CN113705230A

CN113705230A - Artificial intelligence-based policy agreement assessment method, device, equipment and medium

Info

Publication number: CN113705230A
Application number: CN202111014577.4A
Authority: CN
Inventors: 叶向荣
Original assignee: Ping An Property and Casualty Insurance Company of China Ltd
Current assignee: Ping An Property and Casualty Insurance Company of China Ltd
Priority date: 2021-08-31
Filing date: 2021-08-31
Publication date: 2021-11-26
Anticipated expiration: 2041-08-31
Also published as: CN113705230B

Abstract

The application relates to the technical field of artificial intelligence, and discloses a method, a device, equipment and a medium for evaluating a policy speciality based on artificial intelligence, wherein the method comprises the following steps: performing word segmentation on the policy agreement to be evaluated to obtain a target policy agreement phrase set; obtaining a target grammatical error phrase set from the target policy special phrase set; evaluating semantic relevance according to the target policy special phrase set to obtain a target semantic relevance evaluation result; performing risk evaluation on a target policy special offer phrase set according to the acquired policy special offer semantic library to obtain a target risk evaluation result; and generating a special evaluation result corresponding to the target policy according to the target grammar error phrase set, the target semantic correlation evaluation result and the target risk evaluation result. The automatic evaluation of grammar errors, semantic relevance and risk of the order-issuing person is realized, the requirements on the insurance professional level and the special agreement standard of the order-issuing person are reduced, and the accuracy of the special agreement of the policy is improved.

Description

Artificial intelligence-based policy agreement assessment method, device, equipment and medium

Technical Field

The present application relates to the field of artificial intelligence technologies, and in particular, to a method, an apparatus, a device, and a medium for evaluating a policy agreement based on artificial intelligence.

Background

In the insurance industry, since the format terms of the policy cannot completely meet the requirements of the client, special agreement terms (abbreviated as special agreement) need to be tailored in the policy according to the personalized requirements of the client. The specials in the policy play an important role in the execution of processes and risk control of insurance companies and insurance clients with respect to the fulfillment and/or reimbursement of subsequent policies. The traditional special agreement is a term which is manually added or edited by an insurance policy issuing person after communicating with a client, the requirement of the client on understanding the insurance professional level and the special agreement specification is higher due to complex and variable requirements, and the insurance policy is invalid or has potential risk due to the phenomena of ambiguity, conflict with format descriptions in an insurance policy template and the like easily generated when the issuing person makes the special agreement.

Disclosure of Invention

The present application mainly aims to provide an evaluation method, an evaluation device, an evaluation apparatus, and an evaluation medium for policy specialization based on artificial intelligence, which are used to solve the technical problems that the requirements for understanding the insurance professional level and the specialization specification are high due to the complicated and variable requirements of customers, ambiguity is easily generated when an order issuing person makes specialization, and the policy is conflicted with format descriptions in a policy template, so that the policy is invalid or has potential risks.

In order to achieve the above object, the present application proposes an evaluation method of policy agreement based on artificial intelligence, the method comprising:

acquiring a special agreement evaluation request corresponding to a target policy, wherein the special agreement evaluation request carries a policy special agreement to be evaluated;

performing word segmentation on the policy agreement to be evaluated to obtain a target policy agreement phrase set;

finding out phrases with grammar errors from the target policy special agreement phrase set to obtain a target grammar error phrase set;

evaluating semantic relevance according to the target policy special offer phrase set to obtain a target semantic relevance evaluation result;

performing risk assessment on the target policy special offer phrase set according to the acquired policy special offer semantic library to obtain a target risk assessment result;

and generating a special agreement evaluation result corresponding to the target policy according to the target grammar error phrase set, the target semantic correlation evaluation result and the target risk evaluation result.

Further, the step of obtaining a contract evaluation request of the target policy includes:

acquiring an increment special contract checking request corresponding to the target policy;

responding to the increment special contract checking request, and acquiring an edited special contract segment from the target policy to serve as the policy special contract to be evaluated;

and generating the agreement evaluation request according to the policy agreement to be evaluated and the policy identification corresponding to the target policy.

Further, before the step of performing risk assessment on the target policy agreement phrase set according to the obtained policy agreement semantic library to obtain a target risk assessment result, the method further includes:

acquiring a semantic library updating request, wherein the semantic library updating request carries a policy identifier set to be analyzed;

matching a policy information field set from the acquired policy database according to each policy identifier in the policy identifier set to be analyzed to obtain the policy information field set to be analyzed corresponding to each policy identifier;

respectively carrying out word segmentation and stop word deletion on the policy special contract fields in the policy information field set to be analyzed to obtain a policy special contract phrase set to be analyzed;

performing text weight calculation of each phrase according to the policy and special agreement phrase set to be analyzed to obtain a text weight set;

according to the complaint information field and the claim information field corresponding to the policy special offer phrase set to be analyzed, performing risk weight division calculation on each phrase in the policy special offer phrase set to be analyzed to obtain risk weight diversity;

and updating the policy agreement semantic library according to the policy information field set to be analyzed, the policy agreement phrase set to be analyzed, the text weight set and the risk weight diversity.

