CN110288096A - Prediction model training and prediction technique, device, electronic equipment, storage medium - Google Patents
Prediction model training and prediction technique, device, electronic equipment, storage medium Download PDFInfo
- Publication number
- CN110288096A CN110288096A CN201910583402.1A CN201910583402A CN110288096A CN 110288096 A CN110288096 A CN 110288096A CN 201910583402 A CN201910583402 A CN 201910583402A CN 110288096 A CN110288096 A CN 110288096A
- Authority
- CN
- China
- Prior art keywords
- layer
- prediction model
- training
- vehicle
- data
- 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
Links
Classifications
-
- 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
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/08—Insurance
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Finance (AREA)
- General Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- General Engineering & Computer Science (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Technology Law (AREA)
- General Business, Economics & Management (AREA)
- Traffic Control Systems (AREA)
Abstract
The present invention provides a kind of training of prediction model and prediction technique, device, electronic equipment, storage medium, driving safety risk forecast model training method includes: for M vehicle, obtain the data of being in danger of each vehicle, and multiple features of each vehicle are obtained according at least to dynamic data, vehicle data, Driver data and environmental data is driven, using multiple feature of each vehicle as the feature vector of the vehicle, the feature vector of the M vehicle forms sample set；Utilize the sample set and corresponding data training the first prediction model of first layer and the second prediction model of first layer of being in danger；When using sample set training the first prediction model of first layer and the second prediction model of first layer, the output of the second prediction model of output and first layer of the first prediction model of first layer and the corresponding data training training second layer prediction model that is in danger.Method and device provided by the invention realizes effective management to the safe control of the air control of driver and high-quality and non-prime driver.
Description
Technical field
The present invention relates to big data technical field more particularly to a kind of training of prediction model and prediction technique, device, electronics
Equipment, storage medium.
Background technique
Logistics has merged multiple industries such as Transportation Industry, warehousing industry and information industry as infrastructural industry for the national economy,
It is wide to be related to field.Wherein, the driving safety accident for being related to driver's life still needs to pay special attention to.The annual transportation incident in China
Total amount still occupies a high position, and truck causes toll on traffic to account for about one third.In recent years, China's highway was handed over
The ratio of the logical dead sum of death tolls occupied road traffic accident maintains 10% or so.Usually pass through vehicle-mounted OBD (On-
Board Diagnostic) equipment such as equipment or smart phone, vehicle operation data can be obtained by the sensor in equipment, from
And control intuitively can be carried out to vehicle running state.The logistics cargo allocation platform emerging for internet, how based on big number
According to the high non-prime driver of the high-quality driver and rate of being in danger that distinguish driving safety, and to the safe control of the air control of driver
It is also critical issue urgently to be solved.109636654 A of CN, which is disclosed, a kind of judges its Claims Resolution according to the relevant data of car owner
The system of risk class, with machine learning techniques, driving behavior and bus attribute to car owner, situation of insuring carry out risk point
The following Claims Resolution ratio is predicted in analysis.109767145 A of CN discloses a kind of driver's driving efficiency points-scoring system, contains perfect drive
The function of the person's of sailing driving efficiency evaluation and Improving advice, to remind the deficiency and Improving advice of this stroke of user.
In above scheme, 109636654 A of CN vehicle physical attribute, history are accepted insurance settle a claim information and break in traffic rules and regulations letter
Breath is used as independent variable, by random forests algorithm model, to obtain the simple loss ratio of objective function and the frequency that is in danger.Its disadvantage
It is only history static data, and lacks vehicle relevant to driving safety and drive dynamic data in real time.CN 109767145 A
Primarily directed to driver driving efficiency, it is whether environmentally friendly, whether safety and whether fuel-efficient comprehensive score.It is mainly in list
Driver is scored and suggested in secondary driving stroke, and is lacked to a kind of long drives behavior of driver and security risk
Observation.
Past judges that driving safety risk can only be come by some vehicles and driver's essential information and past Claims Resolution information
Simple judgement, so that the prior art is difficult to realize the effective of the safe control of air control to driver and high-quality and non-prime driver
Management.
Summary of the invention
The present invention in order to overcome defect existing for above-mentioned the relevant technologies, provide a kind of prediction model training and prediction technique,
Device, electronic equipment, storage medium, and then overcome the limitation and defect due to the relevant technologies at least to a certain extent and cause
One or more problem.
According to an aspect of the present invention, a kind of driving safety risk forecast model training method is provided, comprising:
For M vehicle, the data of being in danger of each vehicle are obtained, and according at least to driving dynamic data, vehicle data, driving
Member's data and environmental data obtain multiple features of each vehicle, using multiple feature of each vehicle as the feature of the vehicle
Vector, the feature vector of the M vehicle form sample set, and M is the integer greater than 2；
Utilize the sample set and corresponding be in danger data training the first prediction model of first layer and the prediction of first layer second
Model；
When using sample set training the first prediction model of first layer and the second prediction model of first layer, first layer first
The output of the second prediction model of output and first layer of prediction model and the corresponding data training training second layer prediction mould that is in danger
Type.
Optionally, described to utilize the sample set and corresponding data training the first prediction model of first layer and first that is in danger
Layer the second prediction model include:
The sample set is divided into the first training set X and the first test set T by preset ratio；
First training set is divided into N number of first training subset, N is the integer greater than 1；
Based on N number of first training subset and corresponding data training the first prediction model of first layer that is in danger, N number of the is obtained
One forecast set A；
The first test set T is inputted into housebroken first prediction model of first layer, obtains N number of second forecast set
B；
Based on N number of first training subset and corresponding data training the second prediction model of first layer that is in danger, N number of the is obtained
Three forecast set C；
The first test set T is inputted into housebroken second prediction model of first layer, obtains N number of 4th forecast set
D。
Optionally, described to utilize sample set training the first prediction model of first layer and the second prediction model of first layer
When, the output of the second prediction model of output and first layer of the first prediction model of first layer and the corresponding data training that is in danger are trained
Second layer prediction model includes:
N number of first forecast set A and N number of third forecast set C are formed into the second training set P；
The mean value of N number of second forecast set B and the mean value of N number of 4th forecast set D are formed into the second test set Q；
According to the second training set P, the second test set Q and the corresponding data training second layer prediction mould that is in danger
Type.
Optionally, described based on N number of first training subset training, first prediction model of first layer, it is pre- to obtain N number of first
Surveying collection A includes:
For every one first training subset Xi:
Utilize the first training subset X1... Xi-1, Xi+1... XN, training the first prediction model of first layer obtains quasi- first layer the
One prediction model Mod1i；
By the first training subset XiInput quasi- the first prediction model of first layer Mod1iObtain the first forecast set Ai。
Optionally, described that the first test set T is inputted into housebroken first prediction model of first layer, obtain N
A second forecast set B includes:
First test set T is inputted into N number of the first prediction model of quasi- first layer respectively to obtain N number of second forecast set.
Optionally, described based on N number of first training subset training, second prediction model of first layer, it is pre- to obtain N number of third
Surveying collection C includes:
For every one first training subset Xi:
Utilize the first training subset X1... Xi-1, Xi+1... XN, training the second prediction model of first layer obtains quasi- first layer the
One prediction model Mod2i；
By the first training subset XiInput quasi- the second prediction model of first layer Mod2iObtain third forecast set Ci。
Optionally, described that the first test set T is inputted into housebroken second prediction model of first layer, obtain N
A 4th forecast set D includes:
First test set T is inputted into N number of the second prediction model of quasi- first layer respectively to obtain N number of 4th forecast set.
Optionally, the driving dynamic data includes: vehicle location, car speed, vehicle acceleration, in vehicle angular speed
It is one or more；
The vehicle data includes: vehicle dimension, vehicle model, type of vehicle, one or more in vehicle performance；
The Driver data includes: driver's gender, driver's age, one or more in driver's driving age；
The environmental data includes: to drive weather, one or more in road information.
Optionally, multiple features of each vehicle are obtained also according to business datum,
The business datum includes order information and evaluation information, and the order information includes that order starts at the end of
Between, the starting point of order, terminal area in it is one or more, the evaluation information includes the evaluation that driver is obtained.
Optionally, first prediction model of first layer is extreme gradient lift scheme, and the first layer second predicts mould
Type is neural network model, and the second layer prediction model is Logic Regression Models.
According to another aspect of the invention, a kind of driving safety Risk Forecast Method is also provided, using driving as described above
Security risk prediction model training method is sailed, the driving safety Risk Forecast Method includes:
According at least to vehicle to be predicted driving dynamic data, vehicle data, Driver data and environmental data obtain to
Predict multiple features of vehicle；
Multiple features of vehicle to be predicted are inputted into housebroken the first prediction model of first layer, to obtain the first prediction
Value；
Multiple features of vehicle to be predicted are inputted into housebroken the second prediction model of first layer, to obtain the second prediction
Value；
First predicted value and second predicted value are inputted into housebroken second layer prediction model, to obtain to pre-
The risk profile value of measuring car.
According to another aspect of the invention, a kind of driving safety risk forecast model training device is also provided, comprising:
Sample constructs module, for obtaining the data of being in danger of each vehicle for M vehicle, and according at least to driving dynamic
Data, vehicle data, Driver data and environmental data obtain multiple features of each vehicle, by multiple spy of each vehicle
The feature vector as the vehicle is levied, the feature vector of the M vehicle forms sample set, and M is the integer greater than 2；
First training module utilizes the sample set and corresponding be in danger data training the first prediction model of first layer and the
One layer of second prediction model；
Second training module utilizes sample set training the first prediction model of first layer and the second prediction model of first layer
When, the output of the second prediction model of output and first layer of the first prediction model of first layer and the corresponding data training that is in danger are trained
Second layer prediction model.
According to another aspect of the invention, a kind of driving safety risk profile device is also provided, using driving as described above
Security risk prediction model training device is sailed, the driving safety risk profile device includes:
Feature construction module, for according at least to the driving dynamic data of vehicle to be predicted, vehicle data, Driver data
And environmental data obtains multiple features of vehicle to be predicted；
First prediction module predicts mould for multiple features of vehicle to be predicted to be inputted housebroken first layer first
Type, to obtain the first predicted value；
Second prediction module predicts mould for multiple features of vehicle to be predicted to be inputted housebroken first layer second
Type, to obtain the second predicted value；
First predicted value and second predicted value are inputted the housebroken second layer and predict mould by third prediction module
Type, to obtain the risk profile value of vehicle to be predicted.
According to another aspect of the invention, a kind of electronic equipment is also provided, the electronic equipment includes: processor；Storage
Medium, is stored thereon with computer program, and the computer program executes step as described above when being run by the processor.
According to another aspect of the invention, a kind of storage medium is also provided, computer journey is stored on the storage medium
Sequence, the computer program execute step as described above when being run by processor.
Compared with prior art, present invention has an advantage that
On the one hand, by big data obtain driving dynamic data, vehicle data, Driver data, environmental data and go out
Dangerous data are to prediction model and are trained, so that the prediction model can be predicted risk is driven, to realize to driving
Effective management of the safe control of the air control for the person of sailing and high-quality and non-prime driver；On the other hand, by providing two layers of prediction mould
Type, the input by the training output of two prediction models of first layer as second layer prediction model is to train the second layer to predict mould
Type, the prediction model being arranged as a result, by level (performances of the comprehensive multiple prediction models used) compare single prediction model
Predictablity rate it is higher；In another aspect, by the division of training set and test set, the division again of training set with multi-faceted instruction
Practice two prediction models of first layer, so that the prediction that training method in this way increases by two prediction models of first layer is accurate
Rate.
Detailed description of the invention
Its example embodiment is described in detail by referring to accompanying drawing, above and other feature of the invention and advantage will become
It is more obvious.
Fig. 1 shows the flow chart of driving safety risk forecast model training method according to an embodiment of the present invention.
Fig. 2 shows trained the first prediction models of first layer according to an embodiment of the present invention and the second prediction model of first time
Flow chart.
Fig. 3 shows the flow chart of trained second layer prediction model according to an embodiment of the present invention.
Fig. 4 to fig. 6 shows the schematic diagram of the first prediction model of trained first layer according to an embodiment of the present invention.
Fig. 7 shows the flow chart of driving safety risk profile mould method according to an embodiment of the present invention.
Fig. 8 shows the module map of driving safety risk forecast model training device according to an embodiment of the present invention.
Fig. 9 shows the module map of driving safety risk profile mold device according to an embodiment of the present invention.
Figure 10 schematically shows a kind of computer readable storage medium schematic diagram in exemplary embodiment of the present.
Figure 11 schematically shows a kind of electronic equipment schematic diagram in exemplary embodiment of the present.
Specific embodiment
Example embodiment is described more fully with reference to the drawings.However, example embodiment can be with a variety of shapes
Formula is implemented, and is not understood as limited to example set forth herein；On the contrary, thesing embodiments are provided so that the present invention will more
Fully and completely, and by the design of example embodiment comprehensively it is communicated to those skilled in the art.Described feature, knot
Structure or characteristic can be incorporated in any suitable manner in one or more embodiments.
In addition, attached drawing is only schematic illustrations of the invention, it is not necessarily drawn to scale.Identical attached drawing mark in figure
Note indicates same or similar part, thus will omit repetition thereof.Some block diagrams shown in the drawings are function
Energy entity, not necessarily must be corresponding with physically or logically independent entity.These function can be realized using software form
Energy entity, or these functional entitys are realized in one or more hardware modules or integrated circuit, or at heterogeneous networks and/or place
These functional entitys are realized in reason device device and/or microcontroller device.
Flow chart shown in the drawings is merely illustrative, it is not necessary to including all steps.For example, the step of having
It can also decompose, and the step of having can merge or part merges, therefore, the sequence actually executed is possible to according to the actual situation
Change.
Fig. 1 shows the flow chart of driving safety risk forecast model training method according to an embodiment of the present invention.It drives
Security risk prediction model training method includes the following steps:
Step S110: for M vehicle, the data of being in danger of each vehicle are obtained, and according at least to driving dynamic data, vehicle
Data, Driver data and environmental data obtain multiple features of each vehicle, using multiple feature of each vehicle as this
The feature vector of vehicle, the feature vector of the M vehicle form sample set, and M is the integer greater than 2.
Specifically, in each embodiment of the present invention, M is very big integer, so that the present invention can pass through
A large amount of data carry out the model training in subsequent step.
Specifically, data of being in danger include to be in danger to compensate information, whether apply for insurance reason for vehicle in particular time range
The data of compensation.Prediction model output when data of being in danger are as model training, vehicle application settlement of insurance claim (being defined as 1), vehicle
Apply settlement of insurance claim (being defined as 0), in actual prediction, prediction model can export the value between 0-1 as a result, to indicate vehicle
Apply for the probability of settlement of insurance claim.
In various embodiments of the present invention, the driving dynamic data may include vehicle location, car speed, vehicle
Acceleration, vehicle angular speed etc..Specifically, drive dynamic data can by vehicle-mounted monitoring module and monitoring cloud platform come
It obtains.Specifically, car-mounted terminal module obtains in real time drives dynamic data, monitoring cloud platform is for storing car-mounted terminal mould
The driving dynamic data that block obtains can obtain driving for M vehicle by monitoring cloud platform as a result, in step s 110
Sail dynamic data.
The vehicle data may include: vehicle dimension, vehicle model, type of vehicle, vehicle performance etc..Vehicle data is
Vehicle basic static information.Vehicle data can associatedly be stored in above-mentioned prison with vehicles identifications (license plate/distribution unique identification)
Cloud platform is controlled, or is stored in independent cloud platform.
The Driver data may include: driver's gender, driver's age, driver's driving age etc..Driver data
It can be associatedly stored in above-mentioned monitoring cloud platform with driver identification (such as driver's account), or be stored in independent cloud
In platform.Further, in above-mentioned cloud platform, the mapping between driver identification and vehicles identifications can also be stored
Table.
The environmental data may include driving weather, road information etc..Specifically, Weather information includes single stroke
In weather conditions, such as fine day, cloudy day, rainy day, road information includes condition of road surface in single stroke, such as hill path, high speed, city
City's road etc..
In some embodiments of the invention, multiple features of each vehicle can also be obtained according to business datum.It is described
Business datum may include order information and evaluation information.The order information may include order start with the end time, order
Single starting point, terminal area etc..The evaluation information may include the evaluation etc. that driver is obtained.
As a result, according to above-mentioned data, multiple features of available vehicle, for example, the acceleration times in the period, traveling
Highway total time, mileage, driver's favorable comment number, greasy weather drive number etc..System is not limited thereto in the setting of feature, this
Field technical staff can change the setting of features described above according to actual needs.
It further, further include that pretreated step is carried out to feature after obtaining features described above.Pretreatment to feature
It may include that interpolation supplement is carried out to missing values, abnormal point be removed.
In a specific embodiment, after pre-processing feature, can also include the steps that screening feature.Specifically
For, it can be screened in the following way, to each feature, calculate the correlation system between two features according to the following formula
Number:
Wherein, S, T are two selected features, and cov (S, T) indicates the covariance of feature S and T, σSAnd σTIndicate its standard
Difference.It, will if the relative coefficient between a feature and other all features is lower than threshold value (such as 0.001) in each feature
This feature is deleted.The calculation amount of system needed for reducing invention of the invention as a result, reduces system energy consumption.
As a result, by the feature retained after screening, that is, constitute sample set.In sample set can by a vehicle and its feature Lai
Form a sample.In the present embodiment, for obtaining the data of M vehicle, to include M sample in sample set.In some changes
Change in example, sample can also be formed by other forms, and (such as the feature of a stroke and the trip associated vehicle is one
The feature of sample, a driver and the relevant vehicle of the driver is sample etc.), the present invention is not limited thereto
System.
Step S120: the sample set and corresponding data training the first prediction model of first layer and the first layer of being in danger are utilized
Second prediction model.
Step S130: when using sample set training the first prediction model of first layer and the second prediction model of first layer,
The output of the second prediction model of output and first layer of the first prediction model of first layer and the corresponding data training training that is in danger
Two layers of prediction model.
In driving safety risk forecast model training method provided by the invention, on the one hand, obtained by big data
Dynamic data, vehicle data, Driver data, environmental data and data of being in danger are driven to prediction model and to be trained, so as to
The prediction model can be predicted risk is driven, to realize to the safe control of the air control of driver and high-quality and non-prime
Effective management of driver；On the other hand, by providing two layers of prediction model, the training of two prediction models of first layer is defeated
Out as the input of second layer prediction model to train second layer prediction model, the prediction model being arranged as a result, by level is (comprehensive
Close the performance of the multiple prediction models used) it is higher compared to the predictablity rate of single prediction model；In another aspect, passing through training
Collect and the division of test set, the division again of training set are with multi-faceted two prediction models of trained first layer, thus in this way
Training method increase by two prediction models of first layer predictablity rate.
In some embodiments of the invention, the specific implementation of above-mentioned steps S120 may refer to Fig. 2, and Fig. 2 shows roots
According to the flow chart of the first prediction model of training first layer and the second prediction model of first time of the embodiment of the present invention.Fig. 2 is shown altogether
Following steps:
Step S121: the sample set is divided into the first training set X and the first test set T by preset ratio.
For example, sample set is randomly selected the sample of L% as the first training set X, the sample of (100-L) % is as first
Test set T.L is preferably more than 50 less than 100.
Step S122: first training set is divided into N number of first training subset, N is the integer greater than 1.
By taking sample set includes M sample as an example, the sample number in the first training set X is L%*M, then in each training subset
Sample number be L%*M/N.Step S122 can also be divided by the way of stochastical sampling.
Step S123: training the first prediction model of first layer based on N number of first training subset and corresponding data of being in danger,
Obtain N number of first forecast set A.
Specifically, step S123 includes the following steps: for every one first training subset Xi: utilize the first training subset
X1... Xi-1, Xi+1... XN, quasi- the first prediction model of the first layer Mod of training the first prediction model of first layer acquisition1i；By the first instruction
Practice subset XiInput quasi- the first prediction model of first layer Mod1iObtain the first forecast set Ai。
Step S124: the first test set T is inputted into housebroken first prediction model of first layer, is obtained N number of
Second forecast set B.
Specifically, above-mentioned steps S124 includes that the first test set T is inputted to N number of the first prediction model of quasi- first layer respectively
To obtain N number of second forecast set.
The realization of above-mentioned steps S123 and step S124 may refer to fig. 4 to fig. 6.By taking N is equal to 3 as an example, totally three first
Training subset X1, X2, X3。
For the first training subset X1, utilize the first training subset X2, X3, training the first prediction model of first layer obtains quasi-
The first prediction model of first layer Mod11；By the first training subset X1Input quasi- the first prediction model of first layer Mod11It is pre- to obtain first
Survey collection A1.The first test set T is inputted into housebroken quasi- the first prediction model of first layer Mod11, obtain the second prediction
Collect B1。
For the first training subset X2, utilize the first training subset X1, X3, training the first prediction model of first layer obtains quasi-
The first prediction model of first layer Mod12；By the first training subset X1Input quasi- the first prediction model of first layer Mod12It is pre- to obtain first
Survey collection A2.The first test set T is inputted into housebroken quasi- the first prediction model of first layer Mod12, obtain the second prediction
Collect B2。
For the first training subset X3, utilize the first training subset X1, X2, training the first prediction model of first layer obtains quasi-
The first prediction model of first layer Mod13；By the first training subset X1Input quasi- the first prediction model of first layer Mod13It is pre- to obtain first
Survey collection A3.The first test set T is inputted into housebroken quasi- the first prediction model of first layer Mod13, obtain the second prediction
Collect B3。
Above is only the specific execution for schematically describing step S123 and step S124 provided by the invention, the present invention
System is not limited thereto.
Step S125: training the second prediction model of first layer based on N number of first training subset and corresponding data of being in danger,
Obtain N number of third forecast set C.
Specifically, the step S125 is similar with the training method of step S123.The step S125 includes: for every
One first training subset Xi: utilize the first training subset X1... Xi-1, Xi+1... XN, training the second prediction model of first layer acquisition
Quasi- the first prediction model of first layer Mod2i；By the first training subset XiInput quasi- the second prediction model of first layer Mod2iObtain third
Forecast set Ci。
Step S126: the first test set T is inputted into housebroken second prediction model of first layer, is obtained N number of
4th forecast set D.
Specifically, the step S126 may include: that the first test set T is inputted N number of quasi- first layer second in advance respectively
Model is surveyed to obtain N number of 4th forecast set.
In various embodiments of the present invention, first prediction model of first layer can be extreme gradient lift scheme
(XGBOOST), second prediction model of first layer can be neural network model (NN).Extreme gradient lift scheme
(XGBOOST) and neural network model (NN) is existing machine learning model.The first prediction model of first layer of the invention and
System is not limited thereto in the second prediction model of first layer.
In some embodiments of the invention, the specific implementation of above-mentioned steps S130 may refer to Fig. 3, and Fig. 3 shows root
According to the flow chart of the training second layer prediction model of the embodiment of the present invention.Fig. 3 is shown below step altogether:
Step S131: N number of first forecast set A and N number of third forecast set C are formed into the second training set P.
Step S132: by the mean value of N number of second forecast set B and the mean value of N number of 4th forecast set D composition second
Test set Q.
Step S133: according to the second training set P, the second test set Q and the corresponding data training second that is in danger
Layer prediction model.
Specifically, the second training set P and corresponding data of being in danger are for training the model in second layer prediction model to join
Number, the second test set Q are used to test the accuracy of housebroken second layer prediction model.
The second layer prediction model can be Logic Regression Models.It is minimized using loss function, to logistic regression
Model is trained.Loss function are as follows:
Wherein, m is sample size in the second training set P, x(i)For i-th of sample, y(i)For the corresponding number that is in danger of the sample
According to.
Above is only the training method for schematically describing second layer prediction model of the invention, and the present invention is not with this
For limitation.
Upper is only one or more specific implementations provided by the invention, and the present invention is not for limitation.
According to another aspect of the invention, a kind of driving safety Risk Forecast Method is also provided, as shown in fig. 7, Fig. 7 is shown
The flow chart of driving safety risk profile mould method according to an embodiment of the present invention.Driving safety risk provided by the invention is pre-
Survey method is using the driving safety risk forecast model training method as described in figure 1 above to Fig. 3.Fig. 7 is shown below step altogether:
Step S210: according at least to the driving dynamic data of vehicle to be predicted, vehicle data, Driver data and environment number
According to obtaining multiple features of vehicle to be predicted.
Drive the pretreatment mode and step of dynamic data, vehicle data, Driver data and environmental data and feature
The type of each data in S110 is similar with acquisition mode, and it will not be described here.
Step S220: inputting housebroken the first prediction model of first layer for multiple features of vehicle to be predicted, to obtain
First predicted value.
Step S230: inputting housebroken the second prediction model of first layer for multiple features of vehicle to be predicted, to obtain
Second predicted value.
Step S240: inputting housebroken second layer prediction model for first predicted value and second predicted value,
To obtain the risk profile value of vehicle to be predicted.
Specifically, the probability of risk profile value prediction application settlement of insurance claim.It can be by risk according to the risk profile value
Prediction be divided into security risk is high, in security risk, the low three classes of security risk, in order to manage and show, the present invention not with
This is limitation.
In driving safety Risk Forecast Method provided by the invention, on the one hand, the driving dynamic obtained by big data
Data, vehicle data, Driver data, environmental data and data of being in danger to prediction model and are trained, so as to the prediction mould
Type can be predicted risk is driven, to realize to the safe control of the air control of driver and high-quality and non-prime driver
Effectively management；On the other hand, by providing two layers of prediction model, by the training output of two prediction models of first layer as the
The input of two layers of prediction model is to train second layer prediction model, and by the prediction model of level setting, (synthesis is used as a result,
The performance of multiple prediction models) it is higher compared to the predictablity rate of single prediction model；In another aspect, passing through training set and test
The division of collection, the division again of training set are with multi-faceted two prediction models of trained first layer, thus training side in this way
The predictablity rate of formula increase by two prediction models of first layer.
According to another aspect of the invention, a kind of driving safety risk forecast model training device is also provided, Fig. 8 is shown
The module map of driving safety risk forecast model training device according to an embodiment of the present invention.Driving safety risk forecast model instruction
Practicing device 300 includes sample building module 310, the first training module 320 and the second training module 330.
Sample constructs module 310 and is used to obtain the data of being in danger of each vehicle, and dynamic according at least to driving for M vehicle
State data, vehicle data, Driver data and environmental data obtain multiple features of each vehicle, by the multiple of each vehicle
Feature vector of the feature as the vehicle, the feature vector of the M vehicle form sample set, and M is the integer greater than 2.
First training module 320 using the sample set and it is corresponding be in danger data training the first prediction model of first layer and
The second prediction model of first layer；
Second training module 330 predicts mould using sample set training the first prediction model of first layer and first layer second
When type, the output of the second prediction model of output and first layer of the first prediction model of first layer and corresponding data training instruction of being in danger
Practice second layer prediction model.
In driving safety risk forecast model training device provided by the invention, on the one hand, obtained by big data
Dynamic data, vehicle data, Driver data, environmental data and data of being in danger are driven to prediction model and to be trained, so as to
The prediction model can be predicted risk is driven, to realize to the safe control of the air control of driver and high-quality and non-prime
Effective management of driver；On the other hand, by providing two layers of prediction model, the training of two prediction models of first layer is defeated
Out as the input of second layer prediction model to train second layer prediction model, the prediction model being arranged as a result, by level is (comprehensive
Close the performance of the multiple prediction models used) it is higher compared to the predictablity rate of single prediction model；In another aspect, passing through training
Collect and the division of test set, the division again of training set are with multi-faceted two prediction models of trained first layer, thus in this way
Training method increase by two prediction models of first layer predictablity rate.
Fig. 8 is only to show schematically driving safety risk forecast model training device 300 provided by the invention, not
Under the premise of violating present inventive concept, the fractionation of module, increases all within protection scope of the present invention merging.The present invention mentions
The driving safety risk forecast model training device 300 of confession can be by software, hardware, firmware, plug-in unit and any between them
To realize, the present invention is not limited thereto for combination.
According to another aspect of the invention, a kind of driving safety risk profile device is also provided, Fig. 9 is shown according to this hair
The module map of the driving safety risk profile device of bright embodiment.Driving safety risk profile device 400 includes feature construction mould
Block 410, the first prediction module 420, the second prediction module 430 and third prediction module 440.
Feature construction module 410 is used for driving dynamic data, vehicle data, driver's number according at least to vehicle to be predicted
According to and environmental data obtain multiple features of vehicle to be predicted；
First prediction module 420, which is used to multiple features of vehicle to be predicted inputting housebroken first layer first, predicts mould
Type, to obtain the first predicted value；
Second prediction module 430, which is used to multiple features of vehicle to be predicted inputting housebroken first layer second, predicts mould
Type, to obtain the second predicted value；
First predicted value and second predicted value are inputted the housebroken second layer and predicted by third prediction module 440
Model, to obtain the risk profile value of vehicle to be predicted.
In driving safety risk profile device provided by the invention, on the one hand, the driving dynamic obtained by big data
Data, vehicle data, Driver data, environmental data and data of being in danger to prediction model and are trained, so as to the prediction mould
Type can be predicted risk is driven, to realize to the safe control of the air control of driver and high-quality and non-prime driver
Effectively management；On the other hand, by providing two layers of prediction model, by the training output of two prediction models of first layer as the
The input of two layers of prediction model is to train second layer prediction model, and by the prediction model of level setting, (synthesis is used as a result,
The performance of multiple prediction models) it is higher compared to the predictablity rate of single prediction model；In another aspect, passing through training set and test
The division of collection, the division again of training set are with multi-faceted two prediction models of trained first layer, thus training side in this way
The predictablity rate of formula increase by two prediction models of first layer.
Fig. 9 is only to show schematically driving safety risk profile device 400 provided by the invention, without prejudice to this hair
Under the premise of bright design, the fractionation of module, increases all within protection scope of the present invention merging.Driving provided by the invention
Security risk prediction meanss 400 can realize by software, hardware, firmware, plug-in unit and any combination between them, the present invention
It is not limited thereto.
In an exemplary embodiment of the present invention, a kind of computer readable storage medium is additionally provided, meter is stored thereon with
Driving safety risk described in any one above-mentioned embodiment may be implemented when being executed by such as processor in calculation machine program, the program
The step of prediction model training method and/or driving safety Risk Forecast Method.In some possible embodiments, of the invention
Various aspects be also implemented as the form of program product a kind of comprising program code, when described program product is in terminal
When running in equipment, said program code is for making the terminal device execute the above-mentioned driving safety risk profile mould of this specification
Various illustrative embodiments according to the present invention described in type training method and/or driving safety Risk Forecast Method part
Step.
Refering to what is shown in Fig. 10, describing the program product for realizing the above method of embodiment according to the present invention
700, can using portable compact disc read only memory (CD-ROM) and including program code, and can in terminal device,
Such as it is run on PC.However, program product of the invention is without being limited thereto, in this document, readable storage medium storing program for executing can be with
To be any include or the tangible medium of storage program, the program can be commanded execution system, device or device use or
It is in connection.
Described program product can be using any combination of one or more readable mediums.Readable medium can be readable letter
Number medium or readable storage medium storing program for executing.Readable storage medium storing program for executing for example can be but be not limited to electricity, magnetic, optical, electromagnetic, infrared ray or
System, device or the device of semiconductor, or any above combination.The more specific example of readable storage medium storing program for executing is (non exhaustive
List) include: electrical connection with one or more conducting wires, portable disc, hard disk, random access memory (RAM), read-only
Memory (ROM), erasable programmable read only memory (EPROM or flash memory), optical fiber, portable compact disc read only memory
(CD-ROM), light storage device, magnetic memory device or above-mentioned any appropriate combination.
The computer readable storage medium may include in a base band or the data as the propagation of carrier wave a part are believed
Number, wherein carrying readable program code.The data-signal of this propagation can take various forms, including but not limited to electromagnetism
Signal, optical signal or above-mentioned any appropriate combination.Readable storage medium storing program for executing can also be any other than readable storage medium storing program for executing
Readable medium, the readable medium can send, propagate or transmit for by instruction execution system, device or device use or
Person's program in connection.The program code for including on readable storage medium storing program for executing can transmit with any suitable medium, packet
Include but be not limited to wireless, wired, optical cable, RF etc. or above-mentioned any appropriate combination.
The program for executing operation of the present invention can be write with any combination of one or more programming languages
Code, described program design language include object oriented program language-Java, C++ etc., further include conventional
Procedural programming language-such as " C " language or similar programming language.Program code can be fully in tenant
It calculates and executes in equipment, partly executed in tenant's equipment, being executed as an independent software package, partially in tenant's calculating
Upper side point is executed on a remote computing or is executed in remote computing device or server completely.It is being related to far
Journey calculates in the situation of equipment, and remote computing device can pass through the network of any kind, including local area network (LAN) or wide area network
(WAN), it is connected to tenant and calculates equipment, or, it may be connected to external computing device (such as utilize ISP
To be connected by internet).
In an exemplary embodiment of the present invention, a kind of electronic equipment is also provided, which may include processor,
And the memory of the executable instruction for storing the processor.Wherein, the processor is configured to via described in execution
Executable instruction executes driving safety risk forecast model training method and/or driving described in any one above-mentioned embodiment
The step of security risk prediction technique.
Person of ordinary skill in the field it is understood that various aspects of the invention can be implemented as system, method or
Program product.Therefore, various aspects of the invention can be embodied in the following forms, it may be assumed that complete hardware embodiment, complete
The embodiment combined in terms of full Software Implementation (including firmware, microcode etc.) or hardware and software, can unite here
Referred to as circuit, " module " or " system ".
The electronic equipment 500 of this embodiment according to the present invention is described referring to Figure 11.The electricity that Figure 11 is shown
Sub- equipment 500 is only an example, should not function to the embodiment of the present invention and use scope bring any restrictions.
As shown in figure 11, electronic equipment 500 is showed in the form of universal computing device.The component of electronic equipment 500 can be with
Including but not limited to: at least one processing unit 510, at least one storage unit 520, the different system components of connection (including are deposited
Storage unit 520 and processing unit 510) bus 530, display unit 540 etc..
Wherein, the storage unit is stored with program code, and said program code can be held by the processing unit 510
Row, so that the processing unit 510 executes the above-mentioned driving safety risk forecast model training method of this specification and/or drives peace
Described in full Risk Forecast Method part according to the present invention various illustrative embodiments the step of.For example, the processing is single
Member 510 can execute step as shown in Figure 1 to Figure 2.
The storage unit 520 may include the readable medium of volatile memory cell form, such as random access memory
Unit (RAM) 5201 and/or cache memory unit 5202 can further include read-only memory unit (ROM) 5203.
The storage unit 520 can also include program/practical work with one group of (at least one) program module 5205
Tool 5204, such program module 5205 includes but is not limited to: operating system, one or more application program, other programs
It may include the realization of network environment in module and program data, each of these examples or certain combination.
Bus 530 can be to indicate one of a few class bus structures or a variety of, including storage unit bus or storage
Cell controller, peripheral bus, graphics acceleration port, processing unit use any bus structures in a variety of bus structures
Local bus.
Electronic equipment 500 can also be with one or more external equipments 600 (such as keyboard, sensing equipment, bluetooth equipment
Deng) communication, the equipment that also tenant can be enabled interact with the electronic equipment 500 with one or more communicates, and/or with make
Any equipment (such as the router, modulation /demodulation that the electronic equipment 500 can be communicated with one or more of the other calculating equipment
Device etc.) communication.This communication can be carried out by input/output (I/O) interface 550.Also, electronic equipment 500 can be with
By network adapter 560 and one or more network (such as local area network (LAN), wide area network (WAN) and/or public network,
Such as internet) communication.Network adapter 560 can be communicated by bus 530 with other modules of electronic equipment 500.It should
Understand, although not shown in the drawings, other hardware and/or software module can be used in conjunction with electronic equipment 500, including but unlimited
In: microcode, device driver, redundant processing unit, external disk drive array, RAID system, tape drive and number
According to backup storage system etc..
Through the above description of the embodiments, those skilled in the art is it can be readily appreciated that example described herein is implemented
Mode can also be realized by software realization in such a way that software is in conjunction with necessary hardware.Therefore, according to the present invention
The technical solution of embodiment can be embodied in the form of software products, which can store non-volatile at one
Property storage medium (can be CD-ROM, USB flash disk, mobile hard disk etc.) in or network on, including some instructions are so that a calculating
Equipment (can be personal computer, server or network equipment etc.) executes the above-mentioned driving of embodiment according to the present invention
Security risk prediction model training method and/or driving safety Risk Forecast Method.
Compared with prior art, present invention has an advantage that
On the one hand, by big data obtain driving dynamic data, vehicle data, Driver data, environmental data and go out
Dangerous data are to prediction model and are trained, so that the prediction model can be predicted risk is driven, to realize to driving
Effective management of the safe control of the air control for the person of sailing and high-quality and non-prime driver；On the other hand, by providing two layers of prediction mould
Type, the input by the training output of two prediction models of first layer as second layer prediction model is to train the second layer to predict mould
Type, the prediction model being arranged as a result, by level (performances of the comprehensive multiple prediction models used) compare single prediction model
Predictablity rate it is higher；In another aspect, by the division of training set and test set, the division again of training set with multi-faceted instruction
Practice two prediction models of first layer, so that the prediction that training method in this way increases by two prediction models of first layer is accurate
Rate.
Those skilled in the art after considering the specification and implementing the invention disclosed here, will readily occur to of the invention its
Its embodiment.This application is intended to cover any variations, uses, or adaptations of the invention, these modifications, purposes or
Person's adaptive change follows general principle of the invention and including the undocumented common knowledge in the art of the present invention
Or conventional techniques.The description and examples are only to be considered as illustrative, and true scope and spirit of the invention are by appended
Claim is pointed out.
Claims (15)
1. a kind of driving safety risk forecast model training method characterized by comprising
For M vehicle, the data of being in danger of each vehicle are obtained, and according at least to driving dynamic data, vehicle data, driver's number
According to and environmental data obtain multiple features of each vehicle, using multiple feature of each vehicle as the feature of the vehicle to
Amount, the feature vector of the M vehicle form sample set, and M is the integer greater than 2；
Utilize the sample set and corresponding data training the first prediction model of first layer and the second prediction model of first layer of being in danger；
When using sample set training the first prediction model of first layer and the second prediction model of first layer, first layer first is predicted
The output of the second prediction model of output and first layer of model and the corresponding data training training second layer prediction model that is in danger.
2. driving safety risk forecast model training method as described in claim 1, which is characterized in that described to utilize the sample
This collection and trained the first prediction model of first layer of corresponding data of being in danger and the second prediction model of first layer include:
The sample set is divided into the first training set X and the first test set T by preset ratio；
First training set is divided into N number of first training subset, N is the integer greater than 1；
Based on N number of first training subset and corresponding data training the first prediction model of first layer that is in danger, it is pre- to obtain N number of first
Survey collection A；
The first test set T is inputted into housebroken first prediction model of first layer, obtains N number of second forecast set B；
Based on N number of first training subset and corresponding data training the second prediction model of first layer that is in danger, it is pre- to obtain N number of third
Survey collection C；
The first test set T is inputted into housebroken second prediction model of first layer, obtains N number of 4th forecast set D.
3. driving safety risk forecast model training method as claimed in claim 2, which is characterized in that described to utilize the sample
This collection training the first prediction model of first layer and when the second prediction model of first layer, the output of the first prediction model of first layer and the
The output of one layer of second prediction model and the corresponding data training training second layer prediction model that is in danger include:
N number of first forecast set A and N number of third forecast set C are formed into the second training set P；
The mean value of N number of second forecast set B and the mean value of N number of 4th forecast set D are formed into the second test set Q；
According to the second training set P, the second test set Q and the corresponding data training second layer prediction model that is in danger.
4. driving safety risk forecast model training method as claimed in claim 2, which is characterized in that described N number of based on this
First training subset trains the first prediction model of first layer, obtains N number of first forecast set A and includes:
For every one first training subset Xi:
Utilize the first training subset X1... Xi-1, Xi+1... XN, it is pre- that training the first prediction model of first layer obtains quasi- first layer first
Survey model M od1i；
By the first training subset XiInput quasi- the first prediction model of first layer Mod1iObtain the first forecast set Ai。
5. driving safety risk forecast model training method as claimed in claim 4, which is characterized in that described by described first
Test set T inputs housebroken first prediction model of first layer, obtains N number of second forecast set B and includes:
First test set T is inputted into N number of the first prediction model of quasi- first layer respectively to obtain N number of second forecast set.
6. driving safety risk forecast model training method as claimed in claim 2, which is characterized in that described N number of based on this
First training subset trains the second prediction model of first layer, and obtaining N number of third forecast set C includes:
For every one first training subset Xi:
Utilize the first training subset X1... Xi-1, Xi+1... XN, it is pre- that training the second prediction model of first layer obtains quasi- first layer first
Survey model M od2i；
By the first training subset XiInput quasi- the second prediction model of first layer Mod2iObtain third forecast set Ci。
7. driving safety risk forecast model training method as claimed in claim 6, which is characterized in that described by described first
Test set T inputs housebroken second prediction model of first layer, obtains N number of 4th forecast set D and includes:
First test set T is inputted into N number of the second prediction model of quasi- first layer respectively to obtain N number of 4th forecast set.
8. driving safety risk forecast model training method as described in any one of claim 1 to 7, which is characterized in that
The dynamic data that drives includes: vehicle location, car speed, vehicle acceleration, one or more in vehicle angular speed
?；
The vehicle data includes: vehicle dimension, vehicle model, type of vehicle, one or more in vehicle performance；
The Driver data includes: driver's gender, driver's age, one or more in driver's driving age；
The environmental data includes: to drive weather, one or more in road information.
9. driving safety risk forecast model training method as described in any one of claim 1 to 7, which is characterized in that go back root
Multiple features of each vehicle are obtained according to business datum,
The business datum includes order information and evaluation information, the order information include order start with the end time, order
One or more during the starting point of list, terminal are regional, the evaluation information includes the evaluation that driver is obtained.
10. driving safety risk forecast model training method as described in any one of claim 1 to 7, which is characterized in that described
The first prediction model of first layer is extreme gradient lift scheme, and second prediction model of first layer is neural network model, institute
Stating second layer prediction model is Logic Regression Models.
11. a kind of driving safety Risk Forecast Method, which is characterized in that driven using as described in any one of claim 1 to 10
Security risk prediction model training method is sailed, the driving safety Risk Forecast Method includes:
According at least to vehicle to be predicted driving dynamic data, vehicle data, Driver data and environmental data obtain it is to be predicted
Multiple features of vehicle；
Multiple features of vehicle to be predicted are inputted into housebroken the first prediction model of first layer, to obtain the first predicted value；
Multiple features of vehicle to be predicted are inputted into housebroken the second prediction model of first layer, to obtain the second predicted value；
First predicted value and second predicted value are inputted into housebroken second layer prediction model, to obtain vehicle to be predicted
Risk profile value.
12. a kind of driving safety risk forecast model training device characterized by comprising
Sample constructs module, for for M vehicle, obtaining the data of being in danger of each vehicle, and according at least to driving dynamic data,
Vehicle data, Driver data and environmental data obtain multiple features of each vehicle, and multiple feature of each vehicle is made
For the feature vector of the vehicle, the feature vector of the M vehicle forms sample set, and M is the integer greater than 2；
First training module utilizes the sample set and corresponding data training the first prediction model of first layer and the first layer of being in danger
Second prediction model；
Second training module, when using sample set training the first prediction model of first layer and the second prediction model of first layer,
The output of the second prediction model of output and first layer of the first prediction model of first layer and the corresponding data training training that is in danger
Two layers of prediction model.
13. a kind of driving safety risk profile device, which is characterized in that use driving safety risk as claimed in claim 11
Prediction model training device, the driving safety risk profile device include:
Feature construction module, for driving dynamic data, vehicle data, Driver data and the ring according at least to vehicle to be predicted
Border data obtain multiple features of vehicle to be predicted；
First prediction module, for multiple features of vehicle to be predicted to be inputted housebroken the first prediction model of first layer, with
Obtain the first predicted value；
Second prediction module, for multiple features of vehicle to be predicted to be inputted housebroken the second prediction model of first layer, with
Obtain the second predicted value；
First predicted value and second predicted value are inputted housebroken second layer prediction model by third prediction module,
To obtain the risk profile value of vehicle to be predicted.
14. a kind of electronic equipment, which is characterized in that the electronic equipment includes:
Processor；
Memory is stored thereon with computer program, is executed when the computer program is run by the processor as right is wanted
It seeks 1 to 10 described in any item driving safety risk forecast model training methods or executes as claimed in claim 11 drive
Security risk prediction technique.
15. a kind of storage medium, which is characterized in that be stored with computer program, the computer program on the storage medium
Driving safety risk forecast model training method as described in any one of claim 1 to 10 is executed when being run by the processor
Or execute driving safety Risk Forecast Method as claimed in claim 11.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910583402.1A CN110288096B (en) | 2019-06-28 | 2019-06-28 | Prediction model training method, prediction model training device, prediction model prediction method, prediction model prediction device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910583402.1A CN110288096B (en) | 2019-06-28 | 2019-06-28 | Prediction model training method, prediction model training device, prediction model prediction method, prediction model prediction device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110288096A true CN110288096A (en) | 2019-09-27 |
| CN110288096B CN110288096B (en) | 2021-06-08 |
Family
ID=68020315
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910583402.1A Active CN110288096B (en) | 2019-06-28 | 2019-06-28 | Prediction model training method, prediction model training device, prediction model prediction method, prediction model prediction device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110288096B (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110826640A (en) * | 2019-11-12 | 2020-02-21 | 郑州轻工业学院 | Unsupervised sensor array interference characteristic removing method |
| CN110826820A (en) * | 2019-11-29 | 2020-02-21 | 北京世纪高通科技有限公司 | Method and device for predicting vehicle insurance case quantity |
| CN110942248A (en) * | 2019-11-26 | 2020-03-31 | 支付宝(杭州)信息技术有限公司 | Training method and device for transaction wind control network and transaction risk detection method |
| CN111191957A (en) * | 2020-01-07 | 2020-05-22 | 深圳广联赛讯有限公司 | UBI scoring method, UBI scoring device and readable storage medium |
| CN111523560A (en) * | 2020-03-18 | 2020-08-11 | 第四范式（北京）技术有限公司 | Training method, prediction method, device and system for number prediction model of arriving trucks |
| CN111626879A (en) * | 2020-05-28 | 2020-09-04 | 泰康保险集团股份有限公司 | Data prediction method, device, equipment and storage medium |
| CN111959518A (en) * | 2020-08-14 | 2020-11-20 | 北京嘀嘀无限科技发展有限公司 | Data processing method, device and equipment |
| CN112329886A (en) * | 2020-11-26 | 2021-02-05 | 珠海大横琴科技发展有限公司 | Double-license plate recognition method, model training method, device, equipment and storage medium |
| CN112508228A (en) * | 2020-11-03 | 2021-03-16 | 北京理工大学前沿技术研究院 | Driving behavior risk prediction method and system |
| CN112763045A (en) * | 2019-11-06 | 2021-05-07 | 东北大学秦皇岛分校 | Vehicle self-load detection cloud calibration prediction method |
| CN112906139A (en) * | 2021-04-08 | 2021-06-04 | 平安科技（深圳）有限公司 | Vehicle fault risk assessment method and device, electronic equipment and storage medium |
| CN112948749A (en) * | 2021-03-02 | 2021-06-11 | 北京交通大学 | System and method for identifying and predicting risk factors of full-chain logistics |
| WO2021121252A1 (en) * | 2019-12-17 | 2021-06-24 | Beijing Didi Infinity Technology And Development Co., Ltd. | Comment-based behavior prediction |
| CN114299633A (en) * | 2021-12-28 | 2022-04-08 | 中国电信股份有限公司 | Automobile driving accident prediction method and device, electronic equipment and storage medium |
Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105139648A (en) * | 2015-08-11 | 2015-12-09 | 北京奇虎科技有限公司 | Driving habit data generation method and system |
| CN106448215A (en) * | 2016-08-18 | 2017-02-22 | 深圳市永兴元科技股份有限公司 | Vehicle insurance claiming warning method and vehicle insurance claiming warning device |
| CN106651162A (en) * | 2016-12-09 | 2017-05-10 | 思建科技有限公司 | Big data-based driving risk assessment method |
| CN106952002A (en) * | 2017-04-05 | 2017-07-14 | 南京人人保网络技术有限公司 | Driving methods of risk assessment and device based on driving behavior |
| CN106980911A (en) * | 2017-04-05 | 2017-07-25 | 南京人人保网络技术有限公司 | Driving methods of risk assessment and device based on the static factor |
| CN108256714A (en) * | 2016-12-29 | 2018-07-06 | 得道车联网络科技（上海）有限公司 | A kind of wheelpath real-time risk assessment model based on car networking big data |
| CN108764517A (en) * | 2018-04-08 | 2018-11-06 | 中南大学 | A kind of blast furnace molten iron silicon content trend method, equipment and storage medium |
| CN108961086A (en) * | 2018-07-09 | 2018-12-07 | 江苏满运软件科技有限公司 | A kind of early warning analysis method and system of vehicle insurance fraud |
| CN109064002A (en) * | 2018-07-26 | 2018-12-21 | 阿里巴巴集团控股有限公司 | Vehicle risk appraisal procedure, device and equipment |
| CN109118055A (en) * | 2018-07-19 | 2019-01-01 | 众安信息技术服务有限公司 | A kind of driving behavior methods of marking and device |
| CN109242044A (en) * | 2018-09-30 | 2019-01-18 | 江苏满运软件科技有限公司 | Training method, device, storage medium and the electronic equipment of vehicle and goods matching model |
| CN109272165A (en) * | 2018-09-30 | 2019-01-25 | 江苏满运软件科技有限公司 | Register probability predictor method, device, storage medium and electronic equipment |
| CN109472462A (en) * | 2018-10-18 | 2019-03-15 | 中山大学 | A kind of project risk ranking method and device based on the fusion of multi-model storehouse |
| CN109635990A (en) * | 2018-10-12 | 2019-04-16 | 阿里巴巴集团控股有限公司 | A kind of training method, prediction technique, device and electronic equipment |
| CN109658272A (en) * | 2018-12-26 | 2019-04-19 | 江苏数慧信息科技有限公司 | Driving behavior evaluation system and Insurance Pricing system based on driving behavior |
| CN109685273A (en) * | 2018-12-25 | 2019-04-26 | 江苏满运软件科技有限公司 | Brake failure method for early warning, device, electronic equipment, the storage medium of vehicle |
| CN109840660A (en) * | 2017-11-29 | 2019-06-04 | 北京四维图新科技股份有限公司 | A kind of vehicular characteristics data processing method and vehicle risk prediction model training method |
-
2019
- 2019-06-28 CN CN201910583402.1A patent/CN110288096B/en active Active
Patent Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105139648A (en) * | 2015-08-11 | 2015-12-09 | 北京奇虎科技有限公司 | Driving habit data generation method and system |
| CN106448215A (en) * | 2016-08-18 | 2017-02-22 | 深圳市永兴元科技股份有限公司 | Vehicle insurance claiming warning method and vehicle insurance claiming warning device |
| CN106651162A (en) * | 2016-12-09 | 2017-05-10 | 思建科技有限公司 | Big data-based driving risk assessment method |
| CN108256714A (en) * | 2016-12-29 | 2018-07-06 | 得道车联网络科技（上海）有限公司 | A kind of wheelpath real-time risk assessment model based on car networking big data |
| CN106952002A (en) * | 2017-04-05 | 2017-07-14 | 南京人人保网络技术有限公司 | Driving methods of risk assessment and device based on driving behavior |
| CN106980911A (en) * | 2017-04-05 | 2017-07-25 | 南京人人保网络技术有限公司 | Driving methods of risk assessment and device based on the static factor |
| CN109840660A (en) * | 2017-11-29 | 2019-06-04 | 北京四维图新科技股份有限公司 | A kind of vehicular characteristics data processing method and vehicle risk prediction model training method |
| CN108764517A (en) * | 2018-04-08 | 2018-11-06 | 中南大学 | A kind of blast furnace molten iron silicon content trend method, equipment and storage medium |
| CN108961086A (en) * | 2018-07-09 | 2018-12-07 | 江苏满运软件科技有限公司 | A kind of early warning analysis method and system of vehicle insurance fraud |
| CN109118055A (en) * | 2018-07-19 | 2019-01-01 | 众安信息技术服务有限公司 | A kind of driving behavior methods of marking and device |
| CN109064002A (en) * | 2018-07-26 | 2018-12-21 | 阿里巴巴集团控股有限公司 | Vehicle risk appraisal procedure, device and equipment |
| CN109272165A (en) * | 2018-09-30 | 2019-01-25 | 江苏满运软件科技有限公司 | Register probability predictor method, device, storage medium and electronic equipment |
| CN109242044A (en) * | 2018-09-30 | 2019-01-18 | 江苏满运软件科技有限公司 | Training method, device, storage medium and the electronic equipment of vehicle and goods matching model |
| CN109635990A (en) * | 2018-10-12 | 2019-04-16 | 阿里巴巴集团控股有限公司 | A kind of training method, prediction technique, device and electronic equipment |
| CN109472462A (en) * | 2018-10-18 | 2019-03-15 | 中山大学 | A kind of project risk ranking method and device based on the fusion of multi-model storehouse |
| CN109685273A (en) * | 2018-12-25 | 2019-04-26 | 江苏满运软件科技有限公司 | Brake failure method for early warning, device, electronic equipment, the storage medium of vehicle |
| CN109658272A (en) * | 2018-12-26 | 2019-04-19 | 江苏数慧信息科技有限公司 | Driving behavior evaluation system and Insurance Pricing system based on driving behavior |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112763045A (en) * | 2019-11-06 | 2021-05-07 | 东北大学秦皇岛分校 | Vehicle self-load detection cloud calibration prediction method |
| CN110826640A (en) * | 2019-11-12 | 2020-02-21 | 郑州轻工业学院 | Unsupervised sensor array interference characteristic removing method |
| CN110826640B (en) * | 2019-11-12 | 2022-07-05 | 郑州轻工业学院 | Unsupervised sensor array interference characteristic removing method |
| CN110942248A (en) * | 2019-11-26 | 2020-03-31 | 支付宝(杭州)信息技术有限公司 | Training method and device for transaction wind control network and transaction risk detection method |
| CN110942248B (en) * | 2019-11-26 | 2022-05-31 | 支付宝(杭州)信息技术有限公司 | Training method and device for transaction wind control network and transaction risk detection method |
| CN110826820A (en) * | 2019-11-29 | 2020-02-21 | 北京世纪高通科技有限公司 | Method and device for predicting vehicle insurance case quantity |
| WO2021121252A1 (en) * | 2019-12-17 | 2021-06-24 | Beijing Didi Infinity Technology And Development Co., Ltd. | Comment-based behavior prediction |
| CN111191957A (en) * | 2020-01-07 | 2020-05-22 | 深圳广联赛讯有限公司 | UBI scoring method, UBI scoring device and readable storage medium |
| CN111191957B (en) * | 2020-01-07 | 2023-09-22 | 深圳广联赛讯股份有限公司 | UBI scoring method, UBI scoring device and readable storage medium |
| CN111523560A (en) * | 2020-03-18 | 2020-08-11 | 第四范式（北京）技术有限公司 | Training method, prediction method, device and system for number prediction model of arriving trucks |
| CN111626879A (en) * | 2020-05-28 | 2020-09-04 | 泰康保险集团股份有限公司 | Data prediction method, device, equipment and storage medium |
| CN111959518A (en) * | 2020-08-14 | 2020-11-20 | 北京嘀嘀无限科技发展有限公司 | Data processing method, device and equipment |
| CN112508228A (en) * | 2020-11-03 | 2021-03-16 | 北京理工大学前沿技术研究院 | Driving behavior risk prediction method and system |
| CN112329886A (en) * | 2020-11-26 | 2021-02-05 | 珠海大横琴科技发展有限公司 | Double-license plate recognition method, model training method, device, equipment and storage medium |
| CN112948749A (en) * | 2021-03-02 | 2021-06-11 | 北京交通大学 | System and method for identifying and predicting risk factors of full-chain logistics |
| CN112906139A (en) * | 2021-04-08 | 2021-06-04 | 平安科技（深圳）有限公司 | Vehicle fault risk assessment method and device, electronic equipment and storage medium |
| CN114299633A (en) * | 2021-12-28 | 2022-04-08 | 中国电信股份有限公司 | Automobile driving accident prediction method and device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110288096B (en) | 2021-06-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110288096A (en) | Prediction model training and prediction technique, device, electronic equipment, storage medium | |
| CN107784587B (en) | Driving behavior evaluation system | |
| CN109155104B (en) | System and method for recommending estimated arrival time | |
| Zhao et al. | Development of a Representative EV Urban Driving Cycle Based on ak‐Means and SVM Hybrid Clustering Algorithm | |
| JP2021168193A (en) | Intelligent self-adaptive automobile device for measurement and integration of risk based on dynamic score using telematics connection search engine and corresponding method thereof | |
| CN103364006B (en) | For determining the system and method for vehicle route | |
| CN109109784A (en) | The system and method for vehicle residue driving information are provided according to user behavior | |
| CN109844793A (en) | Based on intelligence, the OEM route assembly system of telematics and its corresponding method, real-time risk measurement/risk score for bi-directional scaling drives signaling | |
| CN107784251A (en) | The method evaluated based on image recognition technology driving behavior | |
| CN105868787A (en) | Electric car driving range evaluation method based on working condition identification and fuzzy energy consumption | |
| CN110386145A (en) | A kind of real-time forecasting system of target driver driving behavior | |
| CN107346565A (en) | A kind of vehicle data processing method, device and terminal device | |
| CN110371132A (en) | Driver's adapter tube appraisal procedure and device | |
| CN106781697A (en) | Vehicular adverse weather real-time perception and anticollision method for early warning | |
| CN109543923A (en) | Vehicle queue time forecasting methods, system, equipment and storage medium | |
| Ushakov et al. | Artificial Intelligence as a factor of public transportations system development | |
| CN110379165A (en) | A kind of road type prediction technique, device, equipment and storage medium | |
| CN105321339A (en) | Road traffic intelligent management system | |
| Verma et al. | Avoiding stress driving: Online trip recommendation from driving behavior prediction | |
| Biggi et al. | Artificial intelligence in self-driving cars research and innovation: A scientometric and bibliometric analysis | |
| Feng et al. | Research on the identification and management of vehicle behaviour based on Internet of things technology | |
| Andrade et al. | Simple methodology for the development and analysis of local driving cycles applied in the study of cars and motorcycles in Recife, Brazil | |
| GANDHI et al. | Comparative evaluation of alternate bus rapid transit system (BRTS) planning, operation and design options | |
| CN117400948A (en) | Automobile energy consumption prediction method and device, electronic equipment and storage medium | |
| CN205080745U (en) | Road traffic intelligent management system |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20210326 Address after: No.123, Kaifa Avenue, Guiyang Economic and Technological Development Zone, 550000, Guizhou Province Applicant after: Man Bang Information Consulting Co.,Ltd. Address before: 210012 3-5 / F, building 4, 170-1 software Avenue, Yuhuatai District, Nanjing City, Jiangsu Province Applicant before: JIANGSU MANYUN SOFTWARE TECHNOLOGY Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP01 | Change in the name or title of a patent holder | ||
| CP01 | Change in the name or title of a patent holder |
Address after: No.123, Kaifa Avenue, Guiyang Economic and Technological Development Zone, 550000, Guizhou Province Patentee after: Manbang Information Technology Co.,Ltd. Address before: No.123, Kaifa Avenue, Guiyang Economic and Technological Development Zone, 550000, Guizhou Province Patentee before: Man Bang Information Consulting Co.,Ltd. |