insurance product requirement document is input into a pre-trained entity rule recognition model to obtain each entity in the target insurance product requirement document and an entity type corresponding to each entity, where the method includes:
acquiring target document content used for describing the underwriting rule of the insurance product in the target insurance product requirement document;
and inputting the target document content into a pre-trained entity rule identification model to obtain each target entity in the target document content and a target entity type corresponding to each target entity.
Optionally, the performing structured storage on the obtained entities and the entity types corresponding to the entities includes:
and storing the obtained target entities and the target entity types corresponding to the target entities in an underwriting rule database in a structured data form.
Optionally, before the labeling the entity type of the target word segmentation, the method further includes:
and establishing a synonym dictionary, wherein words belonging to the same entity in the insurance industry are marked as synonyms in the synonym dictionary.
Optionally, the sample insurance product requirement document is an insurance product requirement document written by insurance product personnel.
In a second aspect, an embodiment of the present invention provides an insurance product requirement document processing apparatus, where the apparatus includes:
the insurance product requirement document acquisition module is used for acquiring a target insurance product requirement document to be processed;
the entity and entity type acquisition module is used for inputting the target insurance product requirement document into a pre-trained entity rule recognition model to obtain each entity in the target insurance product requirement document and an entity type corresponding to each entity;
the structured storage module is used for carrying out structured storage on the obtained entities and entity types corresponding to the entities;
wherein the entity rule recognition model is trained based on a sample insurance product requirement document, and the apparatus further comprises: an insurance product requirement document marking module; the insurance product requirement document marking module is used for marking each entity in a training set insurance product requirement document in the sample insurance product requirement document and the entity type corresponding to each entity in the training set insurance product requirement document in the process of training the entity rule identification model;
the insurance product requirement document marking module is specifically used for:
performing word segmentation on the document content of the training set insurance product requirement document in the sample insurance product requirement document to obtain a word segmentation result; determining target word segmentation belonging to the entity in the word segmentation result; and labeling the entity types of the target participles, wherein the target participles belonging to the synonyms are labeled as the same entity type.
Optionally, the apparatus further comprises: an entity rule recognition model training module, the entity rule recognition model training module is specifically configured to:
acquiring a sample insurance product demand document;
dividing the sample insurance product requirement document into a training set insurance product requirement document and a test set insurance product requirement document, wherein the training set insurance product requirement document is used for training a preset neural network model, and the test set insurance product requirement document is used for testing the preset neural network model;
inputting a training set insurance product requirement document labeled with each entity and the entity type corresponding to each entity into a predetermined neural network model, and training the predetermined neural network model to obtain a trained predetermined neural network model;
and inputting the requirement document of the test set insurance product into the trained preset neural network model, and determining the trained preset neural network model as a trained entity rule recognition model when the accuracy of the entity and the entity type output from the trained preset neural network model is greater than the preset accuracy.
Optionally, the insurance product requirement document includes: a purchase rule for an insurance product and/or an underwriting rule for an insurance product.
Optionally, when the insurance product requirement document includes an insurance rule of the insurance product, the entity and entity type obtaining module is specifically configured to:
acquiring target document content used for describing the underwriting rule of the insurance product in the target insurance product requirement document;
and inputting the target document content into a pre-trained entity rule identification model to obtain each target entity in the target document content and a target entity type corresponding to each target entity.
Optionally, the structured storage module is specifically configured to:
and storing the obtained target entities and the target entity types corresponding to the target entities in an underwriting rule database in a structured data form.
Optionally, the apparatus further comprises:
and the synonym dictionary establishing module is used for establishing a synonym dictionary before the entity type of the target participle is marked, wherein words belonging to the same entity in the insurance industry are marked as synonyms in the synonym dictionary.
Optionally, the sample insurance product requirement document is an insurance product requirement document written by insurance product personnel.
In a third aspect, an embodiment of the present invention provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of the insurance product requirement document processing method according to the first aspect when executing the program.
In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the steps of the insurance product requirement document processing method according to the first aspect.
In the technical scheme provided by the embodiment of the application, the electronic device can automatically process the insurance product requirement document, and after the inventor analyzes a large amount of insurance product requirement documents, the inventor finds out the rules included in the insurance product requirement documents, usually some entity rules described through natural language, so that after the entity rule identification module trained in advance identifies each entity of the insurance product requirement documents and the entity type corresponding to each entity, the rule information included in the insurance product requirement documents can be accurately extracted, and the entity types of each entity and each entity in the insurance product requirement documents can be structurally stored, so that the configuration efficiency of the rule information can be improved.
In addition, in the process of training the entity rule recognition model, the target participles belonging to the synonyms are marked as the same entity type, so that the recognition accuracy of the entity type corresponding to each entity can be extracted, and the configuration accuracy of the rule information is further improved. The method and the device for configuring the rule information do not need manual operation as in the prior art, the manual operation is low in configuration efficiency, the problems that the rule information is overlooked and the rule information is mistaken can be caused, and the like are solved.