Further, the step of performing risk weight score calculation on each phrase in the policy agreement phrase set to be analyzed according to the complaint information field and the claim information field corresponding to the policy agreement phrase set to be analyzed to obtain risk weight diversity includes:

performing risk score evaluation on the policy according to the complaint information field and the claim settlement information field in the policy specialization phrase set to be analyzed to obtain a policy risk score;

performing word frequency calculation on each phrase in the policy agreement phrase set to be analyzed to obtain a word frequency set;

and according to the policy risk score and the word frequency set, performing risk weight score calculation on each phrase in the policy speciality phrase set to be analyzed to obtain the risk weight diversity.

Further, the step of performing risk assessment on the target policy agreement phrase set according to the obtained policy agreement semantic library to obtain a target risk assessment result includes:

similarity calculation is carried out on each special phrase in the policy special offer semantic library and each phrase in the target policy special offer phrase set, and a similarity set to be analyzed corresponding to each phrase in the target policy special offer phrase set is obtained;

acquiring a preset similarity threshold, and finding out the similarity greater than the preset similarity threshold from the similarity set to be analyzed to serve as a target similarity set;

acquiring each text weight corresponding to each similarity in the target similarity set from the policy special appointment semantic library to obtain a text weight set to be analyzed;

acquiring a preset text weight threshold, and finding out a text weight larger than the preset text weight threshold from the text weight set to be analyzed to serve as a target text weight set;

and acquiring complaint information fields and claim settlement information fields corresponding to each target text weight set from the policy speciality semantic library to obtain the target risk assessment result.

Further, after the step of generating the agreement evaluation result corresponding to the target policy according to the target grammar error phrase set, the target semantic relevance evaluation result and the target risk evaluation result, the method further includes:

acquiring a special agreement audit request of the target policy;

responding to the special agreement auditing request, and acquiring a policy special agreement from the target policy to serve as a policy special agreement to be audited;

performing word segmentation on the policy agreement to be audited to obtain a policy agreement phrase set to be audited;

finding out phrases with grammar errors from the policy agreement phrase set to be audited to obtain a grammar error phrase set to be processed;

according to the grammar error phrase set to be processed, carrying out problem special contract fragment proportion statistics on the policy special contract to be audited to obtain a target problem special contract fragment proportion;

according to the policy special agreement semantic library, performing risk audit on each phrase in the policy special agreement phrase set to be audited to obtain a risk audit result;

and generating a special agreement auditing result according to the grammar error phrase set to be processed, the target problem special agreement fragment proportion and the risk auditing result.

Further, the step of performing risk audit on each phrase in the policy and agreement phrase set to be audited according to the policy and agreement semantic library to obtain a risk audit result includes:

matching each risk weight score of each phrase in the policy special offer phrase set to be audited from the policy special offer semantic library to serve as risk weight diversity to be analyzed corresponding to each phrase in the policy special offer phrase set to be audited;

acquiring a preset risk weight division threshold value, and finding out a risk weight division larger than the preset risk weight division threshold value from the risk weight diversity to be analyzed to obtain a target risk weight diversity;

acquiring policy information field sets corresponding to the risk weight scores in the target risk weight score set from the policy special appointment semantic library to obtain policy information field sets to be displayed;

performing non-empty target risk weight diversity calculation according to each target risk weight diversity as a risk phrase number;

and taking the risk phrase number and each policy information field set to be displayed as the risk auditing result.

The present application further proposes an evaluation device of a policy agreement based on artificial intelligence, the device comprising:

the system comprises a request acquisition module, a policy evaluation module and a policy evaluation module, wherein the request acquisition module is used for acquiring a special evaluation request corresponding to a target policy, and the special evaluation request carries a policy special to be evaluated;

the word segmentation module is used for segmenting words of the policy speciality to be evaluated to obtain a target policy speciality phrase set;

a target grammar error phrase set determining module, which is used for finding out phrases with grammar errors from the target policy special agreement phrase set to obtain a target grammar error phrase set;

the target semantic correlation evaluation result determining module is used for evaluating semantic correlation according to the target policy special offer phrase set to obtain a target semantic correlation evaluation result;

the target risk assessment result determining module is used for performing risk assessment on the target policy special offer phrase set according to the acquired policy special offer semantic library to obtain a target risk assessment result;

and the evaluation result display module is used for generating a special offer evaluation result corresponding to the target policy according to the target grammar error phrase set, the target semantic relevance evaluation result and the target risk evaluation result.

The present application further proposes a computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the steps of any of the above methods when executing the computer program.

The present application also proposes a computer-readable storage medium having stored thereon a computer program which, when being executed by a processor, carries out the steps of the method of any of the above.

The method comprises the steps of firstly carrying out word segmentation on the policy agreement to be evaluated to obtain a target policy agreement phrase set, secondly finding out phrases with grammatical errors from the target policy agreement phrase set to obtain a target grammatical error phrase set, carrying out evaluation on semantic relevance according to the target policy agreement phrase set to obtain a target semantic relevance evaluation result, carrying out risk evaluation on the target policy agreement phrase set according to an obtained policy agreement semantic library to obtain a target risk evaluation result, and finally generating a special agreement evaluation result corresponding to the target policy according to the target grammatical error phrase set, the target semantic relevance evaluation result and the target risk evaluation result, so that automatic evaluation of grammatical error evaluation, evaluation and risk evaluation on a single operator are realized, Semantic relevance evaluation and risk evaluation reduce the requirements on the insurance professional level of the order-issuing member and the understanding of the special agreement specification, improve the accuracy of the special agreement of the insurance policy and avoid the invalid insurance policy or the potential risk; and moreover, manual inspection of the order-issuing personnel is avoided, and the efficiency of the order-issuing personnel is improved.

