CN108846766A - The self-service Claims Resolution success rate prediction technique of vehicle insurance based on deep learning - Google Patents
The self-service Claims Resolution success rate prediction technique of vehicle insurance based on deep learning Download PDFInfo
- Publication number
- CN108846766A CN108846766A CN201810658529.0A CN201810658529A CN108846766A CN 108846766 A CN108846766 A CN 108846766A CN 201810658529 A CN201810658529 A CN 201810658529A CN 108846766 A CN108846766 A CN 108846766A
- Authority
- CN
- China
- Prior art keywords
- data
- hidden layer
- user
- weighting parameter
- deep learning
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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"
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Finance (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- Accounting & Taxation (AREA)
- Human Resources & Organizations (AREA)
- General Business, Economics & Management (AREA)
- Operations Research (AREA)
- Computational Linguistics (AREA)
- Entrepreneurship & Innovation (AREA)
- Technology Law (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Game Theory and Decision Science (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The present invention relates to a kind of self-service Claims Resolution success rate prediction techniques of vehicle insurance based on deep learning, belong to big data analysis electric powder prediction, this method comprises the following steps：S1：Based on insurance company's user's end subscriber, user data required for structuring user's are drawn a portrait is obtained；S2：Data processing is carried out to acquired user data；S3：Characterization factor extraction is carried out to the user data after processing；S4：The prediction model based on deep learning algorithm is established to predict Claims Resolution success rate.On-line off-line data of the method for the present invention based on user's magnanimity build prediction model using deep learning method, and user's portrait is applied to settle a claim in self-service success rate prediction.Claims Resolution service is provided for insurance company's personalization and provides condition, saves artificial, raising working efficiency and user experience.
Description
Technical field
The invention belongs to big data analysis electric powder prediction, be related to a kind of self-service Claims Resolution of the vehicle insurance based on deep learning at
Power forecasting method.
Background technique
With the development of the times, insurance service turns to centered on consumer from product-centered.And by mutual
The high speed development of networking, more and more users' data have been recorded, and a large amount of on-line off-line data are numerous and complicated application
Bring possibility.Insurance company is product searching target user and is that user's customed product service also becomes a reality.
For subsidiary company strategic level, user's portrait can help enterprise's progress market to see clearly, estimate market scale, thus
Auxiliary formulates phased goal, instructs very important decision, promotes ROI, avoids homogeneity, carries out personal marketing；From product itself
For angle, user's portrait can carry out crowd's subdivision around product, the core crowd of product be determined, to aid in determining whether to produce
Product positioning, optimizes the function point of product；For data management angle, user's portrait helps to establish data assets, excavates number
According to value, keep data analysis more accurate, it might even be possible to carry out data trade, promote data circulation.The product of insurance industry
As a long period product, the conversion ratio that Insurance User buys insurance products again is improved, is that insurance company manages old use
One vital task at family.
Summary of the invention
In view of this, the purpose of the present invention is to provide a kind of self-service Claims Resolution success rate predictions of vehicle insurance based on deep learning
Method, the on-line off-line data based on user's magnanimity build prediction model using deep learning method, and user's portrait is applied to
It settles a claim in self-service success rate prediction.Claims Resolution service is provided for insurance company's personalization and provides condition, saves artificial, raising work
Efficiency and user experience.
In order to achieve the above objectives, the present invention provides the following technical solutions：
The self-service Claims Resolution success rate prediction technique of vehicle insurance based on deep learning, this method comprise the following steps：
S1：Based on insurance company's user's end subscriber, user data required for structuring user's are drawn a portrait is obtained；
S2：Data processing is carried out to acquired user data；
S3：Characterization factor extraction is carried out to the user data after processing；
S4：The prediction model based on deep learning algorithm is established to predict Claims Resolution success rate.
Further, user data described in step S1 includes static data, dynamic data and auxiliary data；
The static data be to the relevant personage's fixed data of Claims Resolution product, photo, name, age, occupation comprising user,
Hobby；
The dynamic data is the usage scenario data of Claims Resolution product；
The auxiliary data is industry experience data.