Drawings
FIG. 1 is a flowchart illustrating steps of a method for processing a requirement document of an insurance product according to an embodiment of the present application;
FIG. 2 is a flowchart illustrating steps for training an entity rule recognition model according to an embodiment of the present disclosure;
FIG. 3 is a flow chart of steps of another insurance product requirement documentation method provided in an embodiment of the present application;
FIG. 4 is a flowchart illustrating steps of training an underwriting rule recognition model according to an embodiment of the present disclosure;
FIG. 5 is a flowchart illustrating steps of another method for processing a requirement document of an insurance product according to an embodiment of the present application;
FIG. 6 is a block diagram of an insurance product requirement document processing apparatus according to an embodiment of the present application;
fig. 7 is a block diagram of an electronic device according to an embodiment of the present application.
Detailed Description
In order to make the aforementioned objects, features and advantages of the present application more comprehensible, the present application is described in further detail with reference to the accompanying drawings and the detailed description.
Currently, when an insurance company releases an insurance product, insurance product personnel write an insurance product requirement document of the insurance product, and the insurance product requirement document usually comprises a plurality of rules. Since the rules included in the insurance product requirement documents of each insurance product are usually inconsistent, for each insurance product requirement document of each insurance product, a technician is required to read the rules, extract the rule information of the rules, and configure the extracted rule information into the rule database.
Since technicians manually read the multiple rules, extract the rule information of the multiple rules, and configure the extracted rule information to the rule database, a large amount of time is consumed, so that the efficiency is low, and the period for developing insurance products is long.
In order to solve the technical problem, an embodiment of the application provides a method and a device for processing insurance product requirement documents, and an electronic device.
In a first aspect, a method for processing a requirement document of an insurance product provided by an embodiment of the present application is first explained in detail.
Referring to fig. 1, a flowchart illustrating steps of a method for processing an insurance product requirement document according to the present application is shown, which may specifically include the following steps:
and S110, acquiring a target insurance product requirement document to be processed.
When an insurance company releases each insurance product, insurance product personnel can write the insurance product requirement document of the insurance product, so that the target insurance product requirement document can be any insurance product requirement document written by the insurance product personnel.
And S120, inputting the target insurance product requirement document into a pre-trained entity rule recognition model to obtain each entity in the target insurance product requirement document and an entity type corresponding to each entity.
After analyzing a large number of insurance product requirement documents, the inventor finds out the rules included in the insurance product requirement documents, which are usually some entity rules described by natural language, wherein the insurance product requirement documents may include: a purchase rule for an insurance product and/or an underwriting rule for an insurance product. Of course, in practical applications, the insurance product requirement document may further include other rules, which are not specifically limited in the embodiment of the present application.
Therefore, after the electronic device obtains the target insurance product requirement document, in order to extract rule information corresponding to the rules included in the insurance product requirement document and improve the accuracy of the determined entities and entity types, the electronic device may input the target insurance product requirement document into a pre-trained entity rule identification model to obtain each entity in the target insurance product requirement document and the entity type corresponding to each entity.
For example, one underwriting rule in the target insurance product rule requirement document is: [ person to be insured ] (customertype) [ age ] (age): [8 years of age ] (startage) (inclusive) - [49 years of age ] (inclusive). Then, the entities of the piece of underwriting rule are: administering to the insured life, age 8, age 49. The entity type corresponding to the insured person can be: customer type, the entity type corresponding to age may be: age, the entity type corresponding to age 8 may be: startage is the starting age, and the entity age corresponding to 49 years of age may be: endage is the ending age.
The entity rule identification model is obtained based on sample insurance product requirement document training, and in the process of training the entity rule identification model, all entities in the training set insurance product requirement documents in the sample insurance product requirement documents and entity types corresponding to all the entities are marked.
In addition, in order to enable the trained entity rule recognition model to more accurately recognize the entities and the entity types, the process of labeling each entity in the training set insurance product requirement document in the sample insurance product requirement document and the entity type corresponding to each entity may be as follows:
firstly, segmenting the document content of the sample insurance product requirement document to obtain a segmentation result, wherein the segmentation result is composed of a plurality of segments; then, in the word segmentation result, determining target word segmentation belonging to the entity, such as the insurable person, age and the like, belonging to the target word segmentation; and finally, marking the entity type of the target word segmentation. And when the entity type of the target participle is labeled, whether the participle belonging to the synonym exists in the target participle can be judged, and if the participle belonging to the synonym exists, the participle belonging to the synonym can be labeled as the same entity type. For example, an insured life is synonymous with an insured life, that is, the insured life and the insured life belong to the same entity, and therefore both synonyms of insured life and insured life are labeled as the same entity type. Therefore, the accuracy of each marked entity and the entity type corresponding to each entity is higher, and the entity type can be more accurately identified by the trained entity rule identification model.