Drawings

FIG. 1 is a schematic flow chart illustrating a method for evaluating policy specialties based on artificial intelligence according to an embodiment of the present application;

FIG. 2 is a block diagram illustrating an exemplary configuration of an apparatus for evaluating an artificial intelligence-based policy agreement according to an embodiment of the present application;

fig. 3 is a block diagram illustrating a structure of a computer device according to an embodiment of the present application.

The implementation, functional features and advantages of the objectives of the present application will be further explained with reference to the accompanying drawings.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.

Referring to fig. 1, in an embodiment of the present application, a method for evaluating an artificial intelligence-based policy agreement is provided, the method including:

s1: acquiring a special agreement evaluation request corresponding to a target policy, wherein the special agreement evaluation request carries a policy special agreement to be evaluated;

s2: performing word segmentation on the policy agreement to be evaluated to obtain a target policy agreement phrase set;

s3: finding out phrases with grammar errors from the target policy special agreement phrase set to obtain a target grammar error phrase set;

s4: evaluating semantic relevance according to the target policy special offer phrase set to obtain a target semantic relevance evaluation result;

s5: performing risk assessment on the target policy special offer phrase set according to the acquired policy special offer semantic library to obtain a target risk assessment result;

s6: and generating a special agreement evaluation result corresponding to the target policy according to the target grammar error phrase set, the target semantic correlation evaluation result and the target risk evaluation result.

In this embodiment, a target policy agreement phrase set is obtained by performing word segmentation on the policy agreement to be evaluated, a phrase with a syntax error is found from the target policy agreement phrase set to obtain a target syntax error phrase set, semantic relevance evaluation is performed according to the target policy agreement phrase set to obtain a target semantic relevance evaluation result, risk evaluation is performed on the target policy agreement phrase set according to an obtained policy agreement semantic library to obtain a target risk evaluation result, and a special agreement evaluation result corresponding to the target policy is generated according to the target syntax error phrase set, the target semantic relevance evaluation result and the target risk evaluation result, so that syntax error evaluation, semantic relevance evaluation and risk evaluation are automatically performed on an order taker, and the insurance professional level, the grammar relevance and the risk evaluation of the order taker are reduced, The requirement of the specification of the appointment for understanding improves the accuracy of the policy appointment and avoids the invalidity or the potential risk of the policy; and moreover, manual inspection of the order-issuing personnel is avoided, and the efficiency of the order-issuing personnel is improved.

For S1, a contract evaluation request corresponding to the target policy input by the user (e.g., the order taker) may be obtained, a contract evaluation request corresponding to the target policy input by the third-party application system may also be obtained, or a contract evaluation request corresponding to the target policy actively triggered by the program implementing the present application may also be obtained. For example, when an order entry symbol is issued, a special offer evaluation request corresponding to a target policy actively triggered by the program of the present application is implemented, which is not specifically limited by the examples herein.

Target warranty, i.e., the warranty that requires the appointment evaluation. Policy, generally referred to as an insurance policy, is a written proof that the insurer has entered into an insurance contract with the applicant.

A contract evaluation request, i.e., a request to evaluate a contract in a target policy.

The policy agreement to be evaluated, i.e., the agreement for which the agreement evaluation request is intended to evaluate. It is understood that the policy agreement to be evaluated may be all of the agreements in the target policy or some of the agreements in the target policy, and is not limited herein.

A policy convention, also known as an insurance special agreement, is a special agreement or term in a policy that is agreed upon by the applicant and the insurance company to be attached to a policy.

For S2, a Principal Component Analysis (PCA) method is adopted to perform word segmentation on the policy agreement to be evaluated, and each phrase obtained by word segmentation is used as a target policy agreement phrase set.

Optionally, a word segmentation dictionary is adopted to segment the policy agreement to be evaluated, and each phrase obtained by word segmentation is used as a target policy agreement phrase set.

The word segmentation dictionary comprises one or more phrases.

And S3, acquiring a grammar error detection model, inputting the target policy-preserving special phrase set into the grammar error detection model for phrase searching of grammar errors, and taking each phrase output by the grammar error detection model as a target grammar error phrase set.

The grammar error detection model is a model obtained based on a dependency tree algorithm and machine learning training. And the grammar error detection model is used for carrying out long-distance grammar error detection.

And S4, acquiring a semantic correlation scoring model, inputting the target policy special terms set into a semantic correlation detection model to evaluate the semantic correlation in the target policy special terms set, and taking evaluation data output by the semantic correlation scoring model as a target semantic correlation evaluation result. That is, the target semantic relevance assessment result is an assessment result of semantic relevance within the policy agreement to be assessed, and the target semantic relevance assessment result is an assessment result of semantic relevance between individual terms in the target policy agreement term set.

The semantic relevance scoring model is a model obtained based on initial model training. The initial model includes convolutional layers, pooling layers, fully-connected layers, and the like.

