CN113781248A - Insurance claim settlement processing method, device and system - Google Patents

Insurance claim settlement processing method, device and system Download PDF

Info

Publication number
CN113781248A
CN113781248A CN202111134366.4A CN202111134366A CN113781248A CN 113781248 A CN113781248 A CN 113781248A CN 202111134366 A CN202111134366 A CN 202111134366A CN 113781248 A CN113781248 A CN 113781248A
Authority
CN
China
Prior art keywords
training
insurance
verification
reimbursement
information
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.)
Granted
Application number
CN202111134366.4A
Other languages
Chinese (zh)
Other versions
CN113781248B (en
Inventor
祁婷
邝智颖
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Bank of China Ltd
Original Assignee
Bank of China Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Bank of China Ltd filed Critical Bank of China Ltd
Priority to CN202111134366.4A priority Critical patent/CN113781248B/en
Publication of CN113781248A publication Critical patent/CN113781248A/en
Application granted granted Critical
Publication of CN113781248B publication Critical patent/CN113781248B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • 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
    • 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
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
    • G06F18/24147—Distances to closest patterns, e.g. nearest neighbour classification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/24323—Tree-organised classifiers
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • 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
    • G06Q20/00—Payment architectures, schemes or protocols
    • G06Q20/38—Payment protocols; Details thereof
    • G06Q20/382—Payment protocols; Details thereof insuring higher security of transaction
    • G06Q20/3829—Payment protocols; Details thereof insuring higher security of transaction involving key management
    • 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
    • G06Q20/00—Payment architectures, schemes or protocols
    • G06Q20/38—Payment protocols; Details thereof
    • G06Q20/40—Authorisation, e.g. identification of payer or payee, verification of customer or shop credentials; Review and approval of payers, e.g. check credit lines or negative lists
    • G06Q20/401—Transaction verification
    • G06Q20/4014—Identity check for transactions

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Accounting & Taxation (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • Strategic Management (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Finance (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • General Business, Economics & Management (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Computer Security & Cryptography (AREA)
  • Human Resources & Organizations (AREA)
  • Development Economics (AREA)
  • Software Systems (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Game Theory and Decision Science (AREA)
  • Medical Informatics (AREA)
  • Technology Law (AREA)
  • Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)

Abstract

The invention provides an insurance claim settlement processing method, device and system, which can be applied to the field of block chains or the field of finance. The method comprises the steps that the validity of an reimbursement certificate is verified firstly under the condition that an insurance claim settlement request is received, a data validity prediction model is called to predict the validity of information to be audited only under the condition that the reimbursement certificate is valid, the waste of computing resources caused by the fact that the data validity prediction model is called to predict all information to be audited is avoided, the information to be audited is packaged and sent to a block chain under the condition that the information to be audited is valid, the insurance claim settlement is completed under the condition that the information to be audited triggers a preset insurance claim settlement intelligent contract, invalid data chain linking is avoided, the whole insurance claim settlement process is completed automatically, the insurance claim settlement process is simplified, the insurance claim settlement period is shortened, and user experience is effectively improved.