Also, the insurance field is a relatively professional field. For example, the insured person and the insured person are the same entity, and need to be marked as synonyms, and the synonyms are put into a synonym dictionary, so that the synonyms can be correctly recognized as the same entity during training, and thus, the action objects (such as the insured person, age and the like) of the entity rules are ensured to be correct. Therefore, in one embodiment, before labeling the entity type of the target segmented word, the insurance product requirement document processing method may further include:
and establishing a synonym dictionary, wherein words belonging to the same entity in the insurance industry are marked as synonyms in the synonym dictionary.
Through the technical scheme provided by the embodiment of the invention, the synonym can be correctly identified as the same entity in the entity rule identification model, so that the action objects (such as insured persons, ages and the like) of the entity rule are ensured to be correct.
For clarity of description of the scheme, the training process of the neural network model will be explained in detail in the following embodiments.
S130, performing structured storage on each obtained entity and the entity type corresponding to each entity.
In order to realize the automatic configuration of the rule information in the insurance product requirement document, after obtaining each entity in the target insurance product requirement document and the entity type corresponding to each entity, the electronic equipment can store the obtained entities and entity types in a structured manner. For example, the entities and entity types may be stored in a rules database.
In the technical scheme provided by the embodiment of the application, the electronic device can automatically process the insurance product requirement document, and after the inventor analyzes a large amount of insurance product requirement documents, the inventor finds out the rules included in the insurance product requirement documents, usually some entity rules described through natural language, so that after the entity rule identification module trained in advance identifies each entity of the insurance product requirement documents and the entity type corresponding to each entity, the rule information included in the insurance product requirement documents can be accurately extracted, and the entity types of each entity and each entity in the insurance product requirement documents can be structurally stored, so that the configuration efficiency of the rule information can be improved.
In addition, in the process of training the entity rule recognition model, the target participles belonging to the synonyms are marked as the same entity type, so that the recognition accuracy of the entity type corresponding to each entity can be extracted, and the configuration accuracy of the rule information is further improved. The method and the device for configuring the rule information do not need manual operation as in the prior art, the manual operation is low in configuration efficiency, the problems that the rule information is overlooked and the rule information is mistaken can be caused, and the like are solved.
For clarity of description of the scheme, the following embodiments will explain the training process of the entity rule recognition model in detail. As shown in fig. 2, the process of training the entity rule recognition model may further include the following steps:
s210, obtaining a sample insurance product requirement document.
When an insurance company releases each insurance product, insurance product personnel write insurance product requirement documents of the insurance product, and in order to enable the identification accuracy of the trained entity identification model to be higher, the sample insurance product requirement documents can be the insurance product requirement documents written by the insurance product personnel.
It can be appreciated that in training a neural network model, a large number of sample insurance product requirement documents need to be collected. The greater the number of collected sample insurance product requirement documents, the higher the recognition accuracy of the trained neural network model is.
S220, dividing the sample insurance product requirement documents into training set insurance product requirement documents and testing set insurance product requirement documents.
The system comprises a training set insurance product requirement document, a test set insurance product requirement document and a test set insurance product requirement document, wherein the training set insurance product requirement document is used for training a preset neural network model, and the test set insurance product requirement document is used for testing the preset neural network model.
Specifically, when the sample insurance product requirement document is obtained, the sample insurance product requirement document can be divided into two parts, one part is a training set insurance product requirement document, the training set insurance product requirement document is used for training a predetermined neural network model, and the other part is a test set insurance product requirement document. The test set insurance product requirement document is used for testing the predetermined neural network model, namely for testing whether the recognition result of the predetermined neural network model is accurate.
And S230, inputting the requirement documents of the training set insurance products marked with the entities and the entity types corresponding to the entities into a preset neural network model, and training the preset neural network model to obtain the trained preset neural network model.
Since the trained entity rule recognition model needs to recognize entities in the insurance product requirement document, as well as entity types. Therefore, each entity in the requirement document of the insurance product in the training set and the entity type corresponding to each entity need to be labeled.
Since the process of labeling each entity in the training set insurance product requirement document in the sample insurance product requirement document and the entity type corresponding to each entity has already been explained in detail in the embodiment shown in fig. 1, no further description is given here.
S240, inputting the requirement documents of the training set insurance products marked with the entities and the entity types corresponding to the entities into a predetermined neural network model, and training the predetermined neural network model to obtain the trained predetermined neural network model.
Specifically, after entity and entity type labeling is carried out on the training set insurance product requirement documents, the training set insurance product requirement documents are input into the predetermined neural network model, and the predetermined neural network model is trained to obtain the trained predetermined neural network model.
And S250, inputting the requirement document of the test set insurance product into the trained preset neural network model, and determining the trained preset neural network model as the trained entity rule recognition model when the accuracy of the entity and the entity type output from the trained preset neural network model is greater than the preset accuracy.