For S5, the policy specialization semantic library may be obtained from a database, or may be obtained from a third party application system.

The policy speciality semantic library comprises: the special phrase, the policy identification, the text weight, the risk weight score, the policy risk score and the policy information field set, wherein the policy information field set comprises: the policy system includes a policy base information field, a policy speciality field, a complaint information field and a claim information field, the speciality phrase is a phrase in the policy speciality field, the text weight is a text weight of the speciality phrase in the policy speciality field, and the risk weight score is a score occupied by the speciality phrase in the policy risk score. The policy identification may be a policy name, policy ID, or other data that uniquely identifies a policy. The policy risk score is the risk score of the policy calculated from the complaint information field and the claim information field. The insurance policy basic information field is basic information of the insurance policy, such as insurance start and end period, insurance target, and is not specifically limited by the examples herein. The policy agreement field is the agreement of the policy. The complaint information field is complaint information of the policy, such as a complaint category, a complaint subject, and a complaint frequency, and is not specifically limited by the example. The claim information field is the claim information of the policy, such as the type of the claim, the amount of the claim, and the number of the claims, which is not limited in this embodiment.

And matching each phrase in the target special policy phrase set into a special phrase with the similarity meeting the similarity requirement in a special policy database of the policy, wherein the text weight corresponding to the matched special phrase meets the text weight requirement, and the complaint information field and the claim information field corresponding to the special phrase meeting the similarity requirement and meeting the text weight are used as target risk assessment results.

And S6, generating a specification by adopting a preset special offer evaluation result, and generating a special offer evaluation result corresponding to the target policy according to the target grammar error phrase set, the target semantic relevance evaluation result and the target risk evaluation result.

And displaying the special assessment result in a preset display mode so as to carry out grammar check, risk retrieval and semantic relevance error reminding on the order-issuing person of the target insurance policy.

The preset display modes include but are not limited to: and displaying the bullet frame and marking the bullet frame in the target insurance policy by adopting a preset format.

Optionally, the present application further includes: acquiring an association request corresponding to the target policy, wherein the association request carries phrases to be associated; matching special phrases which simultaneously accord with the similarity requirement and accord with the text weight in a policy special offer semantic library according to the phrases to be associated, and taking policy special terms corresponding to the matched special phrases as association results corresponding to the phrases to be associated. Therefore, the requirements for understanding the insurance professional level and the special rule specification of the order-issuing person are further reduced, and the efficiency of the order-issuing person is further improved.

In an embodiment, the step of obtaining a contract evaluation request of the target policy includes:

s11: acquiring an increment special contract checking request corresponding to the target policy;

s12: responding to the increment special contract checking request, and acquiring an edited special contract segment from the target policy to serve as the policy special contract to be evaluated;

s13: and generating the agreement evaluation request according to the policy agreement to be evaluated and the policy identification corresponding to the target policy.

The embodiment realizes generation of the special offer evaluation request according to the special offer segment with editing, thereby improving the efficiency of responding to the special offer evaluation request, being beneficial to assisting the order issuing personnel to quickly correct the error of the policy special offer of the target policy, improving the accuracy of the policy special offer and improving the efficiency of the order issuing personnel.

For S11, the increment and special offer check request corresponding to the target policy input by the user may be obtained, the increment and special offer check request corresponding to the target policy may also be obtained from a third-party application system, or the increment and special offer check request corresponding to the target policy actively triggered by the program of the present application may also be implemented. For example, when the order output member edits the target policy, the order output member inputs a punctuation mark to actively trigger the request for checking the incremental property, which is not specifically limited in this example. For another example, when the order output member edits the target policy, the pause duration input by the order output member exceeds the preset pause duration, and the increment special contract checking request is actively triggered, which is not specifically limited in this example.

The method comprises the steps of receiving a request for incremental contract check, namely a request for contract evaluation on a contract fragment with editing.

For S12, in response to the incremental treaty check request, treating text between two punctuations in the target policy as a treaty segment, text before the first punctuation in the target policy as a treaty segment, and text after the last punctuation in the target policy as a treaty segment; and judging whether each special section of the target policy has edition or not according to the policy special offer corresponding to the latest time special offer evaluation request of the target policy, and taking each special section with edition as the policy special offer to be evaluated according to the judgment result. That is, each appointment fragment of the policy agreement to be evaluated is an appointment fragment for which there is an edit.

Punctuation marks are written symbols used to indicate sentence reading and tone.

For S13, when the agreement evaluation request is generated according to the policy agreement to be evaluated and the policy identifier corresponding to the target policy, the policy agreement to be evaluated and the policy identifier corresponding to the target policy are used as parameters of the agreement evaluation request.

In an embodiment, before the step of performing risk assessment on the target policy agreement phrase set according to the obtained policy agreement semantic library to obtain a target risk assessment result, the method further includes:

s511: acquiring a semantic library updating request, wherein the semantic library updating request carries a policy identifier set to be analyzed;

s512: matching a policy information field set from the acquired policy database according to each policy identifier in the policy identifier set to be analyzed to obtain the policy information field set to be analyzed corresponding to each policy identifier;

s513: respectively carrying out word segmentation and stop word deletion on the policy special contract fields in the policy information field set to be analyzed to obtain a policy special contract phrase set to be analyzed;

s514: performing text weight calculation of each phrase according to the policy and special agreement phrase set to be analyzed to obtain a text weight set;

s515: according to the complaint information field and the claim information field corresponding to the policy special offer phrase set to be analyzed, performing risk weight division calculation on each phrase in the policy special offer phrase set to be analyzed to obtain risk weight diversity;

s516: and updating the policy agreement semantic library according to the policy information field set to be analyzed, the policy agreement phrase set to be analyzed, the text weight set and the risk weight diversity.