Further, step S2 is specifically included the following steps：
S21：Source data collection is carried out to the user data；
S22：Integrality processing is carried out to the initial data being collected into, corrects the incomplete property data of initial data；
S23：Uniqueness processing is carried out to integrality treated data, eliminates redundant digit present in data after integrality processing
According to；
S24：Authoritative processing is carried out to uniqueness treated data, the different numbers of identical parameters in unified multi-source data
Value；
S25：Treated that data carry out legitimacy processing to authoritative, casts out the data for obviously not meeting common sense；
S26：Consistency treatment is carried out to legitimacy treated data, integrates substantially identical data.
Further, in step S3, the characterization factor includes：Driver information, information of vehicles, temporal information, address letter
Breath, event information and the fraud factor.
Further, the processing step of the address information is as follows：
S31：The word that address information is last is taken, a kind of address characteristic information is obtained；
S32：Two words that address information is last are taken, two class address characteristic informations are obtained；
S33：Three words that address information is last are taken, three classes address characteristic information is obtained；
S34：By the address characteristic information of above three type according to three classes address characteristic information, two class address characteristic informations, one
The sequence of class address characteristic information matches address information, using the data after matching as effective address information；
S35：According to above-mentioned effective address data, it is in danger characteristic in conjunction with social experience and vehicle insurance, according to the period according to population and vehicle
Congestion level classify to effective address information.
Further, deep learning algorithm described in step S4 is based on deep neural network, and the deep neural network includes
The input layer successively connected entirely, hidden layer and output layer；
Step S4 is specifically included the following steps：
S41：Data comprising user tag are divided into training setAnd test set, and choose a training setIt is trained, the optimization weighting parameter of acquisition input layer to the first hidden layer,；
S42：By training setThe first hidden layer is propagated through forward obtains fisrt feature, and as input
The second hidden layer of training, obtains the optimization weighting parameter of the first hidden layer to the second hidden layer,；
S43：By fisrt featureThe second hidden layer is propagated through forward obtains second feature, and with this
To input training third hidden layer, the optimization weighting parameter of the second hidden layer of acquisition to third hidden layer,；
S44：Second feature is propagated through forward to third hidden layer and obtains second feature, and instructed as input
Practice output layer, obtains optimization output weighting parameter；
S45：By the optimization weighting parameter of input layer to the first hidden layer,, optimization of first hidden layer to the second hidden layer
Weighting parameter,, the optimization weighting parameter of the second hidden layer to third hidden layer,And optimization output weighting parameterMake
For initiation parameter, by training setAs input, the cost function and ladder of whole network are provided according to back-propagating principle
Degree, adjusts weight parameter in known prediction label, obtains the input layer of optimization to the weighting parameter of the first hidden layer,, the weighting parameter of the first hidden layer to the second hidden layer of optimization,, the of optimization
Weighting parameter of two hidden layers to third hidden layer,And the output weighting parameter optimized；
S46：By adjusting good weighting parameter deep neural network to test setIt is predicted, it is accurate to calculate prediction
Rate executes S45 and readjusts weighting parameter if predictablity rate is unsatisfactory for design requirement, if predictablity rate meets design
The deep neural network is then used for actual prediction by demand.
The beneficial effects of the present invention are：The present invention is based on the on-line off-line data of user's magnanimity, using deep learning side
Method builds prediction model, and customer portrait is applied to settle a claim in self-service success rate prediction.Self-service success interest rate of settling a claim is to user
Whether can complete independently mobile phone it is self-service Claims Resolution make prediction, self-service user can be completed by show for customer portrait, insurance public affairs
Self-service Claims Resolution method is recommended first by department, and to the client that cannot complete self-service Claims Resolution, then exempt to recommend self-service Claims Resolution mode.For insurance
Company's personalization provides Claims Resolution service and provides condition, saves artificial, raising working efficiency and user experience.
Detailed description of the invention
In order to keep the purpose of the present invention, technical scheme and beneficial effects clearer, the present invention provides following attached drawing and carries out
Explanation：
Fig. 1 is the method for the present invention flow chart；
Fig. 2 is neural network structure figure of the present invention；
Fig. 3 is prediction result schematic diagram of the present invention.
Specific embodiment
Below in conjunction with attached drawing, a preferred embodiment of the present invention will be described in detail.
As shown in Figure 1, the present invention is a kind of self-service Claims Resolution success rate prediction technique of vehicle insurance based on deep learning, it is specific to wrap
Containing following steps：
Step 1：Based on insurance company's user's end subscriber, user data required for structuring user's are drawn a portrait is obtained；
Step 2：Data processing is carried out to acquired user data；
Step 3：Characterization factor extraction is carried out to the user data after processing
Step 4：The prediction model based on deep learning algorithm is established to predict Claims Resolution success rate.
Wherein in step 1, the source of user data required for obtaining
For car owner crowd, the insurance company client for using insurance company's vehicle insurance Claims Resolution App, the data of structuring user's portrait needs
Composition includes static data, dynamic data and auxiliary data.
Static data（Essential attribute/fixed data）：The people relevant with product such as photo, name, age, occupation, hobby
Object attribute, static data are accepted insurance from insurance company and are obtained in data；
Dynamic data（Usage scenario data）：Specific vehicle is in danger situation, and dynamic data is obtained from insurance company's vehicle insurance data
?；
Auxiliary data（Industry experience data）：For derived data, obtained according to industry experience.
Static data will enrich in product systems, transaction system and the CRM system inside insurance company
Data concentrated, build data warehouse, excavate introduce the factor.Such as car owner's gender, car owner's age, car owner's occupation, vehicle
Vehicle age, car owner's driving age etc..Dynamic data is also settled a claim comprising user in insurance company's vehicle insurance other than the data of being in danger comprising vehicle
The history click data of App.
Auxiliary data, that is, artificial investigational data acquisition, by the empirical data obtained to industry observational study, for example infers
The risk of fraud time of acquisition, the time difference etc. that accident occurs and reports a case to the security authorities.
Step 2 data processing specifically includes the following steps：
2.1：Source data collection is carried out to the user data；
2.2：Integrality processing is carried out to the initial data being collected into, corrects the incomplete property data of initial data；
2.3：Uniqueness processing is carried out to integrality treated data, eliminates redundant digit present in data after integrality processing
According to；
2.4：Authoritative processing is carried out to uniqueness treated data, the different numbers of identical parameters in unified multi-source data
Value；
2.5：Treated that data carry out legitimacy processing to authoritative, casts out the data for obviously not meeting common sense；
2.6：Consistency treatment is carried out to legitimacy treated data, integrates substantially identical data.
The accurate data further to be standardized.
Step 3, characterization factor acquisition include：
Data handling procedure more than, from the n item attribute of initial data, direct or indirect is obtained greater than n features
The factor.Including：
1) driver information；
2) information of vehicles；
3) temporal information；
4) address information；
5) event information；
6) factor etc. is cheated.
Address information therein is because its data volume is huge, the diversity of text feature and the person's of surveying input habit, so that right
The classified use of address information is especially difficult, in the market there is no ready-made, an effective literal address classification method,
For this problem, the present invention also proposes a kind of address processing method, and specific step is as follows：
（1）Firstly, take address information the last character, obtain it is preliminary, including garden, road, street, county, the village, town, gulf, port, hilllock,
Bridge, village, seat, group, number, building, stand, stockaded village, residence, the feature including the lists such as institute (comprising 105 features in the embodiment of the present invention)；
（2）Secondly, take address information most latter two word, obtain it is preliminary, including area of race, master station, army unit, head factory, in one, two
In, in three, backyard, number institute, home, garden, Hua Yuan, bold and unconstrained garden, orchard, park, unit, apartment, prison, left and right, middle school etc. pair
Feature (including 53 features in the embodiment of the present invention) including word；
（3）Again, last three words of address information are taken, obtain three preliminary word features, including gas station, gas station, gas station,
Feature (including 8 features in the embodiment of the present invention) including three word such as bus station, charge station, kindergarten, university city.
（4）Then, address information is matched according to the sequence of three words, two words, a words, the number that will be filtered out
According to as valid data；
（5）It is in danger characteristic in conjunction with practical social experience and vehicle insurance, address information is divided into specific category by the embodiment of the present invention, is wrapped
It includes：It is a kind of：People Duo Che is also more, and people Bi Che is more, whole day crowded section of highway；
Two classes：People Duo Che is also more, peak period congestion on and off duty, other periods people's Multiple Sections；
Three classes：People Duo Che is also more, and speed is slower, peak congested link on and off duty；
Etc..
Step 4：It is established based on deep learning algorithm model
Deep neural network（Deep Neural Networks, hereinafter referred to as DNN）It is the basis of deep learning, inside DNN
Neural net layer can be divided into three classes, input layer, and hidden layer and output layer, in general first layer are input layer, the last layer
It is output layer, and the intermediate number of plies is all hidden layer, is connected entirely between layers, that is to say, that theiAny one of layer
Neuron is centainly connected with layer any one neuron, as shown in Figure 2.
4.1：Data comprising user tag are divided into training setAnd test set, and choose a training setIt is trained, the optimization weighting parameter of acquisition input layer to the first hidden layer,。
4.2：By training setThe first hidden layer is propagated through forward obtains fisrt feature, and as
The second hidden layer of training is inputted, the optimization weighting parameter of the first hidden layer to the second hidden layer is obtained,。
4.3：By fisrt featureThe second hidden layer is propagated through forward obtains second feature, and with
This trains third hidden layer to input, the optimization weighting parameter of the second hidden layer of acquisition to third hidden layer,。
4.4：By second featureThird hidden layer is propagated through forward obtains second feature, and with
This obtains optimization output weighting parameter to input training output layer。
4.5：By the optimization weighting parameter of input layer to the first hidden layer,, the first hidden layer to the second hidden layer
Optimize weighting parameter,, the optimization weighting parameter of the second hidden layer to third hidden layer,And optimization output weight
ParameterAs initiation parameter, by training setAs input, the cost of whole network is provided according to back-propagating principle
Function and gradient adjust weight parameter in known prediction label, obtain the input layer of optimization to the first hidden layer
Weighting parameter,, the weighting parameter of the first hidden layer to the second hidden layer of optimization,, most
Weighting parameter of second hidden layer of optimization to third hidden layer,And the output weighting parameter optimized。
4.6：By adjusting good weighting parameter deep neural network to test setIt is predicted, it is quasi- to calculate prediction
True rate executes S45 and readjusts weighting parameter if predictablity rate is unsatisfactory for design requirement, if predictablity rate satisfaction is set
The deep neural network is then used for actual prediction by meter demand.
Fig. 3 is the specific prediction result schematic diagram of the embodiment of the present invention, as shown in figure 3, pressing on interface comprising single prediction
Button, batch forecast button and clear history record button.
If single is predicted, needs to fill in input data, click start button after the completion of input, provide predicted value.
If it is batch forecast, then excel file can be loaded into read data in batches, and write in the next column of former data
Enter predicted value, it can be with preview file data below interface after select file.
Close button is clicked after prediction, will pop up the name element that dialog box saves this time batch storage.
Finally, it is stated that preferred embodiment above is only to illustrate the technical solution of invention rather than limits, although passing through
Above preferred embodiment is described in detail the present invention, however, those skilled in the art should understand that, can be in shape
Various changes are made in formula and to it in details, without departing from claims of the present invention limited range.
Claims (6)
1. the self-service Claims Resolution success rate prediction technique of the vehicle insurance based on deep learning, it is characterised in that：This method comprises the following steps：
S1：Based on insurance company's user's end subscriber, user data required for structuring user's are drawn a portrait is obtained；
S2：Data processing is carried out to acquired user data；
S3：Characterization factor extraction is carried out to the user data after processing；
S4：The prediction model based on deep learning algorithm is established to predict Claims Resolution success rate.
2. the self-service Claims Resolution success rate prediction technique of the vehicle insurance based on deep learning according to claim 1, it is characterised in that：
User data described in step S1 includes static data, dynamic data and auxiliary data；
The static data is personage's fixed data relevant to Claims Resolution product；
The dynamic data is the usage scenario data of Claims Resolution product；
The auxiliary data is industry experience data.
3. the self-service Claims Resolution success rate prediction technique of the vehicle insurance based on deep learning according to claim 2, it is characterised in that：
Step S2 is specifically included the following steps：
S21：Source data collection is carried out to the user data；
S22：Integrality processing is carried out to the initial data being collected into, corrects the incomplete property data of initial data；
S23：Uniqueness processing is carried out to integrality treated data, eliminates redundant digit present in data after integrality processing
According to；
S24：Authoritative processing is carried out to uniqueness treated data, the different numbers of identical index in unified multi-source data
Value；
S25：Treated that data carry out legitimacy processing to authoritative, casts out the data for obviously not meeting common sense；
S26：Consistency treatment is carried out to legitimacy treated data, integrates substantially identical data.
4. the self-service Claims Resolution success rate prediction technique of the vehicle insurance based on deep learning according to claim 3, it is characterised in that：
In step S3, the characterization factor includes：Driver information, information of vehicles, temporal information, address information and are taken advantage of at event information
Cheat the factor.
5. the self-service Claims Resolution success rate prediction technique of the vehicle insurance based on deep learning according to claim 4, it is characterised in that：
The classifying step of the address information is as follows：
S31：The word that address information is last is taken, a kind of address characteristic information is obtained；
S32：Two words that address information is last are taken, two class address characteristic informations are obtained；
S33：Three words that address information is last are taken, three classes address characteristic information is obtained；
S34：By the address characteristic information of above three type according to three classes address characteristic information, two class address characteristic informations, one
The sequence of class address characteristic information matches address information, using the data after matching as effective address information；
S35：According to above-mentioned effective address data, it is in danger characteristic in conjunction with social experience and vehicle insurance, according to the period according to population and vehicle
Congestion level classify to effective address information.
6. the self-service Claims Resolution success rate prediction technique of the vehicle insurance based on deep learning according to claim 3, it is characterised in that：
Deep learning algorithm described in step S4 is based on deep neural network, and the deep neural network includes the input successively connected entirely
Layer, hidden layer and output layer；
Step S4 is specifically included the following steps：
S41：Data comprising user tag are divided into training setAnd test set, and choose a training set
It is trained, the optimization weighting parameter of acquisition input layer to the first hidden layer,；
S42：By training setThe first hidden layer is propagated through forward obtains fisrt feature, and as input
The second hidden layer of training, obtains the optimization weighting parameter of the first hidden layer to the second hidden layer,；
S43：By fisrt featureThe second hidden layer is propagated through forward obtains second feature, and as
It inputs and trains third hidden layer, the optimization weighting parameter of the second hidden layer of acquisition to third hidden layer,；
S44：By second featureThird hidden layer is propagated through forward obtains second feature, and as
Training output layer is inputted, optimization output weighting parameter is obtained；
S45：By the optimization weighting parameter of input layer to the first hidden layer,, optimization of first hidden layer to the second hidden layer
Weighting parameter,, the optimization weighting parameter of the second hidden layer to third hidden layer,And optimization output weighting parameter
As initiation parameter, by training setAs input, according to back-propagating principle provide the cost function of whole network with
Gradient adjusts weight parameter in known prediction label, and the weight for obtaining input layer to the first hidden layer of optimization is joined
Number,, the weighting parameter of the first hidden layer to the second hidden layer of optimization,, optimization
Weighting parameter of second hidden layer to third hidden layer,And the output weighting parameter optimized；
S46：By adjusting good weighting parameter deep neural network to test setIt is predicted, it is accurate to calculate prediction
Rate executes S45 and readjusts weighting parameter if predictablity rate is unsatisfactory for design requirement, if predictablity rate meets design
The deep neural network is then used for actual prediction by demand.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810658529.0A CN108846766A (en) | 2018-06-25 | 2018-06-25 | The self-service Claims Resolution success rate prediction technique of vehicle insurance based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810658529.0A CN108846766A (en) | 2018-06-25 | 2018-06-25 | The self-service Claims Resolution success rate prediction technique of vehicle insurance based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108846766A true CN108846766A (en) | 2018-11-20 |
Family
ID=64203727
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810658529.0A Pending CN108846766A (en) | 2018-06-25 | 2018-06-25 | The self-service Claims Resolution success rate prediction technique of vehicle insurance based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108846766A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111861765A (en) * | 2020-07-29 | 2020-10-30 | 贵州力创科技发展有限公司 | Intelligent anti-fraud method for vehicle insurance claim settlement |
| JP2021026384A (en) * | 2019-08-01 | 2021-02-22 | ジャパンモード株式会社 | Personal property insurance proposal program and personal property insurance condition inclusion possibility determination program |
| CN116342300A (en) * | 2023-05-26 | 2023-06-27 | 凯泰铭科技（北京）有限公司 | Method, device and equipment for analyzing characteristics of insurance claim settlement personnel |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106780052A (en) * | 2017-01-10 | 2017-05-31 | 上海诺悦智能科技有限公司 | Method and system are recommended in insurance service based on classification customer behavior analysis |
| CN107464186A (en) * | 2017-08-04 | 2017-12-12 | 缪骁 | Claims Resolution pattern based on vehicle insurance customer portrait recommends method |
| CN107610091A (en) * | 2017-07-31 | 2018-01-19 | 阿里巴巴集团控股有限公司 | Vehicle insurance image processing method, device, server and system |
| CN108153879A (en) * | 2017-12-26 | 2018-06-12 | 爱因互动科技发展（北京）有限公司 | The method and device of recommendation information is provided a user by human-computer interaction |
-
2018
- 2018-06-25 CN CN201810658529.0A patent/CN108846766A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106780052A (en) * | 2017-01-10 | 2017-05-31 | 上海诺悦智能科技有限公司 | Method and system are recommended in insurance service based on classification customer behavior analysis |
| CN107610091A (en) * | 2017-07-31 | 2018-01-19 | 阿里巴巴集团控股有限公司 | Vehicle insurance image processing method, device, server and system |
| CN107464186A (en) * | 2017-08-04 | 2017-12-12 | 缪骁 | Claims Resolution pattern based on vehicle insurance customer portrait recommends method |
| CN108153879A (en) * | 2017-12-26 | 2018-06-12 | 爱因互动科技发展（北京）有限公司 | The method and device of recommendation information is provided a user by human-computer interaction |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2021026384A (en) * | 2019-08-01 | 2021-02-22 | ジャパンモード株式会社 | Personal property insurance proposal program and personal property insurance condition inclusion possibility determination program |
| CN111861765A (en) * | 2020-07-29 | 2020-10-30 | 贵州力创科技发展有限公司 | Intelligent anti-fraud method for vehicle insurance claim settlement |
| CN116342300A (en) * | 2023-05-26 | 2023-06-27 | 凯泰铭科技（北京）有限公司 | Method, device and equipment for analyzing characteristics of insurance claim settlement personnel |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Bi et al. | Daily tourism volume forecasting for tourist attractions | |
| KR101692521B1 (en) | Apparatus for providing real estate information and sharing | |
| West et al. | A comparative analysis of neural networks and statistical methods for predicting consumer choice | |
| Fishburn | Utility theory | |
| WO2016091173A1 (en) | User maintenance system and method | |
| CN107735811A (en) | System, methods and procedures for real estate pricing | |
| KR102046501B1 (en) | Service providing apparatus and method for evaluating valuation and supporting upbringing based on diagnosis of start-up company | |
| EP2474945A1 (en) | Analyzing transactional data | |
| CN107251082A (en) | Information processor, information processing method and program | |
| US20210176262A1 (en) | Event monitoring and response system and method | |
| US11367142B1 (en) | Systems and methods for clustering data to forecast risk and other metrics | |
| US11367141B1 (en) | Systems and methods for forecasting loss metrics | |
| CN108846766A (en) | The self-service Claims Resolution success rate prediction technique of vehicle insurance based on deep learning | |
| US20190385080A1 (en) | K-lstm architecture for purchase prediction | |
| CN112131480A (en) | Personalized commodity recommendation method and system based on multilayer heterogeneous attribute network representation learning | |
| CN110175857B (en) | Method and device for determining optimal service | |
| CN117094781B (en) | Intelligent car insurance claim settlement processing method and system | |
| KR20190114581A (en) | Human resource management system using attendance information | |
| CN110472798A (en) | Prediction technique, device and the computer readable storage medium of time series data | |
| KR20220065735A (en) | Method and server for recommending personalized real estate information using the consumer choice model | |
| CN113781139A (en) | Item recommendation method, item recommendation device, equipment and medium | |
| Raza et al. | Performance analysis of deep approaches on airbnb sentiment reviews | |
| CN116308216B (en) | Digital platform system for enterprise service | |
| Aickelin et al. | Using simulation to incorporate dynamic criteria into multiple criteria decision-making | |
| Wah et al. | Predicting car purchase intent using data mining approach |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181120 |