After the trained predetermined neural network model is obtained, in order to test whether the recognition result of the trained predetermined neural network model is accurate, the requirement document of the test set insurance product can be input into the trained predetermined neural network model, whether the accuracy of the entity and the entity type output from the trained predetermined neural network model is greater than the preset accuracy is judged, and when the accuracy of the entity and the entity type output from the trained predetermined neural network model is greater than the preset accuracy, the accuracy of the recognition result of the trained predetermined neural network model is higher, so that the trained predetermined neural network model is determined to be the trained entity rule recognition model. If the accuracy of the entity and the entity type output from the trained predetermined neural network model is less than the preset accuracy, it indicates that the accuracy of the recognition result of the trained predetermined neural network model is not high enough, and therefore, the trained predetermined neural network model needs to be trained again until the accuracy of the entity and the entity type output from the trained predetermined neural network model is greater than the preset accuracy.
The preset accuracy may be determined according to an actual situation, for example, the preset accuracy may be 99%, and the preset accuracy is not specifically limited in the embodiments of the present invention.
Therefore, by the technical scheme provided by the embodiment, the recognition accuracy of the trained entity rule recognition model is higher, so that the entity of the target insurance product requirement document and the entity type corresponding to the entity can be recognized accurately.
Referring to fig. 3, a flow chart showing steps of another insurance product requirement document processing method of the present application may specifically include the following steps:
s310, obtaining a target insurance product requirement document to be processed.
Since step S310 is the same as step S110, step S110 has already been described in detail in the embodiment shown in fig. 1, and therefore step S310 is not described again here.
S320, obtaining the target document content of the underwriting rule of the insurance product in the target insurance product requirement document.
The inventor can find that almost all insurance product requirement documents contain a plurality of pieces of underwriting information of the insurance products by analyzing a large number of existing insurance product requirement documents, namely the underwriting rules are usually included in the insurance product requirement documents.
Therefore, after the electronic device obtains the target insurance product requirement document, the electronic device can intercept the target document content of the insurance product requirement document for describing the underwriting rule of the insurance product.
S330, inputting the target document content into the entity rule recognition model trained in advance to obtain each target entity in the target document content and the target entity type corresponding to each target entity.
After the target document content of the underwriting rule of the insurance product is described in the target insurance product requirement document, the target document content can be input into a pre-trained entity rule recognition model, and then each target entity in the target document content and the target entity type corresponding to each target entity are accurately obtained.
S340, storing each target entity and the target entity type corresponding to each target entity in an underwriting rule database in the form of structured data.
In order to realize the automatic configuration of the insurance product underwriting rules, after obtaining each target entity in the target document content and the target entity type corresponding to each target entity, each target entity and the target entity type corresponding to each target entity can be stored in the underwriting rule database in the form of structured data, so that the automatic configuration of the underwriting rules is realized.
In the technical scheme provided by the embodiment of the application, the electronic device can automatically extract the target document content for describing the underwriting rule in the insurance product requirement document, extract the target entity in the target document content and the target entity type corresponding to each target entity through a pre-trained entity rule identification model, and improve the configuration efficiency of the underwriting rule through structured storage of the target entity and the entity type corresponding to each target entity. In addition, in the process of training the entity rule recognition model, the target participles belonging to the synonyms are marked as the same entity type, so that the recognition accuracy of the entity type corresponding to each entity can be extracted, and the configuration accuracy of the rule information is further improved. The method and the device for configuration of the insurance product do not need manual operation like the prior art, the manual operation is low in configuration efficiency, and the problems that rule information is overlooked and the rule information is mistaken can be caused.
For clarity of description of the scheme, the following describes the training process of the underwriting rule recognition model in detail by taking the neural network model as the underwriting rule recognition model as an example. As shown in fig. 4. The training process of the underwriting rule recognition model can be divided into the following steps:
s410, collecting sample insurance product requirement documents.
Wherein the sample insurance product requirement document may be an insurance product requirement document written by an insurance product staff.
And S420, intercepting document contents for describing the underwriting rules from the sample insurance product requirement document.
For example, taking age restrictions in the underwriting rules as an example, the intercepted document content for describing underwriting rules may be: age of person to be insured: 8 years old (inclusive) -49 years old (inclusive); the insured age is between 18 and the year of the week; alternatively, the insured life is between 18startage and 69 years of age.
S430, dividing the intercepted document content for describing the underwriting rule into test set data and training set data; and annotating entities and entity types in the test set data.
For example, the annotation result may be:
[ insured person ] (customer type) [ age ] (age): age [ 8] (start) (inclusive) to age [49 ] (end) (inclusive).
[ insured ] (customertype) [ age ] (age) is between [18] (start) and [69 years of age ] (end).
[ insured ] (customertype) is between [18] (start) and [69 years of age ] (end).
Where brackets [ ] represent entities and brackets () are entity types.
S440, synonym processing is carried out on the annotation data.