Description

Insurance claim settlement processing method, device and system
Technical Field
The invention relates to the technical field of computers, in particular to an insurance claim settlement processing method, device and system.
Background
In recent years, along with the improvement of the living standard of the people, the insurance concept of people is gradually changed, more and more people pay attention to or buy insurance, so that a guarantee is provided for the future of the people and family, and the strong development of the insurance industry of China is further driven.
Although the variety of insurance in the market is numerous at present, because the insurance claim settlement is processed manually during the insurance reimbursement, the process is complex, the period required by the claim settlement is long, the user experience is seriously influenced, and thus many people still have a disadvantage of disappointing the insurance.
Disclosure of Invention
In view of this, the invention provides an insurance claim settlement processing method, device and system, and the whole insurance claim settlement process is completed automatically, so that the insurance claim settlement process is simplified, the insurance claim settlement period is shortened, and the user experience is effectively improved.
In order to achieve the above purpose, the invention provides the following specific technical scheme:
an insurance claim processing method applied to an insurance claim server, the method comprising:
under the condition that an insurance claim settlement request is received, analyzing the insurance claim settlement request to obtain information to be audited of a user, wherein the information to be audited at least comprises reimbursement vouchers;
verifying the validity of the reimbursement credential;
calling a pre-constructed data effectiveness prediction model to perform effectiveness prediction on the information to be checked under the condition that the reimbursement voucher passes verification, wherein the data effectiveness prediction model is obtained by training a multi-layer machine learning model by using historical reimbursement record data, and training algorithms corresponding to each layer of the multi-layer machine learning model are different;
and packaging and sending the information to be audited to a block chain platform under the condition that the information to be audited is effective, and finishing insurance claim settlement under the condition that the information to be audited triggers a preset insurance claim settlement intelligent contract.
Optionally, the reimbursement certificate is a verifiable certificate which is issued by a user requesting a preset organization by using a distributed Digital Identity (DID), and the verifiable certificate is generated by the preset organization encrypting reimbursement certificate data of the user by using a private key;
verifying the validity of the reimbursement credential, comprising:
and verifying the reimbursement certificate by using the public key of the preset mechanism, and determining whether the reimbursement certificate is valid according to a verification result.
Optionally, constructing the data validity prediction model includes:
preprocessing historical reimbursement record data to obtain a sample data set;
dividing the sample data set into a training set and a test set;
respectively training and testing N training algorithms by utilizing the training set and the test set to obtain the accuracy of each training algorithm, wherein N is an odd number;
the accuracy of each training algorithm is sorted from high to low, the first one
Figure BDA0003281487060000021
Taking the training algorithm as a first training algorithm
Figure BDA0003281487060000022
Using the training algorithm as a second training algorithm, and using the last training algorithm as a third training algorithm;
in a first layer of the model, dividing the training set into K groups, taking K-1 groups as a sub-training set, taking the remaining group as a verification set, respectively performing K-fold cross verification on each first training algorithm, calculating the average value of the prediction result of each first training algorithm for each sample in the verification set to obtain a verification set P 'corresponding to the verification result, and calculating the average value of the prediction result of each first training algorithm for each sample in the test set to obtain a test set T' corresponding to the test result;
in a second layer of the model, dividing a verification set P 'of the first layer of the model into K groups, taking K-1 groups as a sub-training set, taking the rest groups as the verification sets, respectively carrying out K-fold cross verification on each second training algorithm, calculating the average value of the prediction result of each second training algorithm aiming at each sample in the verification sets to obtain a verification set P' corresponding to the verification result, and calculating the average value of the prediction result of each second training algorithm aiming at each sample in a test set T 'of the first layer of the model to obtain a test set T' corresponding to the test result;
in the third layer of the model, dividing the verification set P 'of the second layer of the model into K groups, taking K-1 groups as a sub-training set, taking the rest group as a verification set, carrying out K-fold cross verification on the third training algorithm, testing the third training algorithm by using the test set T' of the second layer of the model, and obtaining the data validity prediction model consisting of the first layer, the second layer and the third layer under the condition that the verification result and the test result meet the preset conditions.
Optionally, the preprocessing the historical reimbursement record data to obtain a sample data set includes:
respectively extracting preset characteristic indexes of each piece of historical reimbursement record data;
respectively carrying out Cartesian product feature combination on preset feature indexes of each piece of historical reimbursement record data, and marking whether each piece of historical reimbursement record data is valid or not to obtain a plurality of sample data;
and removing abnormal values in a plurality of sample data by using a local abnormal factor LOF algorithm to obtain the sample data set.
Optionally, the method further includes:
and prompting the user that the insurance claim is not passed under the condition that the reimbursement voucher is not verified, the information to be audited is invalid or the information to be audited does not trigger the intelligent insurance claim contract.
Optionally, after the insurance claim is completed, the method further includes:
and writing the reimbursement record data into a supervision report, and sending the supervision report to a supervision mechanism system.
An insurance claim processing apparatus applied to an insurance claim server, the apparatus comprising:
the system comprises a claim settlement request analyzing unit, a claim settlement request analyzing unit and a verification processing unit, wherein the claim settlement request analyzing unit is used for analyzing an insurance claim settlement request under the condition that the insurance claim settlement request is received to obtain information to be verified of a user, and the information to be verified at least comprises reimbursement vouchers;
the reimbursement certificate verification unit is used for verifying the validity of the reimbursement certificate;
the data effectiveness prediction unit is used for calling a pre-constructed data effectiveness prediction model to perform effectiveness prediction on the information to be checked under the condition that the reimbursement voucher passes verification, wherein the data effectiveness prediction model is obtained by training a multi-layer machine learning model by using historical reimbursement record data, and training algorithms corresponding to each layer of the multi-layer machine learning model are different;
and the data uplink unit is used for packaging and sending the information to be audited to a block chain platform under the condition that the information to be audited is effective, and finishing insurance claim settlement under the condition that the information to be audited triggers a preset intelligent insurance claim contract.
Optionally, the reimbursement certificate is a verifiable certificate which is issued by a user requesting a preset organization by using a distributed Digital Identity (DID), and the verifiable certificate is generated by the preset organization encrypting reimbursement certificate data of the user by using a private key;
the reimbursement certificate verification unit is specifically configured to verify the reimbursement certificate by using the public key of the preset authority, and determine whether the reimbursement certificate is valid according to a verification result.
Optionally, the apparatus further includes a prediction model building unit, including:
the data preprocessing subunit is used for preprocessing the historical reimbursement record data to obtain a sample data set;
the data dividing subunit is used for dividing the sample data set into a training set and a test set;
the pre-training subunit is used for respectively training and testing N training algorithms by utilizing the training set and the test set to obtain the accuracy of each training algorithm, wherein N is an odd number;
an accuracy ranking subunit, configured to rank the accuracy of each of the training algorithms in order from high to low, and rank the accuracy of each of the training algorithms in order from high to low
Figure BDA0003281487060000041
Taking the training algorithm as a first training algorithm
Figure BDA0003281487060000042
Using the training algorithm as a second training algorithm, and using the last training algorithm as a third training algorithm;
the first training subunit is used for dividing the training set into K groups in a first layer of the model, taking the K-1 group as a sub-training set, taking the rest group as a verification set, respectively performing K-fold cross verification on each first training algorithm, calculating the average value of the prediction result of each first training algorithm aiming at each sample in the verification set to obtain a verification set P 'corresponding to the verification result, and calculating the average value of the prediction result of each first training algorithm aiming at each sample in the test set to obtain a test set T' corresponding to the test result;
the second training subunit is used for dividing the verification set P 'of the first layer of the model into K groups, taking the K-1 group as a sub-training set and taking the rest group as a verification set, respectively carrying out K-fold cross verification on each second training algorithm, calculating the average value of the prediction result of each second training algorithm aiming at each sample in the verification set to obtain a verification set P' corresponding to the verification result, and calculating the average value of the prediction result of each second training algorithm aiming at each sample in the test set T 'of the first layer of the model to obtain a test set T' corresponding to the test result;
and the third training subunit is used for dividing the verification set P 'of the second layer of the model into K groups in the third layer of the model, taking the K-1 group as a sub-training set and taking the rest group as a verification set, carrying out K-fold cross verification on the third training algorithm, testing the third training algorithm by using the test set T' of the second layer of the model, and obtaining the data validity prediction model consisting of the first layer, the second layer and the third layer under the condition that the verification result and the test result meet preset conditions.
Optionally, the data preprocessing subunit is specifically configured to:
respectively extracting preset characteristic indexes of each piece of historical reimbursement record data;
respectively carrying out Cartesian product feature combination on preset feature indexes of each piece of historical reimbursement record data, and marking whether each piece of historical reimbursement record data is valid or not to obtain a plurality of sample data;
and removing abnormal values in a plurality of sample data by using a local abnormal factor LOF algorithm to obtain the sample data set.
Optionally, the apparatus further comprises:
and the prompting unit is used for prompting the user that the insurance claim fails under the condition that the reimbursement voucher fails to be verified, the information to be audited is invalid or the information to be audited does not trigger the intelligent insurance claim contract.
Optionally, the apparatus further comprises:
and the reimbursement record storage unit is used for writing the reimbursement record data into a supervision report form after the insurance claim settlement is completed, and sending the supervision report form to a supervision mechanism system.
An insurance claim processing system, comprising: an insurance claim settlement server and a blockchain platform;
the insurance claim settlement server is used for executing the insurance claim settlement processing method disclosed by the embodiment;
and the block chain platform is used for carrying out contract processing on the information to be checked under the condition that the packed information to be checked sent by the insurance claim server is received, judging whether the information to be checked meets the preset triggering condition of the intelligent insurance claim contract or not, and running the intelligent insurance claim contract to complete insurance claim under the condition that the information to be checked meets the triggering condition of the intelligent insurance claim contract.
Compared with the prior art, the invention has the following beneficial effects:
the invention discloses an insurance claim processing method, which is characterized in that under the condition of receiving an insurance claim request, the validity of a reimbursement certificate is verified firstly, under the condition of the reimbursement certificate being valid, a data validity prediction model is called to predict the validity of information to be audited, the waste of computing resources caused by the fact that all information to be audited needs to be called to predict the validity of the data validity prediction model is avoided, under the condition that the information to be audited is valid, the information to be audited is packaged and sent to a block chain, under the condition that the information to be audited triggers a preset insurance claim intelligent contract, the insurance claim is completed, invalid data chaining is avoided, the whole insurance claim process is completed automatically, the insurance claim flow is simplified, the insurance claim period is shortened, and the user experience is effectively improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
FIG. 1 is a schematic flow chart illustrating a method for processing insurance claims according to an embodiment of the present invention;
fig. 2 is a schematic mechanism diagram of an insurance claim settlement processing apparatus according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention discloses an insurance claim processing method which is applied to an insurance claim server, and the method automatically realizes validity verification of reimbursement certificates and information to be verified after receiving an insurance claim request submitted by a user, and sends the information to be verified to a block chain by packaging under the condition that the information to be verified is valid, and completes insurance claim under the condition that the information to be verified triggers a preset insurance claim intelligent contract, thereby avoiding invalid data chaining.
Specifically, referring to fig. 1, the insurance claim settlement processing method disclosed in this embodiment includes the following steps:
s101: under the condition that an insurance claim settlement request is received, analyzing the insurance claim settlement request to obtain information to be audited of a user, wherein the information to be audited at least comprises reimbursement vouchers;
the user can submit the information to be checked and verified required by the insurance claim through the client, the information to be checked and verified corresponding to different dangerous types is different, the information to be checked and verified required by different dangerous types can be preset in the background, for example, medical insurance information and vehicle maintenance records need to be filled, and the insurance claim request is generated after the user submits all the information to be checked and verified.
Analyzing the insurance claim settlement request to obtain information to be audited of the user, wherein the information to be audited at least comprises reimbursement vouchers required by the insurance claim settlement, such as various invoices, and the information to be audited also comprises information of the user, insurance purchase records and the like.
S102: verifying the validity of the reimbursement voucher;
in order to ensure that the reimbursement voucher required for insurance claim settlement is true and valid, the validity of the reimbursement voucher needs to be verified.
Wherein, an optional implementation is: and identifying the seal in the reimbursement certificate through an image identification technology to verify the validity of the reimbursement certificate.
In addition, the validity of the reimbursement certificate can be verified by utilizing a distributed identity verification technology. Taking medical insurance as an example, a user, a medical institution and related authentication party institutions and insurance institutions register distributed digital identity DID in a distributed identity verification system, and the DID is signed by an identity endorsement party and then issued to each institution or individual. When a user needs to apply for insurance claim settlement, a verifiable certificate issued by a distributed digital identity DID is requested to a preset organization, the preset organizations corresponding to different dangerous varieties are different, if the preset organization corresponding to medical insurance is a medical organization, the preset organization corresponding to car insurance is a 4S store, the preset organization encrypts reimbursement certificate data of the user by using a private key to generate the verifiable certificate, and taking the medical insurance as an example, the reimbursement certificate data comprises medical record, consumption record, accident occurrence reason and other data of the user. And after receiving the verifiable certificate, the insurance claim settlement server verifies the reimbursement certificate by using the public key of the preset institution, and determines whether the reimbursement certificate is valid according to the verification result.
S103: calling a pre-constructed data effectiveness prediction model to perform effectiveness prediction on information to be audited under the condition that the reimbursement voucher passes verification;
the data effectiveness prediction model is obtained by training a multi-layer machine learning model by using historical reimbursement record data, wherein training algorithms corresponding to each layer of the multi-layer machine learning model are different.
Specifically, the method for constructing the data effectiveness prediction model comprises the following steps:
1. preprocessing historical reimbursement record data to obtain a sample data set;
preprocessing historical reimbursement record data, firstly respectively extracting preset characteristic indexes of each piece of historical reimbursement record data, then respectively carrying out Cartesian product characteristic combination on the preset characteristic indexes of each piece of historical reimbursement record data, marking whether each piece of historical reimbursement record data is effective or not to obtain a plurality of sample data, and finally removing abnormal values in the plurality of sample data by using a local abnormal factor LOF (local Outlier factor) algorithm to obtain the sample data set.
2. Dividing a sample data set into a training set and a test set;
3. respectively training and testing N training algorithms by using a training set and a test set to obtain the accuracy of each training algorithm, wherein N is an odd number;
the training algorithm may be: adaboost, LR, RF (Random Forest), XGBoost, KNN (K-nearest neighbor classification algorithm), SVR (Support Vector Regression), MLP (Multilayer Perceptron, Multilayer Perceptron algorithm), and the like.
Adaboost is an iterative algorithm, and the core idea thereof is to train different classifiers (weak classifiers) aiming at the same training set, and then to assemble the weak classifiers to form a stronger final classifier (strong classifier).
LR (logistic regression) is a logistic regression algorithm.
Both RF and XGBoost belong to Ensemble Learning (Ensemble Learning), the purpose of which is to improve the generalization ability and robustness of a single learner by combining the prediction results of multiple base learners.
4. The accuracy of each training algorithm is sorted from high to low, the first one
Figure BDA0003281487060000081
The training algorithm is used as the first training algorithm