In the embodiment, the policy agreement semantic library is updated before the acquired policy agreement semantic library, so that a basis is provided for risk assessment based on the policy agreement semantic library.

For S511, the semantic library update request input by the user may be acquired, or the semantic library update request may be acquired from a third-party application system.

Semantic library update requests, i.e., requests to update the policy agreement semantic library.

The set of policy identifications to be analyzed, that is, the set of policy identifications from which the semantic vault update request is intended to update policy agreement semantic data to the policy agreement semantic vault. The set of policy identifications to be analyzed includes one or more policy identifications.

For S512, the policy database includes: a policy identification and a set of policy information fields.

And searching each policy identifier in the policy identifier set to be analyzed in each policy identifier of the policy database, and taking the policy information field set corresponding to each policy identifier searched in the policy database as a policy information field set to be analyzed.

For step S513, performing word segmentation on the policy special offer fields in the policy information field set to be analyzed, and taking each phrase obtained by word segmentation as a policy special offer phrase set of stop words to be deleted; and acquiring a stop word set, deleting each stop word in the stop word set from the policy special phrase set of the stop word to be deleted, and taking the policy special phrase set of the stop word to be deleted after the deletion processing as the policy special phrase set to be analyzed.

The stop word set includes one or more stop words. Stop Words refer to that in information retrieval, in order to save storage space and improve search efficiency, some characters or Words are automatically filtered before or after processing natural language data (or text), and the characters or Words are called Stop Words.

Optionally, a principal component analysis method is adopted to perform word segmentation on the policy agreement fields in the policy information field set to be analyzed, and each phrase obtained by word segmentation is used as the policy agreement phrase set to be deduplicated.

Optionally, a word segmentation dictionary is adopted to perform word segmentation on the policy and speciality fields in the policy information field set to be analyzed, and each phrase obtained by word segmentation is used as the policy and speciality phrase set to be deduplicated.

And S514, performing text weight calculation of each phrase according to the policy-keeping special offer phrase set to be analyzed by adopting a document frequency method, and taking each calculated text weight as a text weight set.

That is, the text weight is the text weight of the phrase in one policy agreement field.

And S515, according to the complaint information fields and the settlement information fields corresponding to the policy special offer phrase sets to be analyzed, performing risk score evaluation on the policy corresponding to the policy special offer phrase sets to be analyzed, then performing risk weight score calculation of each phrase according to the policy special offer phrase sets to be analyzed and the risk scores, and taking each calculated risk weight score as a risk weight diversity.

For S516, the policy information field set to be analyzed, the policy agreement phrase set to be analyzed, the text weight set, and the risk weight diversity corresponding to the same policy identifier are used as association data, and the association data is updated to the policy agreement semantic library.

In an embodiment, the step of performing risk weight score calculation on each phrase in the policy agreement phrase set to be analyzed according to the complaint information field and the claim information field corresponding to the policy agreement phrase set to be analyzed to obtain the risk weight diversity includes:

s5151: performing risk score evaluation on the policy according to the complaint information field and the claim settlement information field in the policy specialization phrase set to be analyzed to obtain a policy risk score;

s5152: performing word frequency calculation on each phrase in the policy agreement phrase set to be analyzed to obtain a word frequency set;

s5153: and according to the policy risk score and the word frequency set, performing risk weight score calculation on each phrase in the policy speciality phrase set to be analyzed to obtain the risk weight diversity.

In this embodiment, risk score evaluation is performed on the policy corresponding to the policy special offer phrase set to be analyzed according to the complaint information field and the claim information field corresponding to the policy special offer phrase set to be analyzed, and then risk weight score calculation is performed on each phrase according to the policy special offer phrase set to be analyzed and the risk score, so that risk weight score of each phrase is obtained, and a basis is provided for risk evaluation based on the policy special offer semantic library.

And S5151, performing risk score evaluation on the policy according to the complaint information field and the claim settlement information field in the policy speciality phrase set to be analyzed by adopting a preset risk score evaluation model, and taking the score obtained by evaluation as the policy risk score.

The preset risk score evaluation model may be a model for classification prediction, and each classification label is a risk score range. It is understood that the start value of the risk score range may be used as the risk score corresponding to the classification label, the end value of the risk score range may be used as the risk score corresponding to the classification label, and the average value of the risk score range may be used as the risk score corresponding to the classification label.

It is understood that the preset risk score evaluation model may also be other models, and is not limited herein.

For S5152, performing word frequency calculation in the policy agreement phrase set to be analyzed on each phrase in the policy agreement phrase set to be analyzed, and taking each calculated word frequency as a word frequency set.

Word frequency, a commonly used weighting technique for intelligence retrieval and text mining, is used to evaluate the degree of repetition of a word for a set of domain documents in a document or corpus.

For S5153, performing multiplication calculation on the word frequency corresponding to the target phrase in the word frequency set and the policy-preserving risk score, and taking the data obtained by the multiplication calculation as the risk weight corresponding to the target phrase, where the target phrase is any phrase in the policy-preserving special phrase set to be analyzed.

In an embodiment, the step of performing risk assessment on the target policy agreement phrase set according to the obtained policy agreement semantic library to obtain a target risk assessment result includes:

s521: similarity calculation is carried out on each special phrase in the policy special offer semantic library and each phrase in the target policy special offer phrase set, and a similarity set to be analyzed corresponding to each phrase in the target policy special offer phrase set is obtained;

s522: acquiring a preset similarity threshold, and finding out the similarity greater than the preset similarity threshold from the similarity set to be analyzed to serve as a target similarity set;

s523: acquiring each text weight corresponding to each similarity in the target similarity set from the policy special appointment semantic library to obtain a text weight set to be analyzed;

s524: acquiring a preset text weight threshold, and finding out a text weight larger than the preset text weight threshold from the text weight set to be analyzed to serve as a target text weight set;

s525: and acquiring complaint information fields and claim settlement information fields corresponding to each target text weight set from the policy speciality semantic library to obtain the target risk assessment result.

In the embodiment, the special phrases with the similarity meeting the similarity requirement are matched in the special policy semantic library according to each phrase in the target special policy phrase set, the text weight corresponding to the matched special phrases meets the text weight requirement, and the complaint information fields and the claim information fields corresponding to the special phrases with the similarity meeting the text weight meeting the similarity requirement are used as the target risk assessment result, so that the automatic risk assessment of the order issuing personnel is realized, the requirements on understanding the insurance professional level and the special specification of the order issuing personnel are reduced, the accuracy of the special policy is improved, and the invalid policy or the potential risk is avoided.

For S521, a KMP (Knuth-Morria-Pratt) algorithm is adopted to perform similarity calculation between each special phrase in the policy agreement semantic library and each phrase in the target policy agreement phrase set, and each similarity corresponding to one phrase in the target policy agreement phrase set is used as a similarity set to be analyzed. That is, the number of similarity sets to be analyzed is the same as the number of phrases of the target policy agreement phrase set.

It is understood that other algorithms may be used to calculate the similarity between each offer phrase in the policy agreement semantic library and each phrase in the target policy agreement phrase set, and are not limited herein.

For S522, the preset similarity threshold may be obtained from the database, or may be obtained from a third-party application system, or may be written in a program implementing the present application. The preset similarity threshold is a specific numerical value.

And finding out the similarity greater than the preset similarity threshold from the similarity set to be analyzed, and taking each found similarity as a target similarity set. That is, the similarities in the target similarity set are all greater than the preset similarity threshold.

For step S523, find out the special phrase corresponding to each similarity in the target similarity set, and use each text weight corresponding to each found special phrase in the policy special semantic library as a text weight set to be analyzed. And finding out a text weight set to be analyzed corresponding to the special phrase with the similarity meeting the similarity requirement (namely, being greater than the preset similarity threshold).

For S524, the preset text weight threshold may be obtained from the database, may also be obtained from a third-party application system, and may also be written in a program implementing the present application. The preset text weight threshold is a specific numerical value.

And finding out the text weights which are larger than the preset text weight threshold value from the text weight set to be analyzed, and taking the found text weights which are larger than the preset text weight threshold value as a target text weight set. That is, the text weight in the target text weight set is greater than the preset text weight threshold. And finding out a target text weight set corresponding to the special phrases with the similarity meeting the similarity requirement (namely, greater than the preset similarity threshold) and the text weight meeting the text weight requirement (namely, greater than the preset text weight threshold).

For step S525, complaint information fields and claim information fields corresponding to the respective text weights in each target text weight set are obtained from the policy speciality semantic library, and the obtained complaint information fields and the obtained claim information fields are used as the target risk assessment result.

In an embodiment, after the step of generating the agreement evaluation result corresponding to the target policy according to the target grammar error phrase set, the target semantic relevance evaluation result, and the target risk evaluation result, the method further includes:

s71: acquiring a special agreement audit request of the target policy;

s72: responding to the special agreement auditing request, and acquiring a policy special agreement from the target policy to serve as a policy special agreement to be audited;

s73: performing word segmentation on the policy agreement to be audited to obtain a policy agreement phrase set to be audited;

s74: finding out phrases with grammar errors from the policy agreement phrase set to be audited to obtain a grammar error phrase set to be processed;

s75: according to the grammar error phrase set to be processed, carrying out problem special contract fragment proportion statistics on the policy special contract to be audited to obtain a target problem special contract fragment proportion;

s76: according to the policy special agreement semantic library, performing risk audit on each phrase in the policy special agreement phrase set to be audited to obtain a risk audit result;

s77: and generating a special agreement auditing result according to the grammar error phrase set to be processed, the target problem special agreement fragment proportion and the risk auditing result.

According to the embodiment, the automatic reminding of syntax error audit, semantic correlation audit and risk audit of the auditor is realized, the requirements on the insurance professional level and the special agreement specification of the auditor are reduced, the auditing accuracy of the auditor is improved, the accuracy of the insurance policy special agreement is further improved, and the insurance policy invalidation or the potential risk is avoided; and only the auditor is required to check the special agreement audit result, so that the efficiency of the auditor is improved.

For S71, a contract auditing request of the target policy sent by an individual may be obtained, or a contract auditing request of the target policy may be obtained from a third-party application system.

A contract audit request, i.e., a request to audit a contract in the target policy.

For S72, in response to the agreement audit request, obtaining a policy agreement from the target policy, and taking the obtained policy agreement as the policy agreement to be audited.

The policy specials to be reviewed are all specials in the target policy.

And S73, performing word segmentation on the policy agreement to be audited by adopting a principal component analysis method, and taking each phrase obtained by word segmentation as a policy agreement phrase set to be audited.

Optionally, a word segmentation dictionary is adopted to segment the policy agreement to be audited, and the policy agreement phrase to be audited is collected.

And S74, inputting the warranty special agreement phrase set to be audited into the grammar error detection model to search grammatical error phrases, and taking each phrase output by the grammar error detection model as a grammar error phrase set to be processed.

For S75, according to the grammar error phrase set to be processed, searching the special contract segment with problems for the policy special contract to be audited; dividing the number of the searched special about segments with problems by the total number of the special about segments in the policy to be audited, and taking the data obtained by the division as the proportion of the special about segments with the target problems.

For S76, matching each phrase in the policy and special offer phrase set to be audited in the policy and special offer semantic library to obtain special phrases whose similarity meets the similarity requirement, and the text weight corresponding to the matched special phrases meets the text weight requirement, and using the policy information field set and the special phrase number corresponding to the special phrases meeting the similarity requirement and the text weight as the risk audit result.

And S77, generating a specification by adopting a preset special agreement audit result, and generating a special agreement audit result according to the grammar error phrase set to be processed, the target problem special agreement fragment proportion and the risk audit result.

In an embodiment, the step of performing risk review on each phrase in the policy agreement phrase set to be reviewed according to the policy agreement semantic library to obtain a risk review result includes:

s761: matching each risk weight score of each phrase in the policy special offer phrase set to be audited from the policy special offer semantic library to serve as risk weight diversity to be analyzed corresponding to each phrase in the policy special offer phrase set to be audited;

s762: acquiring a preset risk weight division threshold value, and finding out a risk weight division larger than the preset risk weight division threshold value from the risk weight diversity to be analyzed to obtain a target risk weight diversity;

s763: acquiring policy information field sets corresponding to the risk weight scores in the target risk weight score set from the policy special appointment semantic library to obtain policy information field sets to be displayed;

s764: performing non-empty target risk weight diversity calculation according to each target risk weight diversity as a risk phrase number;

s765: and taking the risk phrase number and each policy information field set to be displayed as the risk auditing result.

In this embodiment, each phrase in the policy special offer set to be audited is matched into a special phrase whose similarity meets the requirement of similarity in the policy special offer semantic library, and the text weight corresponding to the matched special phrase meets the requirement of text weight, and the policy information field set and the number of special phrases corresponding to the special phrases meeting the requirement of similarity and meeting the requirement of text weight are used as a risk audit result, so that the automated reminding of risk audit of auditors is realized, the requirements of understanding the insurance professional level and special agreement specifications of the auditors are reduced, the audit accuracy of the auditors is improved, the accuracy of policy special offers is further improved, and the condition that the policy is invalid or has potential risk is avoided.

For S761, matching each risk weight score of each phrase in the policy agreement phrase set to be examined from the policy agreement semantic library, and taking each risk weight score matched to each phrase as a risk weight diversity to be analyzed. That is, the number of risk weight diversities to be analyzed is the same as the number of phrases in the policy agreement phrase set to be reviewed.

For step S762, the preset risk weight score threshold may be obtained from the database, or may be obtained from the third-party application system, or may be written in the program implementing the present application. The preset risk weight threshold is a specific value.

And finding out risk weight distribution larger than the preset risk weight distribution threshold value from the risk weight diversity to be analyzed, and taking each found risk weight distribution larger than the preset risk weight distribution threshold value as target risk weight diversity.

And S763, acquiring policy information field sets corresponding to the risk weight scores in the target risk weight score set from the policy speciality semantic library, and taking the acquired policy information field sets as policy information field sets to be displayed.

For S764, performing a target risk weight diversity calculation that is not empty according to each target risk weight diversity, thereby determining a number of phrases in the policy agreement phrase set to be reviewed for which the existing risk weights are greater than the preset risk weight division threshold.

For S765, directly taking the risk phrase number and each policy information field set to be presented as the risk review result.

Referring to fig. 2, the present application also proposes an artificial intelligence-based policy agreement evaluation apparatus, the apparatus comprising:

a request obtaining module 100, configured to obtain a special offer evaluation request corresponding to a target policy, where the special offer evaluation request carries a policy special offer to be evaluated;

a word segmentation module 200, configured to perform word segmentation on the policy agreement to be evaluated, so as to obtain a target policy agreement phrase set;

a target grammatical error phrase set determination module 300, configured to find phrases with grammatical errors from the target policy agreement phrase set, so as to obtain a target grammatical error phrase set;

a target semantic relevance evaluation result determining module 400, configured to perform semantic relevance evaluation according to the target policy special offer phrase set to obtain a target semantic relevance evaluation result;

a target risk assessment result determining module 500, configured to perform risk assessment on the target policy special offer phrase set according to the obtained policy special offer semantic library, to obtain a target risk assessment result;

and an evaluation result display module 600, configured to generate a special agreement evaluation result corresponding to the target policy according to the target grammar error phrase set, the target semantic relevance evaluation result, and the target risk evaluation result.

Referring to fig. 3, a computer device, which may be a server and whose internal structure may be as shown in fig. 3, is also provided in the embodiment of the present application. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the computer designed processor is used to provide computational and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used for storing data such as an evaluation method of the policy agreement based on artificial intelligence. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a method for artificial intelligence based policy agreement assessment. The evaluation method of the policy agreement based on artificial intelligence comprises the following steps: acquiring a special agreement evaluation request corresponding to a target policy, wherein the special agreement evaluation request carries a policy special agreement to be evaluated; performing word segmentation on the policy agreement to be evaluated to obtain a target policy agreement phrase set; finding out phrases with grammar errors from the target policy special agreement phrase set to obtain a target grammar error phrase set; evaluating semantic relevance according to the target policy special offer phrase set to obtain a target semantic relevance evaluation result; performing risk assessment on the target policy special offer phrase set according to the acquired policy special offer semantic library to obtain a target risk assessment result; and generating a special agreement evaluation result corresponding to the target policy according to the target grammar error phrase set, the target semantic correlation evaluation result and the target risk evaluation result.

An embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon, the computer program, when executed by a processor, implementing a method for artificial intelligence-based policy agreement assessment, comprising the steps of: acquiring a special agreement evaluation request corresponding to a target policy, wherein the special agreement evaluation request carries a policy special agreement to be evaluated; performing word segmentation on the policy agreement to be evaluated to obtain a target policy agreement phrase set; finding out phrases with grammar errors from the target policy special agreement phrase set to obtain a target grammar error phrase set; evaluating semantic relevance according to the target policy special offer phrase set to obtain a target semantic relevance evaluation result; performing risk assessment on the target policy special offer phrase set according to the acquired policy special offer semantic library to obtain a target risk assessment result; and generating a special agreement evaluation result corresponding to the target policy according to the target grammar error phrase set, the target semantic correlation evaluation result and the target risk evaluation result.

The executed evaluation method of the policy special agreement based on artificial intelligence obtains a target policy special agreement phrase set by segmenting the policy special agreement to be evaluated, finds out phrases with grammar errors from the target policy special agreement phrase set to obtain a target grammar error phrase set, evaluates semantic relevance according to the target policy special agreement phrase set to obtain a target semantic relevance evaluation result, evaluates risk of the target policy special agreement phrase set according to the obtained policy special agreement semantic library to obtain a target risk evaluation result, and generates a special agreement evaluation result corresponding to the target policy according to the target grammar error phrase set, the target semantic relevance evaluation result and the target risk evaluation result, thereby realizing automatic evaluation of grammar error evaluation, semantic relevance evaluation and risk evaluation of order-takers, the requirement for understanding the insurance professional level and the special agreement specification of the order-issuing member is reduced, the accuracy of the special agreement of the insurance policy is improved, and the invalid insurance policy or potential risk is avoided; and moreover, manual inspection of the order-issuing personnel is avoided, and the efficiency of the order-issuing personnel is improved.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium provided herein and used in the examples may include non-volatile and/or volatile memory. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), double-rate SDRAM (SSRSDRAM), Enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and bus dynamic RAM (RDRAM).

It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, apparatus, article, or method 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, apparatus, article, or method. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, apparatus, article, or method that includes the element.

The above description is only a preferred embodiment of the present application, and not intended to limit the scope of the present application, and all modifications of equivalent structures and equivalent processes, which are made by the contents of the specification and the drawings of the present application, or which are directly or indirectly applied to other related technical fields, are also included in the scope of the present application.

Claims

1. A method for artificial intelligence based assessment of policy specialties, the method comprising:

2. The artificial intelligence based policy agreement assessment method according to claim 1 wherein said step of obtaining a policy assessment request for a target policy comprises:

3. The method according to claim 1, wherein before the step of performing risk assessment on the target policy agreement phrase set according to the obtained policy agreement semantic library to obtain a target risk assessment result, the method further comprises:

4. The method according to claim 3, wherein the step of performing risk weight score calculation on each phrase in the policy agreement phrase set to be analyzed according to the complaint information fields and the claim information fields corresponding to the policy agreement phrase set to be analyzed to obtain risk weight diversity comprises:

5. The method according to claim 1, wherein the step of performing risk assessment on the target policy agreement phrase set according to the obtained policy agreement semantic library to obtain a target risk assessment result comprises:

6. The artificial intelligence based policy agreement assessment method according to claim 1, wherein said step of generating a agreement assessment result corresponding to said target policy based on said target set of grammatical error phrases, said target semantic relevance assessment result and said target risk assessment result further comprises, after said step of:

acquiring a special agreement audit request of the target policy;

7. The method according to claim 6, wherein the step of performing a risk review on each phrase in the set of terms of the policy agreement to be reviewed according to the semantic library of the policy agreement to obtain a risk review result comprises:

8. An apparatus for artificial intelligence based assessment of policy specialties, the apparatus comprising:

9. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor implements the steps of the method of any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.