Specifically, natural language entity recognition is based on word segmentation, and insurance is a relatively specialized domain. For example, the insured person and the insured person are the same entity, and need to be marked as synonyms, and the synonyms are put into a synonym dictionary, so that the synonyms can be correctly recognized as the same entity during training, and thus, the action objects (such as the insured person, age and the like) of the underwriting rules are ensured to be correct.
S450, training an underwriting rule recognition model.
Specifically, the document content labeled through the two steps S430 and S440 and used for describing the underwriting rule is used for training the underwriting rule recognition model.
And S460, testing the underwriting rule identification model.
Specifically, the test set data is input to the underwriting rule identification model generated in step S450 to determine the identification accuracy of the underwriting rule identification model. For example, insurance product requirements documents describe: age of insured: 18-60 years old, and the identification model can identify: insured, age, 18, 60 years of age.
And S470, judging whether the test result achieves the expected effect. That is, it is determined whether the recognition accuracy reaches a preset recognition accuracy. If yes, go to step S480, otherwise, go back to step S430.
Specifically, if the test result reaches the expected result, that is, the identification accuracy reaches the preset identification accuracy, it indicates that the identification accuracy of the underwriting rule identification model is high, and therefore, S480 is executed to obtain the final underwriting rule identification model, otherwise, the step S430 is returned to, that is, the underwriting rule identification model is continuously optimized.
And S480, acquiring a final underwriting rule identification model.
Based on the embodiment shown in fig. 4, after the underwriting rule identification model is available, the underwriting rule in the new insurance product requirement document can be automatically identified by using the underwriting rule identification model. The following will describe in detail a specific embodiment of the automated writing of the underwriting rule into the underwriting rule base, and as shown in fig. 5, the following steps may be performed:
and S510, acquiring a new insurance product requirement document.
S520, acquiring document contents used for describing the underwriting rules in the new insurance product requirement documents.
S530, inputting the document content for describing the underwriting rule into the trained underwriting rule identification model, and identifying structured underwriting rule data, namely the entity and the entity type, in the document content of the underwriting rule.
And S540, writing the structured underwriting rule data into an underwriting rule database.
Through the above four steps S510 to S540, the automatic configuration of the underwriting rules in the insurance product requirement document can be completed.
It is noted that, for simplicity of explanation, the method embodiments are described as a series of acts or combination of acts, but those skilled in the art will appreciate that the present application is not limited by the order of acts, as some steps may, in accordance with the present application, occur in other orders and concurrently. Further, those skilled in the art will also appreciate that the embodiments described in the specification are exemplary and that no action is necessarily required in this application.
In a second aspect, an embodiment of the present invention provides an insurance product requirement document processing apparatus.
Referring to fig. 6, a block diagram of an insurance product requirement document processing apparatus according to the present application is shown, and the apparatus may specifically include the following modules:
an insurance product requirement document acquisition module 610, configured to acquire a target insurance product requirement document to be processed;
an entity and entity type obtaining module 620, configured to input the target insurance product requirement document into a pre-trained entity rule recognition model, so as to obtain each entity in the target insurance product requirement document and an entity type corresponding to each entity;
a structured storage module 630, configured to perform structured storage on the obtained entities and entity types corresponding to the entities;
wherein the entity rule recognition model is trained based on a sample insurance product requirement document, and the apparatus further comprises: an insurance product requirement document marking module; the insurance product requirement document marking module is used for marking each entity in a training set insurance product requirement document in the sample insurance product requirement document and the entity type corresponding to each entity in the training set insurance product requirement document in the process of training the entity rule identification model;
the insurance product requirement document marking module is specifically used for:
performing word segmentation on the document content of the training set insurance product requirement document in the sample insurance product requirement document to obtain a word segmentation result; determining target word segmentation belonging to the entity in the word segmentation result; and labeling the entity types of the target participles, wherein the target participles belonging to the synonyms are labeled as the same entity type.
In the technical scheme provided by the embodiment of the application, the electronic device can automatically process the insurance product requirement document, and after the inventor analyzes a large amount of insurance product requirement documents, the inventor finds out the rules included in the insurance product requirement documents, usually some entity rules described through natural language, so that after the entity rule identification module trained in advance identifies each entity of the insurance product requirement documents and the entity type corresponding to each entity, the rule information included in the insurance product requirement documents can be accurately extracted, and the entity types of each entity and each entity in the insurance product requirement documents can be structurally stored, so that the configuration efficiency of the rule information can be improved.
In addition, in the process of training the entity rule recognition model, the target participles belonging to the synonyms are marked as the same entity type, so that the recognition accuracy of the entity type corresponding to each entity can be extracted, and the configuration accuracy of the rule information is further improved. The method and the device for configuring the rule information do not need manual operation as in the prior art, the manual operation is low in configuration efficiency, the problems that the rule information is overlooked and the rule information is mistaken can be caused, and the like are solved.