Figure BDA0003281487060000082
Taking the training algorithm as a second training algorithm, and taking the last training algorithm as a third training algorithm;
5. in the first layer of the model, dividing the training set into K groups, taking the K-1 group as a sub-training set, taking the remaining group as a verification set, respectively performing K-fold cross verification on each first training algorithm, calculating the average value of the prediction result of each first training algorithm for each sample in the verification set to obtain a verification set P 'corresponding to the verification result, and calculating the average value of the prediction result of each first training algorithm for each sample in the test set to obtain a test set T' corresponding to the test result;
6. in the second layer of the model, dividing the verification set P 'of the first layer of the model into K groups, taking K-1 groups as a sub-training set, taking the rest groups as the verification sets, respectively carrying out K-fold cross verification on each second training algorithm, calculating the average value of the prediction result of each second training algorithm aiming at each sample in the verification sets to obtain the verification set P' corresponding to the verification result, and calculating the average value of the prediction result of each second training algorithm aiming at each sample in the test set T 'of the first layer of the model to obtain the test set T' corresponding to the test result;
7. in the third layer of the model, the verification set P 'of the second layer of the model is divided into K groups, K-1 groups are used as sub-training sets, the rest groups are used as verification sets, K-fold cross verification is carried out on the third training algorithm, the third training algorithm is tested by using the test set T' of the second layer of the model, and a data validity prediction model consisting of the first layer, the second layer and the third layer is obtained under the condition that the verification result and the test result meet preset conditions.
The above predictions are expressed as percentages.
It should be noted that, since data features transition from low-level semantics to high-level semantics from the first layer to the third layer, a strong classifier training algorithm is used in the first layer of the model, a training algorithm lower than the first layer is used in the second layer of the model, and the lowest training algorithm is used in the third layer of the model, so that over-fitting is prevented while advantages of different training algorithms are combined, and a better prediction effect is achieved.
S104: and under the condition that the information to be audited is effective, packaging and sending the information to be audited to the block chain platform, and under the condition that the information to be audited triggers a preset insurance claim intelligent contract, completing the insurance claim.
And prompting the user that the insurance claim fails under the condition that the reimbursement voucher fails to be verified or the information to be audited is invalid or the information to be audited does not trigger the intelligent insurance claim contract.
And the block chain platform performs contract processing on the information to be checked under the condition that the packed information to be checked sent by the insurance claim server is received, judges whether the information to be checked meets the preset triggering condition of the intelligent insurance claim contract, and operates the intelligent insurance claim contract to complete insurance claim under the condition that the information to be checked meets the triggering condition of the intelligent insurance claim contract.
Data on the blockchain is traceable and not tamperable. And after the insurance claims are finished, writing the reimbursement record data into a supervision report form, and sending the supervision report form to a supervision mechanism system. The supervisory authority system can check out bad transactions in time by adding a block chain or by supervising reports.
In the method for processing the insurance claim, when an insurance claim request is received, validity of a reimbursement certificate is verified, a data validity prediction model is called to predict validity of information to be audited only when the reimbursement certificate is valid, so that waste of computing resources caused by the fact that all information to be audited needs to be called to predict the validity of the data validity prediction model is avoided, the information to be audited is packed and sent to a block chain when the information to be audited is valid, the insurance claim is completed under the condition that the information to be audited triggers a preset intelligent insurance claim contract, chain linking of invalid data is avoided, the whole insurance claim process is completed automatically, an insurance claim process is simplified, an insurance claim period is shortened, and user experience is effectively improved.
Based on the insurance claim processing method disclosed in the foregoing embodiment, this embodiment correspondingly discloses an insurance claim processing apparatus applied to an insurance claim server, please refer to fig. 2, where the apparatus includes:
the claim settlement request analyzing unit 201 is configured to, in a case that an insurance claim settlement request is received, analyze the insurance claim settlement request to obtain information to be audited of a user, where the information to be audited at least includes reimbursement vouchers;
an reimbursement credential verification unit 202 for verifying the validity of the reimbursement credential;
the data effectiveness prediction unit 203 is configured to, under the condition that the reimbursement certificate passes verification, invoke a pre-constructed data effectiveness prediction model to perform effectiveness prediction on the information to be audited, where the data effectiveness prediction model is obtained by training a multi-layer machine learning model by using historical reimbursement record data, and training algorithms corresponding to each layer of the multi-layer machine learning model are different;
and the data uplink unit 204 is configured to package and send the information to be audited to the block chain platform under the condition that the information to be audited is valid, and complete insurance claim settlement under the condition that the information to be audited triggers a preset intelligent insurance claim contract.
Optionally, the reimbursement certificate is a verifiable certificate which is issued by a user requesting a preset organization by using a distributed Digital Identity (DID), and the verifiable certificate is generated by the preset organization encrypting reimbursement certificate data of the user by using a private key;
the reimbursement certificate verification unit 202 is specifically configured to verify the reimbursement certificate by using the public key of the preset authority, and determine whether the reimbursement certificate is valid according to a verification result.