Optionally, the apparatus further comprises: an entity rule recognition model training module, the entity rule recognition model training module is specifically configured to:
acquiring a sample insurance product demand document;
dividing the sample insurance product requirement document into a training set insurance product requirement document and a test set insurance product requirement document, wherein the training set insurance product requirement document is used for training a preset neural network model, and the test set insurance product requirement document is used for testing the preset neural network model;
inputting a training set insurance product requirement document labeled with each entity and the entity type corresponding to each entity into a predetermined neural network model, and training the predetermined neural network model to obtain a trained predetermined neural network model;
and inputting the requirement document of the test set insurance product into the trained preset neural network model, and determining the trained preset neural network model as a trained entity rule recognition model when the accuracy of the entity and the entity type output from the trained preset neural network model is greater than the preset accuracy.
Optionally, the insurance product requirement document includes: a purchase rule for an insurance product and/or an underwriting rule for an insurance product.
Optionally, when the insurance product requirement document includes an insurance rule of the insurance product, the entity and entity type obtaining module is specifically configured to:
acquiring target document content used for describing the underwriting rule of the insurance product in the target insurance product requirement document;
and inputting the target document content into a pre-trained entity rule identification model to obtain each target entity in the target document content and a target entity type corresponding to each target entity.
Optionally, the structured storage module is specifically configured to:
and storing the obtained target entities and the target entity types corresponding to the target entities in an underwriting rule database in a structured data form.
Optionally, the apparatus further comprises:
and the synonym dictionary establishing module is used for establishing a synonym dictionary before the entity type of the target participle is marked, wherein words belonging to the same entity in the insurance industry are marked as synonyms in the synonym dictionary.
Optionally, the sample insurance product requirement document is an insurance product requirement document written by insurance product personnel.
For the device embodiment, since it is basically similar to the method embodiment, the description is simple, and for the relevant points, refer to the partial description of the method embodiment.
In a third aspect, an embodiment of the present invention provides an electronic device, as shown in fig. 7, including a memory 710, a processor 720, and a computer program stored on the memory and executable on the processor, where the processor implements the steps of the insurance product requirement document processing method according to the first aspect when executing the program.
In the technical scheme provided by the embodiment of the application, the electronic device can automatically process the insurance product requirement document, and after the inventor analyzes a large amount of insurance product requirement documents, the inventor finds out the rules included in the insurance product requirement documents, usually some entity rules described through natural language, so that after the entity rule identification module trained in advance identifies each entity of the insurance product requirement documents and the entity type corresponding to each entity, the rule information included in the insurance product requirement documents can be accurately extracted, and the entity types of each entity and each entity in the insurance product requirement documents can be structurally stored, so that the configuration efficiency of the rule information can be improved.
In addition, in the process of training the entity rule recognition model, the target participles belonging to the synonyms are marked as the same entity type, so that the recognition accuracy of the entity type corresponding to each entity can be extracted, and the configuration accuracy of the rule information is further improved. The method and the device for configuring the rule information do not need manual operation as in the prior art, the manual operation is low in configuration efficiency, the problems that the rule information is overlooked and the rule information is mistaken can be caused, and the like are solved.
In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the steps of the insurance product requirement document processing method according to the first aspect.
In the technical scheme provided by the embodiment of the application, the electronic device can automatically process the insurance product requirement document, and after the inventor analyzes a large amount of insurance product requirement documents, the inventor finds out the rules included in the insurance product requirement documents, usually some entity rules described through natural language, so that after the entity rule identification module trained in advance identifies each entity of the insurance product requirement documents and the entity type corresponding to each entity, the rule information included in the insurance product requirement documents can be accurately extracted, and the entity types of each entity and each entity in the insurance product requirement documents can be structurally stored, so that the configuration efficiency of the rule information can be improved.
In addition, in the process of training the entity rule recognition model, the target participles belonging to the synonyms are marked as the same entity type, so that the recognition accuracy of the entity type corresponding to each entity can be extracted, and the configuration accuracy of the rule information is further improved. The method and the device for configuring the rule information do not need manual operation as in the prior art, the manual operation is low in configuration efficiency, the problems that the rule information is overlooked and the rule information is mistaken can be caused, and the like are solved.
The embodiments in the present specification are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, apparatus, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, terminal devices (systems), and computer program products according to the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing terminal to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing terminal to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing terminal to cause a series of operational steps to be performed on the computer or other programmable terminal to produce a computer implemented process such that the instructions which execute on the computer or other programmable terminal provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While the preferred embodiments of the present application have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all alterations and modifications as fall within the scope of the application.
Finally, it should also be noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or terminal. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or terminal that comprises the element.
The insurance product requirement document processing method and the insurance product requirement document processing device provided by the application are introduced in detail, and specific examples are applied in the text to explain the principle and the implementation of the application, and the description of the above embodiments is only used to help understand the method and the core idea of the application. Meanwhile, for a person skilled in the art, according to the idea of the present application, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present application.