Optionally, the apparatus further includes a prediction model building unit, including:
the data preprocessing subunit is used for preprocessing the historical reimbursement record data to obtain a sample data set;
the data dividing subunit is used for dividing the sample data set into a training set and a test set;
the pre-training subunit is used for respectively training and testing N training algorithms by utilizing the training set and the test set to obtain the accuracy of each training algorithm, wherein N is an odd number;
an accuracy ranking subunit, configured to rank the accuracy of each of the training algorithms in order from high to low, and rank the accuracy of each of the training algorithms in order from high to low
Figure BDA0003281487060000101
Taking the training algorithm as a first training algorithm
Figure BDA0003281487060000102
Using the training algorithm as a second training algorithm, and using the last training algorithm as a third training algorithm;
the first training subunit is used for dividing the training set into K groups in a first layer of the model, taking the K-1 group as a sub-training set, taking the rest group as a verification set, respectively performing K-fold cross verification on each first training algorithm, calculating the average value of the prediction result of each first training algorithm aiming at each sample in the verification set to obtain a verification set P 'corresponding to the verification result, and calculating the average value of the prediction result of each first training algorithm aiming at each sample in the test set to obtain a test set T' corresponding to the test result;
the second training subunit is used for dividing the verification set P 'of the first layer of the model into K groups, taking the K-1 group as a sub-training set and taking the rest group as a verification set, respectively carrying out K-fold cross verification on each second training algorithm, calculating the average value of the prediction result of each second training algorithm aiming at each sample in the verification set to obtain a verification set P' corresponding to the verification result, and calculating the average value of the prediction result of each second training algorithm aiming at each sample in the test set T 'of the first layer of the model to obtain a test set T' corresponding to the test result;
and the third training subunit is used for dividing the verification set P 'of the second layer of the model into K groups in the third layer of the model, taking the K-1 group as a sub-training set and taking the rest group as a verification set, carrying out K-fold cross verification on the third training algorithm, testing the third training algorithm by using the test set T' of the second layer of the model, and obtaining the data validity prediction model consisting of the first layer, the second layer and the third layer under the condition that the verification result and the test result meet preset conditions.
Optionally, the data preprocessing subunit is specifically configured to:
respectively extracting preset characteristic indexes of each piece of historical reimbursement record data;
respectively carrying out Cartesian product feature combination on preset feature indexes of each piece of historical reimbursement record data, and marking whether each piece of historical reimbursement record data is valid or not to obtain a plurality of sample data;
and removing abnormal values in a plurality of sample data by using a local abnormal factor LOF algorithm to obtain the sample data set.
Optionally, the apparatus further comprises:
and the prompting unit is used for prompting the user that the insurance claim fails under the condition that the reimbursement voucher fails to be verified, the information to be audited is invalid or the information to be audited does not trigger the intelligent insurance claim contract.
Optionally, the apparatus further comprises:
and the reimbursement record storage unit is used for writing the reimbursement record data into a supervision report form after the insurance claim settlement is completed, and sending the supervision report form to a supervision mechanism system.
According to the insurance claim processing device disclosed by the embodiment, under the condition that an insurance claim request is received, the validity of a reimbursement certificate is verified firstly, a data validity prediction model is called to predict the validity of information to be audited under the condition that the reimbursement certificate is valid, the waste of computing resources caused by the fact that all information to be audited needs to be called to predict the validity of the data validity prediction model is avoided, under the condition that the information to be audited is valid, the information to be audited is packaged and sent to a block chain, under the condition that the information to be audited triggers a preset insurance claim intelligent contract, insurance claim is completed, invalid data chaining is avoided, the whole insurance claim process is completed automatically, an insurance claim flow is simplified, an insurance claim period is shortened, and user experience is effectively improved.
The embodiment also discloses an insurance claim settlement processing system, which comprises: an insurance claim settlement server and a blockchain platform;
the insurance claim settlement server is used for executing the insurance claim settlement processing method disclosed by the embodiment;
and the block chain platform is used for carrying out contract processing on the information to be checked under the condition that the packed information to be checked sent by the insurance claim server is received, judging whether the information to be checked meets the preset triggering condition of the intelligent insurance claim contract or not, and running the intelligent insurance claim contract to complete insurance claim under the condition that the information to be checked meets the triggering condition of the intelligent insurance claim contract.
It should be noted that the insurance claim settlement processing method, device and system provided by the invention can be applied to the field of block chains or the financial field. The above description is only an example, and does not limit the application field of the insurance claim settlement processing method, apparatus and system provided by the present invention.
The embodiments in the present description 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. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description.
It is further 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 apparatus 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 apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module may reside in Random Access Memory (RAM), memory, Read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
The above embodiments can be combined arbitrarily, and the features described in the embodiments in the present specification can be replaced or combined with each other in the above description of the disclosed embodiments, so that those skilled in the art can implement or use the present application.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (10)

1. An insurance claim processing method is applied to an insurance claim server, and the method comprises the following steps:
under the condition that an insurance claim settlement request is received, analyzing the insurance claim settlement request to obtain information to be audited of a user, wherein the information to be audited at least comprises reimbursement vouchers;
verifying the validity of the reimbursement credential;
calling a pre-constructed data effectiveness prediction model to perform effectiveness prediction on the information to be checked under the condition that the reimbursement voucher passes verification, wherein the data effectiveness prediction model is obtained by training a multi-layer machine learning model by using historical reimbursement record data, and training algorithms corresponding to each layer of the multi-layer machine learning model are different;
and packaging and sending the information to be audited to a block chain platform under the condition that the information to be audited is effective, and finishing insurance claim settlement under the condition that the information to be audited triggers a preset insurance claim settlement intelligent contract.
2. The method according to claim 1, wherein the reimbursement certificate is a verifiable certificate issued by a user requesting a default organization by using a distributed Digital Identity (DID), and the verifiable certificate is generated by the default organization encrypting reimbursement certificate data of the user by using a private key;
verifying the validity of the reimbursement credential, comprising:
and verifying the reimbursement certificate by using the public key of the preset mechanism, and determining whether the reimbursement certificate is valid according to a verification result.
3. The method of claim 1, wherein constructing the data validity prediction model comprises:
preprocessing historical reimbursement record data to obtain a sample data set;
dividing the sample data set into a training set and a test set;
respectively training and testing N training algorithms by utilizing the training set and the test set to obtain the accuracy of each training algorithm, wherein N is an odd number;
the accuracy of each training algorithm is sorted from high to low, the first one
Figure FDA0003281487050000011
Taking the training algorithm as a first training algorithm
Figure FDA0003281487050000012
Using the training algorithm as a second training algorithm, and using the last training algorithm as a third training algorithm;
in a first layer of the model, dividing the training set into K groups, taking K-1 groups as a sub-training set, taking the remaining group as a verification set, respectively performing K-fold cross verification on each first training algorithm, calculating the average value of the prediction result of each first training algorithm for each sample in the verification set to obtain a verification set P 'corresponding to the verification result, and calculating the average value of the prediction result of each first training algorithm for each sample in the test set to obtain a test set T' corresponding to the test result;
in a second layer of the model, dividing a verification set P 'of the first layer of the model into K groups, taking K-1 groups as a sub-training set, taking the rest groups as the verification sets, respectively carrying out K-fold cross verification on each second training algorithm, calculating the average value of the prediction result of each second training algorithm aiming at each sample in the verification sets to obtain a verification set P' corresponding to the verification result, and calculating the average value of the prediction result of each second training algorithm aiming at each sample in a test set T 'of the first layer of the model to obtain a test set T' corresponding to the test result;
in the third layer of the model, dividing the verification set P 'of the second layer of the model into K groups, taking K-1 groups as a sub-training set, taking the rest group as a verification set, carrying out K-fold cross verification on the third training algorithm, testing the third training algorithm by using the test set T' of the second layer of the model, and obtaining the data validity prediction model consisting of the first layer, the second layer and the third layer under the condition that the verification result and the test result meet the preset conditions.
4. The method of claim 3, wherein the pre-processing the historical reimbursement record data to obtain a sample data set comprises:
respectively extracting preset characteristic indexes of each piece of historical reimbursement record data;
respectively carrying out Cartesian product feature combination on preset feature indexes of each piece of historical reimbursement record data, and marking whether each piece of historical reimbursement record data is valid or not to obtain a plurality of sample data;
and removing abnormal values in a plurality of sample data by using a local abnormal factor LOF algorithm to obtain the sample data set.
5. The method of claim 1, further comprising:
and prompting the user that the insurance claim is not passed under the condition that the reimbursement voucher is not verified, the information to be audited is invalid or the information to be audited does not trigger the intelligent insurance claim contract.
6. The method of claim 1, wherein after completing an insurance claim, the method further comprises:
and writing the reimbursement record data into a supervision report, and sending the supervision report to a supervision mechanism system.
7. An insurance claim processing apparatus applied to an insurance claim server, the apparatus comprising:
the system comprises a claim settlement request analyzing unit, a claim settlement request analyzing unit and a verification processing unit, wherein the claim settlement request analyzing unit is used for analyzing an insurance claim settlement request under the condition that the insurance claim settlement request is received to obtain information to be verified of a user, and the information to be verified at least comprises reimbursement vouchers;
the reimbursement certificate verification unit is used for verifying the validity of the reimbursement certificate;
the data effectiveness prediction unit is used for calling a pre-constructed data effectiveness prediction model to perform effectiveness prediction on the information to be checked under the condition that the reimbursement voucher passes verification, wherein the data effectiveness prediction model is obtained by training a multi-layer machine learning model by using historical reimbursement record data, and training algorithms corresponding to each layer of the multi-layer machine learning model are different;
and the data uplink unit is used for packaging and sending the information to be audited to a block chain platform under the condition that the information to be audited is effective, and finishing insurance claim settlement under the condition that the information to be audited triggers a preset intelligent insurance claim contract.
8. The apparatus according to claim 7, wherein the reimbursement certificate is a verifiable certificate issued by a user requesting a default organization by using a distributed Digital Identity (DID), and the verifiable certificate is generated by the default organization encrypting reimbursement certificate data of the user by using a private key;
the reimbursement certificate verification unit is specifically configured to verify the reimbursement certificate by using the public key of the preset authority, and determine whether the reimbursement certificate is valid according to a verification result.
9. The apparatus of claim 7, further comprising a predictive model building unit comprising:
the data preprocessing subunit is used for preprocessing the historical reimbursement record data to obtain a sample data set;
the data dividing subunit is used for dividing the sample data set into a training set and a test set;
the pre-training subunit is used for respectively training and testing N training algorithms by utilizing the training set and the test set to obtain the accuracy of each training algorithm, wherein N is an odd number;
an accuracy ranking subunit, configured to rank the accuracy of each of the training algorithms in order from high to low, and rank the accuracy of each of the training algorithms in order from high to low
Figure FDA0003281487050000031
Taking the training algorithm as a first training algorithm
Figure FDA0003281487050000032
Using the training algorithm as a second training algorithm, and using the last training algorithm as a third training algorithm;
the first training subunit is used for dividing the training set into K groups in a first layer of the model, taking the K-1 group as a sub-training set, taking the rest group as a verification set, respectively performing K-fold cross verification on each first training algorithm, calculating the average value of the prediction result of each first training algorithm aiming at each sample in the verification set to obtain a verification set P 'corresponding to the verification result, and calculating the average value of the prediction result of each first training algorithm aiming at each sample in the test set to obtain a test set T' corresponding to the test result;
the second training subunit is used for dividing the verification set P 'of the first layer of the model into K groups, taking the K-1 group as a sub-training set and taking the rest group as a verification set, respectively carrying out K-fold cross verification on each second training algorithm, calculating the average value of the prediction result of each second training algorithm aiming at each sample in the verification set to obtain a verification set P' corresponding to the verification result, and calculating the average value of the prediction result of each second training algorithm aiming at each sample in the test set T 'of the first layer of the model to obtain a test set T' corresponding to the test result;
and the third training subunit is used for dividing the verification set P 'of the second layer of the model into K groups in the third layer of the model, taking the K-1 group as a sub-training set and taking the rest group as a verification set, carrying out K-fold cross verification on the third training algorithm, testing the third training algorithm by using the test set T' of the second layer of the model, and obtaining the data validity prediction model consisting of the first layer, the second layer and the third layer under the condition that the verification result and the test result meet preset conditions.
10. An insurance claim processing system, comprising: an insurance claim settlement server and a blockchain platform;
the insurance claim server is used for executing an insurance claim processing method according to claims 1-6;
and the block chain platform is used for carrying out contract processing on the information to be checked under the condition that the packed information to be checked sent by the insurance claim server is received, judging whether the information to be checked meets the preset triggering condition of the intelligent insurance claim contract or not, and running the intelligent insurance claim contract to complete insurance claim under the condition that the information to be checked meets the triggering condition of the intelligent insurance claim contract.
CN202111134366.4A 2021-09-27 2021-09-27 Insurance claim processing method, device and system Active CN113781248B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111134366.4A CN113781248B (en) 2021-09-27 2021-09-27 Insurance claim processing method, device and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111134366.4A CN113781248B (en) 2021-09-27 2021-09-27 Insurance claim processing method, device and system