Claims (10)
1. An insurance product requirement document processing method, characterized in that the method comprises:
acquiring a target insurance product requirement document to be processed;
inputting the target insurance product requirement document into a pre-trained entity rule recognition model to obtain each entity in the target insurance product requirement document and an entity type corresponding to each entity;
carrying out structured storage on each obtained entity and the entity type corresponding to each entity;
the entity rule identification model is obtained based on sample insurance product requirement document training, and in the process of training the entity rule identification model, all entities in a training set insurance product requirement document in the sample insurance product requirement document and entity types corresponding to all the entities are marked;
the labeling of each entity in the training set insurance product requirement documents in the sample insurance product requirement documents and the entity type corresponding to each entity includes:
performing word segmentation on the document content of the training set insurance product requirement document in the sample insurance product requirement document to obtain a word segmentation result; determining target word segmentation belonging to the entity in the word segmentation result; and labeling the entity types of the target participles, wherein the target participles belonging to the synonyms are labeled as the same entity type.
2. The method of claim 1, wherein the process of training the entity rule recognition model further comprises:
acquiring a sample insurance product demand document;
dividing the sample insurance product requirement document into a training set insurance product requirement document and a test set insurance product requirement document, wherein the training set insurance product requirement document is used for training a preset neural network model, and the test set insurance product requirement document is used for testing the preset neural network model;
inputting a training set insurance product requirement document labeled with each entity and the entity type corresponding to each entity into a predetermined neural network model, and training the predetermined neural network model to obtain a trained predetermined neural network model;
and inputting the requirement document of the test set insurance product into the trained preset neural network model, and determining the trained preset neural network model as a trained entity rule recognition model when the accuracy of the entity and the entity type output from the trained preset neural network model is greater than the preset accuracy.
3. The method of claim 1, wherein the insurance product requirement document comprises: a purchase rule for an insurance product and/or an underwriting rule for an insurance product.
4. The method according to claim 3, wherein when the insurance product requirement document includes an insurance rule of an insurance product, the inputting the target insurance product requirement document into a pre-trained entity rule recognition model to obtain each entity in the target insurance product requirement document and an entity type corresponding to each entity includes:
acquiring target document content used for describing the underwriting rule of the insurance product in the target insurance product requirement document;
and inputting the target document content into a pre-trained entity rule identification model to obtain each target entity in the target document content and a target entity type corresponding to each target entity.
5. The method according to claim 4, wherein the structured storage of the obtained entities and the entity types corresponding to the entities comprises:
and storing the obtained target entities and the target entity types corresponding to the target entities in an underwriting rule database in a structured data form.
6. The method of any one of claims 1 to 5, wherein prior to said annotating an entity type of said target participle, said method further comprises:
and establishing a synonym dictionary, wherein words belonging to the same entity in the insurance industry are marked as synonyms in the synonym dictionary.
7. The method of any one of claims 1 to 5, wherein the sample insurance product requirement document is an insurance product requirement document that has been drafted by an insurance product staff.
8. An insurance product requirement document processing apparatus, the apparatus comprising:
the insurance product requirement document acquisition module is used for acquiring a target insurance product requirement document to be processed;
the entity and entity type acquisition module is used for inputting the target insurance product requirement document into a pre-trained entity rule recognition model to obtain each entity in the target insurance product requirement document and an entity type corresponding to each entity;
the structured storage module is used for carrying out structured storage on the obtained entities and entity types corresponding to the entities;
wherein the entity rule recognition model is trained based on a sample insurance product requirement document, and the apparatus further comprises: an insurance product requirement document marking module; the insurance product requirement document marking module is used for marking each entity in a training set insurance product requirement document in the sample insurance product requirement document and the entity type corresponding to each entity in the training set insurance product requirement document in the process of training the entity rule identification model;
the insurance product requirement document marking module is specifically used for:
performing word segmentation on the document content of the training set insurance product requirement document in the sample insurance product requirement document to obtain a word segmentation result; determining target word segmentation belonging to the entity in the word segmentation result; and labeling the entity types of the target participles, wherein the target participles belonging to the synonyms are labeled as the same entity type.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the steps of the insurance product requirement document processing method according to any one of claims 1 to 7 are implemented when the processor executes the program.