Publications (2)

Publication Number Publication Date
CN113781248A true CN113781248A (en) 2021-12-10
CN113781248B CN113781248B (en) 2024-08-20

Family

ID=78853776

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111134366.4A Active CN113781248B (en) 2021-09-27 2021-09-27 Insurance claim processing method, device and system

Country Status (1)

Country Link
CN (1) CN113781248B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116797376A (en) * 2023-08-26 2023-09-22 杭州字节方舟科技有限公司 Asset management method, system, server and storage medium based on block chain

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090182696A1 (en) * 2008-01-10 2009-07-16 Deutsche Telekom Ag Stacking schema for classification tasks
CN105279382A (en) * 2015-11-10 2016-01-27 成都数联易康科技有限公司 Medical insurance abnormal data on-line intelligent detection method
CN107798514A (en) * 2017-11-22 2018-03-13 阿里巴巴集团控股有限公司 The method and apparatus that Claims Resolution is realized based on credit
CN109242538A (en) * 2018-08-14 2019-01-18 平安医疗健康管理股份有限公司 Medical insurance pricing method, device, computer equipment and storage medium
CN110956497A (en) * 2019-11-27 2020-04-03 桂林电子科技大学 Method for predicting repeated purchasing behavior of user of electronic commerce platform
CN111507851A (en) * 2020-04-23 2020-08-07 腾讯科技（深圳）有限公司 Block chain-based medical insurance claim settlement processing method, device and system and storage medium
CN113344437A (en) * 2021-06-29 2021-09-03 中国平安人寿保险股份有限公司 Claim settlement service processing method and device, computer equipment and storage medium

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090182696A1 (en) * 2008-01-10 2009-07-16 Deutsche Telekom Ag Stacking schema for classification tasks
CN105279382A (en) * 2015-11-10 2016-01-27 成都数联易康科技有限公司 Medical insurance abnormal data on-line intelligent detection method
CN107798514A (en) * 2017-11-22 2018-03-13 阿里巴巴集团控股有限公司 The method and apparatus that Claims Resolution is realized based on credit
CN109242538A (en) * 2018-08-14 2019-01-18 平安医疗健康管理股份有限公司 Medical insurance pricing method, device, computer equipment and storage medium
CN110956497A (en) * 2019-11-27 2020-04-03 桂林电子科技大学 Method for predicting repeated purchasing behavior of user of electronic commerce platform
CN111507851A (en) * 2020-04-23 2020-08-07 腾讯科技（深圳）有限公司 Block chain-based medical insurance claim settlement processing method, device and system and storage medium
CN113344437A (en) * 2021-06-29 2021-09-03 中国平安人寿保险股份有限公司 Claim settlement service processing method and device, computer equipment and storage medium

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116797376A (en) * 2023-08-26 2023-09-22 杭州字节方舟科技有限公司 Asset management method, system, server and storage medium based on block chain
CN116797376B (en) * 2023-08-26 2023-12-01 杭州字节方舟科技有限公司 Asset management method, system, server and storage medium based on block chain

Also Published As

Publication number Publication date
CN113781248B (en) 2024-08-20

Similar Documents

Publication Publication Date Title
US20220094709A1 (en) Automatic Machine Learning Vulnerability Identification and Retraining
CN111681091B (en) Financial risk prediction method and device based on time domain information and storage medium
US20210374749A1 (en) User profiling based on transaction data associated with a user
US11151573B2 (en) Intelligent chargeback processing platform
US20180046939A1 (en) Automated Machine Learning Feature Processing
US12034734B2 (en) Systems and methods for identifying synthetic identities associated with network communications
Zhou et al. Fraud detection within bankcard enrollment on mobile device based payment using machine learning
US20220351284A1 (en) System and method for the rapid, flexible approval and disbursement of a loan
US11397950B2 (en) Systems and methods for authenticating an electronic transaction
CN111861174A (en) Credit assessment method for user portrait
CN112465627A (en) Financial loan auditing method and system based on block chain and machine learning
CN111461859A (en) Loan pairing system and method
US12061671B2 (en) Data compression techniques for machine learning models
CN111583017A (en) Risk strategy generation method and device based on guest group positioning and electronic equipment
Wu et al. Application analysis of credit scoring of financial institutions based on machine learning model
CN113298121A (en) Message sending method and device based on multi-data source modeling and electronic equipment
CN113781248B (en) Insurance claim processing method, device and system
CN114090985A (en) Method and device for setting user operation permission in block chain and electronic equipment
Dong et al. Real-time Fraud Detection in e-Market Using Machine Learning Algorithms.
CN110751316A (en) Method and device for predicting resolution result and terminal equipment
US20210090136A1 (en) Ai to ai communication
CN117575773A (en) Method, device, computer equipment and storage medium for determining service data
CN117115833A (en) Certificate classification method, device, equipment and storage medium
US20240161117A1 (en) Trigger-Based Electronic Fund Transfers
CN113706258B (en) Product recommendation method, device, equipment and storage medium based on combined model

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
GR01 Patent grant
GR01 Patent grant