10. A computer-readable storage medium, characterized in that a computer program is stored thereon, which computer program, when being executed by a processor, carries out the steps of the insurance product requirement document processing method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010172659.0A CN111444718A (en) | 2020-03-12 | 2020-03-12 | Insurance product demand document processing method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010172659.0A CN111444718A (en) | 2020-03-12 | 2020-03-12 | Insurance product demand document processing method and device and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111444718A true CN111444718A (en) | 2020-07-24 |
Family
ID=71648695
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010172659.0A Pending CN111444718A (en) | 2020-03-12 | 2020-03-12 | Insurance product demand document processing method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111444718A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112328210A (en) * | 2020-11-18 | 2021-02-05 | 平安普惠企业管理有限公司 | Neural network-based demand document updating method, device, equipment and medium |
| CN112446792A (en) * | 2020-12-01 | 2021-03-05 | 中国人寿保险股份有限公司 | Benefit demonstration generation method and device, electronic equipment and storage medium |
| CN113627189A (en) * | 2021-08-17 | 2021-11-09 | 青岛全掌柜科技有限公司 | Entity identification information extraction, storage and display method for insurance clauses |
| CN113902568A (en) * | 2021-10-30 | 2022-01-07 | 平安科技（深圳）有限公司 | Method for identifying green asset proportion and related product |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106933802A (en) * | 2017-02-24 | 2017-07-07 | 黑龙江特士信息技术有限公司 | A kind of social security class entity recognition method and device towards multi-data source |
| CN109871545A (en) * | 2019-04-22 | 2019-06-11 | 京东方科技集团股份有限公司 | Name entity recognition method and device |
| CN110276054A (en) * | 2019-05-16 | 2019-09-24 | 湖南大学 | A kind of insurance text structure implementation method |
| CN110704633A (en) * | 2019-09-04 | 2020-01-17 | 平安科技（深圳）有限公司 | Named entity recognition method and device, computer equipment and storage medium |
| CN110765754A (en) * | 2019-09-16 | 2020-02-07 | 平安科技（深圳）有限公司 | Text data typesetting method and device, computer equipment and storage medium |
-
2020
- 2020-03-12 CN CN202010172659.0A patent/CN111444718A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106933802A (en) * | 2017-02-24 | 2017-07-07 | 黑龙江特士信息技术有限公司 | A kind of social security class entity recognition method and device towards multi-data source |
| CN109871545A (en) * | 2019-04-22 | 2019-06-11 | 京东方科技集团股份有限公司 | Name entity recognition method and device |
| CN110276054A (en) * | 2019-05-16 | 2019-09-24 | 湖南大学 | A kind of insurance text structure implementation method |
| CN110704633A (en) * | 2019-09-04 | 2020-01-17 | 平安科技（深圳）有限公司 | Named entity recognition method and device, computer equipment and storage medium |
| CN110765754A (en) * | 2019-09-16 | 2020-02-07 | 平安科技（深圳）有限公司 | Text data typesetting method and device, computer equipment and storage medium |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112328210A (en) * | 2020-11-18 | 2021-02-05 | 平安普惠企业管理有限公司 | Neural network-based demand document updating method, device, equipment and medium |
| CN112446792A (en) * | 2020-12-01 | 2021-03-05 | 中国人寿保险股份有限公司 | Benefit demonstration generation method and device, electronic equipment and storage medium |
| CN113627189A (en) * | 2021-08-17 | 2021-11-09 | 青岛全掌柜科技有限公司 | Entity identification information extraction, storage and display method for insurance clauses |
| CN113902568A (en) * | 2021-10-30 | 2022-01-07 | 平安科技（深圳）有限公司 | Method for identifying green asset proportion and related product |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111401609B (en) | Prediction method and prediction device for traffic flow time series | |
| CN108256074B (en) | Verification processing method and device, electronic equipment and storage medium | |
| CN111444718A (en) | Insurance product demand document processing method and device and electronic equipment | |
| US8577884B2 (en) | Automated analysis and summarization of comments in survey response data | |
| US20200257659A1 (en) | Method and apparatus for determing description information, electronic device and computer storage medium | |
| CN112163553B (en) | Material price accounting method, device, storage medium and computer equipment | |
| CN112579466B (en) | Method and device for generating test cases and computer readable storage medium | |
| Martens et al. | Extracting and analyzing context information in user-support conversations on twitter | |
| CN116244410B (en) | Index data analysis method and system based on knowledge graph and natural language | |
| US10782942B1 (en) | Rapid onboarding of data from diverse data sources into standardized objects with parser and unit test generation | |
| CN110866394A (en) | Company name identification method and device, computer equipment and readable storage medium | |
| CN109558580B (en) | Text analysis method and device | |
| CN113779983B (en) | Text data processing method and device, storage medium and electronic device | |
| CN112328812B (en) | Domain knowledge extraction method and system based on self-adjusting parameters and electronic equipment | |
| TWI777163B (en) | Form data detection method, computer device and storage medium | |
| Born et al. | Sequence models for document structure identification in an undeciphered script | |
| CN113901793A (en) | Event extraction method and device combining RPA and AI | |
| CN111488737B (en) | Text recognition method, device and equipment | |
| CN112989795A (en) | Text information extraction method and device, computer equipment and storage medium | |
| CN113722421A (en) | Contract auditing method and system and computer readable storage medium | |
| CN113642318B (en) | Method, system, storage medium and device for correcting English article | |
| CN113111155B (en) | Information display method, device, equipment and storage medium | |
| CN114049528B (en) | Brand name identification method and equipment | |
| CN115659182B (en) | Model updating method, device and equipment | |
| CN115130459A (en) | Method and system for automatic information abstraction processing of document |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |