CN116452008A - Second-hand vehicle risk prediction method and system based on polynomial modeling - Google Patents
Second-hand vehicle risk prediction method and system based on polynomial modeling Download PDFInfo
- Publication number
- CN116452008A CN116452008A CN202310712379.8A CN202310712379A CN116452008A CN 116452008 A CN116452008 A CN 116452008A CN 202310712379 A CN202310712379 A CN 202310712379A CN 116452008 A CN116452008 A CN 116452008A
- Authority
- CN
- China
- Prior art keywords
- data
- real
- polynomial
- hand
- discrete
- 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
- 238000000034 method Methods 0.000 title claims abstract description 83
- 238000012360 testing method Methods 0.000 claims abstract description 94
- 238000013507 mapping Methods 0.000 claims abstract description 25
- 238000004140 cleaning Methods 0.000 claims description 14
- 238000013479 data entry Methods 0.000 claims description 13
- 238000012423 maintenance Methods 0.000 claims description 12
- 238000013506 data mapping Methods 0.000 claims description 10
- 238000012217 deletion Methods 0.000 claims description 8
- 230000037430 deletion Effects 0.000 claims description 8
- 230000001133 acceleration Effects 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 5
- 238000001914 filtration Methods 0.000 claims description 4
- 230000003993 interaction Effects 0.000 claims description 4
- 238000012545 processing Methods 0.000 description 7
- 238000001514 detection method Methods 0.000 description 5
- 238000012502 risk assessment Methods 0.000 description 5
- 239000013598 vector Substances 0.000 description 5
- 238000010606 normalization Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000001303 quality assessment method Methods 0.000 description 2
- 230000001502 supplementing effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
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
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- 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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- General Physics & Mathematics (AREA)
- Game Theory and Decision Science (AREA)
- Marketing (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Educational Administration (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Data Mining & Analysis (AREA)
- Operations Research (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a second-hand vehicle risk prediction method and a second-hand vehicle risk prediction system based on polynomial modeling, and relates to the field of second-hand vehicle risk prediction, wherein the method comprises the following steps: mapping the acquired real-time data and historical data of the test second hand cart into discrete variables respectively; based on a polynomial regression method, a polynomial objective function between a risk prediction value of the test second-hand vehicle and a plurality of discrete variables respectively mapped by real-time data and historical data is established; fitting the polynomial objective function to obtain model parameters of the polynomial objective function; the real-time data and the historical data of the target second hand cart are input into the polynomial objective function, the risk prediction value of the target second hand cart is output, the problem that the second hand cart transaction is affected due to low accuracy and reliability of the second hand cart risk prediction result caused by the prior art is effectively solved, the accuracy and reliability of the second hand cart risk prediction result are effectively improved, and the second hand cart transaction can be conveniently and quickly carried out.
Description
Technical Field
The invention relates to the field of second-hand vehicle risk prediction, in particular to a second-hand vehicle risk prediction method and system based on polynomial modeling.
Background
The second-hand vehicle risk assessment prediction plays an important role in the identification of the value of the second-hand vehicle and the normal and rapid transaction of the second-hand vehicle.
However, the potential quality problem of the existing second-hand vehicle is that one type of the existing second-hand vehicle usually adopts manual analysis data, and final risk assessment is carried out according to personal experience conditions, but the mode has large manual subjective factors, so that the actual quality and risk assessment of the second-hand vehicle are affected; another way is a second-hand vehicle detection method as disclosed in application number CN202110949699.6, which discloses obtaining historical detection data and historical maintenance data of a vehicle, and obtaining a fault defect of the vehicle according to the historical detection data and the historical maintenance data of the vehicle; the comprehensive analysis of the fault defects of the vehicles and the like can be carried out to obtain comprehensive detection results of the vehicles, and although the detection results of the second hand vehicles can also be obtained, the quality condition of the second hand vehicles is subjected to risk assessment prediction by considering a certain historical data, so that the risk prediction results of the second hand vehicles are not accurate enough, and the actual quality and risk assessment of the second hand vehicles are affected.
The scheme is provided for solving the problems of low accuracy and reliability of the secondary handcart risk prediction result in the prior art.
Disclosure of Invention
In order to solve the problems in the prior art, the invention innovatively provides a second-hand vehicle risk prediction and system based on polynomial modeling, which effectively solves the problems that the accuracy and reliability of the second-hand vehicle risk prediction result are not high and the second-hand vehicle transaction is influenced due to the prior art, effectively improves the accuracy and reliability of the second-hand vehicle risk prediction result and is convenient for normal and rapid second-hand vehicle transaction.
The first aspect of the invention provides a second-hand vehicle risk prediction method based on multivariate data polynomial modeling, which comprises the following steps:
acquiring real-time data and historical data of the test cart, and mapping the acquired real-time data and historical data of the test cart into discrete variables respectively, wherein each type of real-time data or each type of historical data corresponds to one discrete variable;
based on a polynomial regression method, a polynomial objective function between a risk prediction value of the test second-hand vehicle and a plurality of discrete variables respectively mapped by real-time data and historical data is established;
fitting the established polynomial objective function aiming at the real-time sample data and the history sample data of the test second hand cart to obtain model parameters of the polynomial objective function;
And acquiring real-time data and historical data of the target second hand vehicle, inputting the real-time data and the historical data of the acquired target second hand vehicle into a polynomial objective function, and outputting a risk prediction value of the target second hand vehicle.
Optionally, the real-time data of the test second hand cart comprises vehicle acceleration time, braking distance, steering wheel angle, steering torque and steering angular speed.
Optionally, the historical data of the test second hand cart comprises maintenance times, vehicle accident times, vehicle age, driving mileage and vehicle use environment.
Optionally, before mapping the acquired real-time data and the historical data of the test cart into discrete variables, the method further includes:
data cleaning is carried out on the real-time data and the historical data, wherein the data cleaning is specifically carried out on the real-time data and the historical data: and deleting the history data with completely overlapped data entries, and filling the deleted history data or real-time data by adopting an average value under the same data.
Optionally, before establishing the polynomial objective function between the risk prediction value of the test second-hand vehicle and the discrete variables mapped by the real-time data and the historical data respectively based on the polynomial regression method, the method further comprises:
Calculating the variance value of each discrete variable of different types of historical data or real-time data mapping, and deleting the discrete variable with variance smaller than a preset variance threshold; the variance value of each discrete variable of the different types of historical data or real-time data mapping is calculated specifically as follows:，
wherein ,for the variance value of the current discrete variable mapped by a certain type of historical data or real-time data, n is the number of historical data or real-time data entries corresponding to the current discrete variable, < >>Representing the value of the ith historical data or real-time data corresponding to the current discrete variable,/or->Representing the average value of all the historical data or real-time data corresponding to the current discrete variable.
Further, before establishing the polynomial objective function between the risk prediction value of the test cart and the plurality of discrete variables respectively mapped by the real-time data and the historical data based on the polynomial regression method, the method further comprises:
calculating a correlation coefficient between any two discrete variables, and deleting the discrete variables with the correlation coefficients larger than a preset correlation coefficient threshold; the method specifically calculates the correlation coefficient between any two discrete variables as follows:，
wherein ,represents the ith historical data corresponding to the current discrete variable Or the value of the real-time data,/>Representing the value of the ith historical data or real-time data corresponding to the current another discrete variable, n is the entry number of the historical data or real-time data corresponding to the current two discrete variables,/for>Is the correlation coefficient between two discrete variables.
Further, based on a polynomial regression method, establishing a polynomial objective function between a risk prediction value of the test second-hand vehicle and a plurality of discrete variables respectively mapped by real-time data and historical data specifically comprises:
after variable deletion and filtration, the number of the remaining discrete variables is k, and the polynomial objective function is:
，
wherein k is the number of discrete variables remaining after the deletion of the variables,is a constant term->For the primary term coefficient->Is a polynomial coefficient of d and f, representing the interaction between the d-th discrete variable and the f-th discrete variable, h is the polynomial degree,is the value of the d-th discrete variable, +.>To the power f of the value of the d-th discrete variable.
Further, the polynomial coefficientsThe range of the values of (2) is determined by polynomial degree h.
Optionally, fitting the polynomial objective function established according to the real-time sample data and the historical sample data of the test cart to obtain model parameters of the polynomial objective function specifically includes:
Fitting the real-time sample data and the history sample data of the test second hand cart to the established polynomial objective function through a plurality of discrete variables respectively converted, and obtaining model parameters of the polynomial objective function after sparsification through regularization treatment after fitting; the regularization process specifically includes:
，
wherein m is the total number of real-time sample data and historical sample data of the test cart corresponding to the remaining discrete variable,is the predicted value of the first real-time sample data or the historical sample data of the test second cart corresponding to the input residual discrete variable, +.>Is the actual output value of the first real-time sample data or the historical sample data of the test second cart corresponding to the input residual discrete variable, +.>Is a regularization parameter, min imize is a minimization function.
The second aspect of the invention provides a second-hand vehicle risk prediction system based on polynomial modeling, comprising:
the acquisition module is used for acquiring real-time data and historical data of the test second hand cart, and mapping the acquired real-time data and historical data of the test second hand cart into discrete variables respectively, wherein each type of real-time data or each type of historical data corresponds to one discrete variable;
The building module is used for building a polynomial objective function between a risk prediction value of the test second-hand vehicle and a plurality of discrete variables respectively mapped by the real-time data and the historical data based on a polynomial regression method;
the fitting module is used for fitting the established polynomial objective function aiming at the real-time sample data and the history sample data of the test second hand cart to obtain model parameters of the polynomial objective function;
and the output module is used for acquiring real-time data and historical data of the target second hand vehicle, inputting the acquired real-time data and the historical data of the target second hand vehicle into the polynomial objective function and outputting a risk prediction value of the target second hand vehicle.
The technical scheme adopted by the invention comprises the following technical effects:
1. according to the invention, the acquired real-time data and historical data of the test second hand cart are respectively mapped into discrete variables; based on a polynomial regression method, a polynomial objective function between a risk prediction value of the test second-hand vehicle and a plurality of discrete variables respectively mapped by real-time data and historical data is established; fitting the established polynomial objective function aiming at the real-time sample data and the history sample data of the test second hand cart to obtain model parameters of the polynomial objective function; the method comprises the steps of acquiring real-time data and historical data of a target second hand cart, inputting the real-time data and the historical data of the acquired target second hand cart into a polynomial objective function, outputting a risk prediction value of the target second hand cart, effectively solving the problems that the accuracy and the reliability of the risk prediction result of the second hand cart are low and the transaction of the second hand cart is affected due to the fact that the prior art, effectively improving the accuracy and the reliability of the risk prediction result of the second hand cart, and facilitating the normal and rapid progress of the transaction of the second hand cart.
2. According to the technical scheme, the sample data not only comprises the historical data of the second hand cart, but also comprises the real-time data of the second hand cart, the data of two dimensions are established, the risk prediction of the final vehicle is realized according to the polynomial model, and the differences of artificial subjective factors and experience assessment are reduced; and because the final vehicle risk prediction is influenced by a plurality of variables and the variables are in discrete relation, the technical scheme of the invention adopts a polynomial regression method, and the number of times of independent variables is increased to map data to a high-dimensional space, thereby realizing nonlinear prediction and improving the model prediction precision of a polynomial objective function.
3. Before the obtained real-time data and history data of the test cart are respectively mapped into discrete variables, the technical scheme of the invention further comprises the following steps: data cleaning is carried out on the real-time data and the historical data, wherein the data cleaning is specifically carried out on the real-time data and the historical data: and deleting the history data with completely overlapped data entries, and filling the deleted history data or the real-time data by adopting an average value under the same data to ensure the validity of the real-time sample data and the history sample data.
4. In the technical scheme of the invention, before the polynomial regression method is based on the polynomial objective function between the risk prediction value of the test second-hand vehicle and a plurality of discrete variables respectively mapped by the real-time data and the historical data is established, the method further comprises the following steps: calculating the variance value of each discrete variable of different types of historical data or real-time data mapping, and deleting the discrete variable with variance smaller than a preset variance threshold; the accuracy and reliability in the actual prediction result are avoided from being affected by fluctuations in the variable values.
5. In the technical scheme of the invention, before the polynomial regression method is based on the polynomial objective function between the risk prediction value of the test second-hand vehicle and a plurality of discrete variables respectively mapped by the real-time data and the historical data is established, the method further comprises the following steps: calculating a correlation coefficient between any two discrete variables, and deleting the discrete variables with the correlation coefficients larger than a preset correlation coefficient threshold; further, the influence on the accuracy and reliability of the actual prediction result is avoided because of the feature redundancy when the trend of the change values of the two discrete variables is consistent, namely, the feature redundancy exists.
6. According to the technical scheme, the method for obtaining the model parameters of the polynomial objective function specifically comprises the following steps of: fitting the discrete variables respectively converted from the real-time sample data and the history sample data of the test second hand cart to the established polynomial objective function, and obtaining model parameters of the polynomial objective function after sparsification by regularization treatment after fitting; the L1 norms of the polynomial parameter vectors are limited by adopting L1 regularization, so that the purpose of sparsification is achieved, and the model complexity of the polynomial objective function is reduced.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention as claimed.
Drawings
For a clearer description of embodiments of the invention or of the solutions of the prior art, reference will be made to the accompanying drawings, which are used in the description of the embodiments or of the prior art, and it will be obvious to those skilled in the art that other drawings can be obtained from these without inventive labour.
FIG. 1 is a schematic flow chart of a method according to an embodiment of the present invention;
FIG. 2 is a schematic flow chart of a method according to an embodiment of the present invention;
FIG. 3 is a schematic flow chart (III) of a method according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of a second system according to an embodiment of the present invention.
Detailed Description
In order to clearly illustrate the technical features of the present solution, the present invention will be described in detail below with reference to the following detailed description and the accompanying drawings. The following disclosure provides many different embodiments, or examples, for implementing different structures of the invention. In order to simplify the present disclosure, components and arrangements of specific examples are described below. Furthermore, the present invention may repeat reference numerals and/or letters in the various examples. This repetition is for the purpose of simplicity and clarity and does not in itself dictate a relationship between the various embodiments and/or configurations discussed. It should be noted that the components illustrated in the figures are not necessarily drawn to scale. Descriptions of well-known components and processing techniques and processes are omitted so as to not unnecessarily obscure the present invention.
Example 1
As shown in fig. 1, the invention provides a second-hand vehicle risk prediction method based on polynomial modeling, which comprises the following steps:
s1, acquiring real-time data and historical data of a test second-hand cart, and mapping the acquired real-time data and historical data of the test second-hand cart into discrete variables respectively, wherein each type of real-time data or each type of historical data corresponds to one discrete variable;
s3, based on a polynomial regression method, establishing a polynomial objective function between a risk prediction value of the test second-hand vehicle and a plurality of discrete variables respectively mapped by real-time data and historical data;
s5, fitting the established polynomial objective function aiming at the real-time sample data and the historical sample data of the test second hand cart to obtain model parameters of the polynomial objective function;
and S7, acquiring real-time data and historical data of the target second hand truck, inputting the acquired real-time data and the historical data of the target second hand truck into a polynomial objective function, and outputting a risk prediction value of the target second hand truck.
In step S1, the real-time data of the test second hand cart includes real-time data of different types (items) of vehicle acceleration time, braking distance, steering wheel angle, steering torque, and steering angular velocity; the historical data of the second hand vehicle comprises historical data of different types (items) of maintenance times, vehicle accident times, vehicle age, driving mileage and vehicle use environment; and then, mapping the acquired real-time data and historical data of the test second hand cart into discrete variables respectively, wherein each type of real-time data or each type of historical data corresponds to one discrete variable.
Specifically, mapping the entry information in the historical data to actual numerical representation, such as maintenance, vehicle accident, vehicle age and quantifiable data of driving mileage data, directly adopting numerical representation, such as enumeration of discretized data of the vehicle use environment into urban commute, field driving and long-distance driving, and respectively dividing the data into 3 (a plurality of (also can be used), and the invention is not limited by the invention) levels (1, 2 and 3) for representation;
for historical data, considering that discrete variables exist, each item of historical data is coded and mapped by using Label Encoding, and the mapping value results are shown in the following table I:
table one: historical data coding mapping table
In the real-time data, the vehicle acceleration time, the braking distance, the steering wheel rotation angle, the steering torque and the steering angular speed are directly quantized by adopting values, namely, the real-time data are mapped: the vehicle acceleration time, braking distance, steering wheel angle, steering torque and steering angular velocity are tested as shown in the following table two:
and (II) table: real-time data encoding mapping table
Further, in step S1, after acquiring the real-time data and the history data of the test cart, before mapping the acquired real-time data and history data of the test cart into discrete variables, the method further includes:
Data cleaning is carried out on the real-time data and the historical data, wherein the data cleaning is specifically carried out on the real-time data and the historical data: and deleting the history data with completely overlapped data entries, and filling the deleted history data or real-time data by adopting an average value under the same data (the same type or the same item).
Specifically, for the provided historical data or real-time data, deleting the data with completely overlapped data entries in the historical data, and filling the deleted historical data or real-time data by adopting an average value under the same data. And processing the collected data, including removing repeated data, filling missing data and normalizing.
The normalization processing is carried out, wherein the maximum and minimum values in the current attribute are obtained by adopting a mode of independently normalizing the data of each attribute, and linear transformation is carried out to the interval of [0,1 ]; the missing values are calculated first, the missing values are filled, and then the attributes in the data are listed as discrete variables in prediction.
Further, as shown in fig. 2, the second-hand vehicle risk prediction method based on polynomial modeling provided by the technical scheme of the present invention further includes, before step S3:
S21, calculating the variance value of each discrete variable mapped by different types of historical data, and deleting the discrete variables with variances smaller than a preset variance threshold; the variance value of each discrete variable of the different types of historical data or real-time data mapping is calculated specifically as follows:，
wherein ,for the variance value of the current discrete variable mapped by a certain type of historical data or real-time data, n is the number of historical data or real-time data entries corresponding to the current discrete variable, < >>Representing the value of the ith historical data or real-time data corresponding to the current discrete variable,/or->Represents the average of all the historical data or real-time data corresponding to the current discrete variable (all the historical data or real-time data under the type or item corresponding to the current discrete variable).
Considering the effectiveness of each discrete variable on final risk prediction, the fluctuation of the discrete variable value also represents the contribution in the actual prediction result, so that the variance value of each discrete variable after different types of historical data or real-time data mapping is calculated in different types of historical data, and the discrete variable with smaller variance (smaller than a preset variance threshold value, which can be flexibly adjusted according to the actual situation) is deleted.
Further, as shown in fig. 3, the second-hand vehicle risk prediction method based on polynomial modeling provided by the technical scheme of the present invention further includes, before step S3:
s22, calculating a correlation coefficient between any two discrete variables, and deleting the discrete variables with the correlation coefficients larger than a preset correlation coefficient threshold; the method specifically calculates the correlation coefficient between any two discrete variables as follows:，
wherein ,a value representing the ith history data or real-time data corresponding to the current discrete variable,/-, for example>Representing the value of the ith historical data or real-time data corresponding to the current another discrete variable, n is the entry number of the historical data or real-time data corresponding to the current two discrete variables,/for>Is the correlation coefficient between two discrete variables.
Considering the correlation between the discrete variables, when the trend of the variation values of the two discrete variables is consistent, feature redundancy exists, the correlation between the two discrete variables is calculated first, if the correlation between the two discrete variables is larger than 0.7, the correlation is high, and the discrete variables are deleted. Specifically, in the technical scheme of the invention, the pearson correlation coefficient method is adopted to analyze the relation between every two discrete variables, and other types of correlation coefficient methods can also be adopted to calculate the correlation coefficient, so that the invention is not limited.
In step S3, based on a polynomial regression method, a polynomial objective function between a risk prediction value of the test cart and a plurality of discrete variables mapped by real-time data and historical data is established specifically as follows:
after variable deletion and filtration, the number of the remaining discrete variables is k, and the polynomial objective function is:
，
wherein k is the number of discrete variables remaining after the deletion of the variables,is a constant term->For the primary term coefficient->Is a polynomial coefficient of d and f, representing the interaction between the d-th discrete variable and the f-th discrete variable, h is the polynomial degree,is the value of the d-th discrete variable, +.>To the power f of the value of the d-th discrete variable, in particular, when f=1,/and>namely, is，/>Namely +.>。
The final vehicle risk prediction is influenced by a plurality of discrete variables, and the discrete variables are in discrete relation, so that a polynomial regression method is adopted, and data are mapped to a high-dimensional space by increasing the times of independent variables, so that nonlinear prediction is realized, and the model prediction precision is improved.
Polynomial coefficientsThe range of values of (2) is determined by polynomial degree h, i.e. +.>The range of the values of (a) is specificallyThe method comprises the steps of carrying out a first treatment on the surface of the For example, when polynomial degree h=2, ++ >The value range of (2) is +.>When polynomial degree h=3, ++>The value range of (2) is +.>And so on.
In step S5, the established polynomial objective function is fitted to the real-time sample data and the history sample data of the test cart, and the model parameters of the polynomial objective function are specifically:
fitting the real-time sample data and the history sample data of the test second hand cart to the established polynomial objective function through a plurality of discrete variables respectively converted, and obtaining model parameters of the polynomial objective function after sparsification through regularization treatment after fitting; the regularization process specifically includes:
，
wherein m is the total number of real-time sample data and historical sample data of the test cart corresponding to the remaining discrete variable,is the predicted value of the first real-time sample data or the historical sample data of the test second cart corresponding to the input residual discrete variable,/>is the actual output value of the first real-time sample data or the historical sample data of the test second cart corresponding to the input residual discrete variable, +.>Is a regularization parameter, +.>For the primary term coefficient->Is a polynomial coefficient, min im is a minimization function.
Specifically, due to the increase of the number of independent variables and the number of polynomial times, the model complexity is high, the problem of model overfitting easily occurs, regularization processing is performed on the target upper expression, the parameter size is limited, the model complexity is reduced, and the calculation formula is as follows:
the regularization parameters are used for controlling the regularization intensity, the larger the value is, the larger the regularization influence is, the L1 norm of the L1 regularization limiting polynomial parameter vector is adopted for achieving the purpose of sparsification, and the complexity of the model is reduced.
In step S7, the collected historical data and real-time data (discrete variable values after mapping respectively) of the vehicle to be predicted are input into a model of the fitted polynomial objective function to obtain a predicted vehicle risk coefficient, and corresponding prompting operation is performed according to the obtained predicted second-hand vehicle risk coefficient, for example, when the obtained predicted second-hand vehicle risk coefficient is smaller than a preset risk coefficient threshold, prompting that the second-hand vehicle risk to be predicted is small; and when the obtained predicted risk coefficient of the second-hand vehicle is not smaller than a preset risk coefficient threshold value, prompting that the risk of the second-hand vehicle to be predicted is large, and carrying out risk early warning.
For a clearer illustration of the solution, take an X brand vehicle Q as an example:
The scrapping period of the vehicle Q is 15 years, the vehicle age is 5 years, the total maintenance is 5 times, the driving mileage is 8 ten thousand kilometers, the longest (max) driving in the database is 60 ten thousand kilometers, the minimum (min) driving is 0 ten thousand kilometers, the service environment is city commute, and the corresponding historical data are shown in the following table III:
table three: historical data encoding mapping table of vehicle Q
The vehicle Q hundred kilometers accelerates for 8s, the actual test value is 9s, the normal vehicle braking distance is 48m, the actual test is 55m, the steering angular velocity is 21 radians/s, the normal steering angular velocity is 20 radians/s, the steering torque is 3500, the steering wheel angle is 35 degrees in the normal range of 2000-5000 degrees, and the angle degree of the normal vehicle is 30-40 degrees. According to the test data, the corresponding history data is shown in the following table four:
table four: real-time data encoding mapping table of vehicle Q
Supplementing by the missing value: if the maintenance frequency information is absent in the historical data, the missing data is filled by taking an average value according to the data items of different vehicles Q (same model) of the brand X in the database.
Normalization: for the maximum and minimum values of each variable of the vehicle Q under the brand X in the database, the variable value of the vehicle Q under the brand X is mapped and transformed into [0,1] through (X-min)/(max-min).
Polynomial prediction: and taking the target second-hand vehicle, namely the discrete variable values after the historical data and the real-time data of the vehicle Q are mapped respectively, into a fitted polynomial objective function, outputting a predicted second-hand vehicle risk coefficient of the current vehicle Q (target second-hand vehicle) as z, and determining the risk coefficient grade of the vehicle Q according to the range of the z value, namely excellent, good, general and poor. Wherein, four different risk factor class ranges are divided as shown in the following table five:
table five: risk coefficient grade range dividing table
Specifically, the target second-hand vehicle, i.e. the discrete variable values mapped by the historical data and the real-time data of the vehicle Q, is brought into the fitted polynomial objective function, and the predicted second-hand vehicle risk coefficient of the current vehicle Q (target second-hand vehicle) is output to be 0.82, i.e. the corresponding risk coefficient grade of the current vehicle Q (target second-hand vehicle) is good. When the obtained predicted second-hand vehicle risk coefficient is smaller than a preset risk coefficient threshold value (which can be 0.5 or 0.8), prompting that the second-hand vehicle risk to be predicted is small; when the obtained predicted risk coefficient of the second-hand vehicle is not smaller than a preset risk coefficient threshold value (which can be 0.5 or 0.8), prompting that the risk of the second-hand vehicle to be predicted is large, and carrying out risk early warning.
According to the technical scheme, from historical data and real-time data of the second-hand vehicle, different numerical quantization operations are carried out on the two types of data, the quantized data are used as variables of a polynomials, modeling is carried out from multiple dimensions, nonlinear mapping capacity of a model is enhanced, model complexity of the polynomials is optimized through regularization, prediction accuracy of the model is improved, and problems existing in artificial subjective factor assessment of vehicle risks are effectively improved.
According to the invention, the acquired real-time data and historical data of the test second hand cart are respectively mapped into discrete variables; based on a polynomial regression method, a polynomial objective function between a risk prediction value of the test second-hand vehicle and a plurality of discrete variables respectively mapped by real-time data and historical data is established; fitting the established polynomial objective function aiming at the real-time sample data and the history sample data of the test second hand cart to obtain model parameters of the polynomial objective function; the method comprises the steps of acquiring real-time data and historical data of a target second hand cart, inputting the real-time data and the historical data of the acquired target second hand cart into a polynomial objective function, outputting a risk prediction value of the target second hand cart, effectively solving the problems that the accuracy and the reliability of the risk prediction result of the second hand cart are low and the transaction of the second hand cart is affected due to the fact that the prior art, effectively improving the accuracy and the reliability of the risk prediction result of the second hand cart, and facilitating the normal and rapid progress of the transaction of the second hand cart.
According to the technical scheme, the sample data not only comprises the historical data of the second hand cart, but also comprises the real-time data of the second hand cart, the data of two dimensions are established, the risk prediction of the final vehicle is realized according to the polynomial model, and the differences of artificial subjective factors and experience assessment are reduced; and because the final vehicle risk prediction is influenced by a plurality of variables and the variables are in discrete relation, the technical scheme of the invention adopts a polynomial regression method, and the number of times of independent variables is increased to map data to a high-dimensional space, thereby realizing nonlinear prediction and improving the model prediction precision of a polynomial objective function.
Before the obtained real-time data and history data of the test cart are respectively mapped into discrete variables, the technical scheme of the invention further comprises the following steps: data cleaning is carried out on the real-time data and the historical data, wherein the data cleaning is specifically carried out on the real-time data and the historical data: and deleting the history data with completely overlapped data entries, and filling the deleted history data or the real-time data by adopting an average value under the same data to ensure the validity of the real-time sample data and the history sample data.
In the technical scheme of the invention, before the polynomial regression method is based on the polynomial objective function between the risk prediction value of the test second-hand vehicle and a plurality of discrete variables respectively mapped by the real-time data and the historical data is established, the method further comprises the following steps: calculating the variance value of each discrete variable of different types of historical data or real-time data mapping, and deleting the discrete variable with variance smaller than a preset variance threshold; the accuracy and reliability in the actual prediction result are avoided from being affected by fluctuations in the variable values.
In the technical scheme of the invention, before the polynomial regression method is based on the polynomial objective function between the risk prediction value of the test second-hand vehicle and a plurality of discrete variables respectively mapped by the real-time data and the historical data is established, the method further comprises the following steps: calculating a correlation coefficient between any two discrete variables, and deleting the discrete variables with the correlation coefficients larger than a preset correlation coefficient threshold; further, the influence on the accuracy and reliability of the actual prediction result is avoided when the trend of the change values of the two variables is consistent, namely, when feature redundancy exists.
According to the technical scheme, the method for obtaining the model parameters of the polynomial objective function specifically comprises the following steps of: fitting the real-time sample data and the history sample data of the test second hand cart to the established polynomial objective function through a plurality of discrete variables respectively converted, and obtaining model parameters of the polynomial objective function after sparsification through regularization treatment after fitting; the L1 norms of the polynomial parameter vectors are limited by adopting L1 regularization, so that the purpose of sparsification is achieved, and the model complexity of the polynomial objective function is reduced.
Example two
As shown in fig. 4, the technical solution of the present invention further provides a second-hand vehicle risk prediction system based on polynomial modeling, including:
the acquisition module 101 acquires real-time data and historical data of the test second hand cart, and maps the acquired real-time data and historical data of the test second hand cart into discrete variables respectively, wherein each type of real-time data or each type of historical data corresponds to one discrete variable;
the establishing module 102 is used for establishing a polynomial objective function between a risk prediction value of the test second-hand vehicle and a plurality of discrete variables respectively mapped by the real-time data and the historical data based on a polynomial regression method;
the fitting module 103 is used for fitting the established polynomial objective function aiming at the real-time sample data and the historical sample data of the test second hand cart to obtain model parameters of the polynomial objective function;
and the output module 104 is used for acquiring real-time data and historical data of the target second hand vehicle, inputting the acquired real-time data and the historical data of the target second hand vehicle into the polynomial objective function and outputting a risk prediction value of the target second hand vehicle.
Wherein, in the acquisition module 101, the real-time data of the test cart includes, but is not limited to, vehicle acceleration time, braking distance (representing one type or one item of historical data), steering wheel angle, steering torque, steering angular velocity; the historical data of the test second hand vehicle includes, but is not limited to, maintenance times, vehicle accident times, vehicle age, driving mileage and vehicle use environment; and then, mapping the acquired real-time data and historical data of the test second hand cart into discrete variables respectively, wherein each type of real-time data or each type of historical data corresponds to one discrete variable.
Specifically, mapping the entry information in the historical data to actual numerical representation, such as maintenance, vehicle accident, vehicle age and quantifiable data of driving mileage data, directly adopts numerical representation, such as enumeration of discrete data of the vehicle use environment into city commute, field driving and long distance driving, and respectively dividing the data into 3 grades (1, 2 and 3) for representation.
Further, after acquiring the real-time data and the history data of the test cart, before mapping the acquired real-time data and history data of the test cart into discrete variables, the method further includes:
data cleaning is carried out on the real-time data and the historical data, wherein the data cleaning is specifically carried out on the real-time data and the historical data: and deleting the history data with completely overlapped data entries, and filling the deleted history data or real-time data by adopting an average value under the same data.
Specifically, for the provided historical data or real-time data, deleting the data with completely overlapped data entries in the historical data, and filling the deleted historical data or real-time data by adopting an average value under the same data. And processing the collected data, including removing repeated data, filling missing data and normalizing.
The normalization processing is carried out, wherein the maximum and minimum values in the current attribute are obtained by adopting a mode of independently normalizing the data of each attribute, and linear transformation is carried out to the interval of [0,1 ]; the missing values are calculated first, the missing values are filled, and then the attributes in the data are listed as discrete variables in prediction.
Further, before the building module 102 executes, it is further required to filter and delete the discrete variables of the mapping transformation, that is, firstly calculate the variance value of each discrete variable of the different types of historical data mapping, and delete the discrete variables with variances smaller than the preset variance threshold; the variance value of each discrete variable of the different types of historical data or real-time data mapping is calculated specifically as follows:，
wherein ,for the variance value of the current discrete variable mapped by a certain type of historical data or real-time data, n is the number of historical data or real-time data entries corresponding to the current discrete variable, < >>Representing the value of the ith historical data or real-time data corresponding to the current discrete variable,/or->Represents the average of all the historical data or real-time data corresponding to the current discrete variable (all the historical data or real-time data under the type or item corresponding to the current discrete variable).
Then calculating a correlation coefficient between any two discrete variables, and deleting the discrete variables with the correlation coefficients larger than a preset correlation coefficient threshold; the method specifically calculates the correlation coefficient between any two discrete variables as follows:，
wherein ,a value representing the ith history data or real-time data corresponding to the current discrete variable,/-, for example>Representing the current other departureThe value of the ith historical data or real-time data corresponding to the scattered variable, n is the entry number of the historical data or real-time data corresponding to the current two discrete variables, < + >>Is the correlation coefficient between two discrete variables.
In the establishing module 102, based on a polynomial regression method, a polynomial objective function between a risk prediction value of the test cart and a plurality of discrete variables respectively mapped by real-time data and historical data is established specifically as follows:
after variable deletion and filtration, the number of the remaining discrete variables is k, and the polynomial objective function is:
，
wherein k is the number of discrete variables remaining after the deletion of the variables,is a constant term->For the primary term coefficient->Is a polynomial coefficient of d and f, representing the interaction between the d-th discrete variable and the f-th discrete variable, h is the polynomial degree,is the value of the d-th discrete variable, +. >To the power f of the value of the d-th discrete variable, in particular, when f=1,/and>namely, is，/>Namely +.>。
The final vehicle risk prediction is influenced by a plurality of discrete variables, and the discrete variables are in discrete relation, so that a polynomial regression method is adopted, and data are mapped to a high-dimensional space by increasing the times of independent variables, so that nonlinear prediction is realized, and the model prediction precision is improved.
Polynomial coefficientsThe range of values of (2) is determined by polynomial degree h, i.e. +.>The range of the values of (a) is specificallyThe method comprises the steps of carrying out a first treatment on the surface of the For example, when polynomial degree h=2, ++>The value range of (2) is +.>When polynomial degree h=3, ++>The value range of (2) is +.>And so on. />
In the fitting module 103, the established polynomial objective function is fitted to the real-time sample data and the history sample data of the test cart, so that model parameters of the polynomial objective function are specifically:
fitting the real-time sample data and the history sample data of the test second hand cart to the established polynomial objective function through a plurality of discrete variables respectively converted, and obtaining model parameters of the polynomial objective function after sparsification through regularization treatment after fitting; the regularization process specifically includes:
，
Wherein m is the total number of real-time sample data and historical sample data of the test cart corresponding to the remaining discrete variable,is the predicted value of the first real-time sample data or the historical sample data of the test second cart corresponding to the input residual discrete variable, +.>Is the actual output value of the first real-time sample data or the historical sample data of the test second cart corresponding to the input residual discrete variable, +.>Is a regularization parameter, +.>For the primary term coefficient->Is a polynomial coefficient, min im is a minimization function.
Specifically, due to the increase of the number of independent variables and the number of polynomial times, the model complexity is high, the problem of model overfitting easily occurs, regularization processing is performed on the target upper expression, the parameter size is limited, the model complexity is reduced, and the calculation formula is as follows:
the regularization parameters are used for controlling the regularization intensity, the larger the value is, the larger the regularization influence is, the L1 norm of the L1 regularization limiting polynomial parameter vector is adopted for achieving the purpose of sparsification, and the complexity of the model is reduced.
In the output module 104, input collected historical data and real-time data (discrete variable values after mapping respectively) of the vehicle to be predicted into a model of a polynomial objective function which is already fitted to obtain a predicted vehicle risk coefficient, and perform corresponding prompting operation according to the obtained predicted second-hand vehicle risk coefficient, for example, when the obtained predicted second-hand vehicle risk coefficient is smaller than a preset risk coefficient threshold, prompting that the second-hand vehicle risk to be predicted is small; and when the obtained predicted risk coefficient of the second-hand vehicle is not smaller than a preset risk coefficient threshold value, prompting that the risk of the second-hand vehicle to be predicted is large, and carrying out risk early warning.
For a clearer illustration of the solution, take an X brand vehicle Q as an example:
the scrapping period of the vehicle Q is 15 years, the vehicle age is 5 years, the total maintenance is 5 times, the driving mileage is 8 ten thousand kilometers, the longest (max) driving in the database is 60 ten thousand kilometers, the minimum (min) driving is 0 ten thousand kilometers, the service environment is city commute, and the corresponding historical data are shown in the following table III:
table three: historical data encoding mapping table of vehicle Q
The vehicle Q hundred kilometers accelerates for 8s, the actual test value is 9s, the normal vehicle braking distance is 48m, the actual test is 55m, the steering angular velocity is 21 radians/s, the normal steering angular velocity is 20 radians/s, the steering torque is 3500, the steering wheel angle is 35 degrees in the normal range of 2000-5000 degrees, and the angle degree of the normal vehicle is 30-40 degrees. According to the test data, the corresponding history data is shown in the following table four:
table four: real-time data encoding mapping table of vehicle Q
Supplementing by the missing value: if the maintenance frequency information is absent in the historical data, the missing data is filled by taking an average value according to the data items of different vehicles Q (same model) of the brand X in the database.
Normalization: for the maximum and minimum values of each variable of the vehicle Q under the brand X in the database, the variable value of the vehicle Q under the brand X is mapped and transformed into [0,1] through (X-min)/(max-min).
Polynomial prediction: and taking the target second-hand vehicle, namely the discrete variable values after the historical data and the real-time data of the vehicle Q are mapped respectively, into a fitted polynomial objective function, outputting a predicted second-hand vehicle risk coefficient of the current vehicle Q (target second-hand vehicle) as z, and determining the risk coefficient grade of the vehicle Q according to the range of the z value, namely excellent, good, general and poor. Wherein, four different risk factor class ranges are divided as shown in the following table five:
table five: risk coefficient grade range dividing table
Specifically, the target second-hand vehicle, i.e. the discrete variable values mapped by the historical data and the real-time data of the vehicle Q, is brought into the fitted polynomial objective function, and the predicted second-hand vehicle risk coefficient of the current vehicle Q (target second-hand vehicle) is output to be 0.82, i.e. the corresponding risk coefficient grade of the current vehicle Q (target second-hand vehicle) is good. When the obtained predicted second-hand vehicle risk coefficient is smaller than a preset risk coefficient threshold value (which can be 0.5 or 0.8), prompting that the second-hand vehicle risk to be predicted is small; when the obtained predicted risk coefficient of the second-hand vehicle is not smaller than a preset risk coefficient threshold value (which can be 0.5 or 0.8), prompting that the risk of the second-hand vehicle to be predicted is large, and carrying out risk early warning.
According to the invention, the acquired real-time data and historical data of the test second hand cart are respectively mapped into discrete variables; based on a polynomial regression method, a polynomial objective function between a risk prediction value of the test second-hand vehicle and a plurality of discrete variables respectively mapped by real-time data and historical data is established; fitting the established polynomial objective function aiming at the real-time sample data and the history sample data of the test second hand cart to obtain model parameters of the polynomial objective function; the method comprises the steps of acquiring real-time data and historical data of a target second hand cart, inputting the real-time data and the historical data of the acquired target second hand cart into a polynomial objective function, outputting a risk prediction value of the target second hand cart, effectively solving the problems that the accuracy and the reliability of the risk prediction result of the second hand cart are low and the transaction of the second hand cart is affected due to the fact that the prior art, effectively improving the accuracy and the reliability of the risk prediction result of the second hand cart, and facilitating the normal and rapid progress of the transaction of the second hand cart.
According to the technical scheme, the sample data not only comprises the historical data of the second hand cart, but also comprises the real-time data of the second hand cart, the data of two dimensions are established, the risk prediction of the final vehicle is realized according to the polynomial model, and the differences of artificial subjective factors and experience assessment are reduced; and because the final vehicle risk prediction is influenced by a plurality of variables and the variables are in discrete relation, the technical scheme of the invention adopts a polynomial regression method, and the number of times of independent variables is increased to map data to a high-dimensional space, thereby realizing nonlinear prediction and improving the model prediction precision of a polynomial objective function.
Before the obtained real-time data and history data of the test cart are respectively mapped into discrete variables, the technical scheme of the invention further comprises the following steps: data cleaning is carried out on the real-time data and the historical data, wherein the data cleaning is specifically carried out on the real-time data and the historical data: and deleting the history data with completely overlapped data entries, and filling the deleted history data or the real-time data by adopting an average value under the same data to ensure the validity of the real-time sample data and the history sample data.
In the technical scheme of the invention, before the polynomial regression method is based on the polynomial objective function between the risk prediction value of the test second-hand vehicle and a plurality of discrete variables respectively mapped by the real-time data and the historical data is established, the method further comprises the following steps: calculating the variance value of each discrete variable of different types of historical data or real-time data mapping, and deleting the discrete variable with variance smaller than a preset variance threshold; the accuracy and reliability in the actual prediction result are avoided from being affected by fluctuations in the variable values.
In the technical scheme of the invention, before the polynomial regression method is based on the polynomial objective function between the risk prediction value of the test second-hand vehicle and a plurality of discrete variables respectively mapped by the real-time data and the historical data is established, the method further comprises the following steps: calculating a correlation coefficient between any two discrete variables, and deleting the discrete variables with the correlation coefficients larger than a preset correlation coefficient threshold; further, the influence on the accuracy and reliability of the actual prediction result is avoided when the trend of the change values of the two variables is consistent, namely, when feature redundancy exists.
According to the technical scheme, the method for obtaining the model parameters of the polynomial objective function specifically comprises the following steps of: fitting the real-time sample data and the history sample data of the test second hand cart to the established polynomial objective function through a plurality of discrete variables respectively converted, and obtaining model parameters of the polynomial objective function after sparsification through regularization treatment after fitting; the L1 norms of the polynomial parameter vectors are limited by adopting L1 regularization, so that the purpose of sparsification is achieved, and the model complexity of the polynomial objective function is reduced.
While the foregoing description of the embodiments of the present invention has been presented in conjunction with the drawings, it should be understood that it is not intended to limit the scope of the invention, but rather, it is intended to cover all modifications or variations within the scope of the invention as defined by the claims of the present invention.
Claims (10)
1. The second-hand vehicle risk prediction method based on polynomial modeling is characterized by comprising the following steps of:
acquiring real-time data and historical data of the test cart, and mapping the acquired real-time data and historical data of the test cart into discrete variables respectively, wherein each type of real-time data or each type of historical data corresponds to one discrete variable;
Based on a polynomial regression method, a polynomial objective function between a risk prediction value of the test second-hand vehicle and a plurality of discrete variables respectively mapped by real-time data and historical data is established;
fitting the established polynomial objective function aiming at the real-time sample data and the history sample data of the test second hand cart to obtain model parameters of the polynomial objective function;
and acquiring real-time data and historical data of the target second hand vehicle, inputting the real-time data and the historical data of the acquired target second hand vehicle into a polynomial objective function, and outputting a risk prediction value of the target second hand vehicle.
2. The method for predicting the risk of a second-hand vehicle based on polynomial modeling according to claim 1, wherein the real-time data of the test second-hand vehicle comprises vehicle acceleration time, braking distance, steering wheel angle, steering torque and steering angular velocity.
3. The method for predicting the risk of a second-hand vehicle based on polynomial modeling of claim 1, wherein the historical data of the test second-hand vehicle comprises a maintenance number, a number of vehicle accidents, a vehicle age, a driving mileage and a vehicle use environment.
4. A method for predicting risk of a second-hand vehicle based on polynomial modeling according to any one of claims 1-3, further comprising, before mapping the acquired real-time data and history data of the test second-hand vehicle into discrete variables, respectively:
Data cleaning is carried out on the real-time data and the historical data, wherein the data cleaning is specifically carried out on the real-time data and the historical data: and deleting the history data with completely overlapped data entries, and filling the deleted history data or real-time data by adopting an average value under the same data.
5. A second-hand car risk prediction method based on polynomial modeling according to any one of claims 1-3, characterized in that before establishing a polynomial objective function between the risk prediction value of the test second-hand car and discrete variables mapped respectively by real-time data and historical data based on a polynomial regression method, the method further comprises:
calculating the variance value of each discrete variable mapped by different types of historical data respectively, and deleting the discrete variables with variances smaller than a preset variance threshold; the variance value of each discrete variable of the different types of historical data or real-time data mapping is calculated specifically as follows:，
wherein ,for the variance value of the current discrete variable mapped by a certain type of historical data or real-time data, n is the number of historical data or real-time data entries corresponding to the current discrete variable, < >>Representing the value of the ith historical data or real-time data corresponding to the current discrete variable,/or- >Representing the average value of all the historical data or real-time data corresponding to the current discrete variable.
6. The method for predicting the risk of a second-hand vehicle based on polynomial modeling according to claim 5, wherein before establishing the polynomial objective function between the risk prediction value of the second-hand vehicle and the discrete variables mapped by the real-time data and the historical data respectively based on the polynomial regression method, the method further comprises:
calculating a correlation coefficient between any two discrete variables, and deleting the discrete variables with the correlation coefficients larger than a preset correlation coefficient threshold; the method specifically calculates the correlation coefficient between any two discrete variables as follows:，
wherein ,a value representing the ith history data or real-time data corresponding to the current discrete variable,/-, for example>Representing the value of the ith historical data or real-time data corresponding to the current another discrete variable, n is the entry number of the historical data or real-time data corresponding to the current two discrete variables,/for>Is the correlation coefficient between two discrete variables.
7. The method for predicting the risk of the second-hand vehicle based on polynomial modeling according to claim 6, wherein the step of establishing a polynomial objective function between the risk prediction value of the test second-hand vehicle and a plurality of discrete variables respectively mapped by real-time data and historical data based on a polynomial regression method specifically comprises:
After variable deletion and filtration, the number of the remaining discrete variables is k, and the polynomial objective function is:
，
wherein k is the number of discrete variables remaining after the deletion of the variables,is a constant term->For the primary term coefficient->Is the polynomial coefficient of d and f, representing the interaction between the d-th discrete variable and the f-th discrete variable, h is the polynomial degree, +.>Is the value of the d-th discrete variable, +.>To the power f of the value of the d-th discrete variable.
8. The method for predicting the risk of a second-hand vehicle based on polynomial modeling of claim 7, wherein the polynomial coefficientsThe range of the values of (2) is determined by polynomial degree h.
9. The method for predicting the risk of a second-hand vehicle based on polynomial modeling according to claim 7, wherein fitting the established polynomial objective function to the real-time sample data and the history sample data of the second-hand vehicle to obtain the model parameters of the polynomial objective function specifically comprises:
fitting the real-time sample data and the history sample data of the test second hand cart to the established polynomial objective function through a plurality of discrete variables respectively converted, and obtaining model parameters of the polynomial objective function after sparsification through regularization treatment after fitting; the regularization process specifically includes:
，
Wherein m is the total number of real-time sample data and historical sample data of the test cart corresponding to the remaining discrete variable,is the predicted value of the first real-time sample data or the historical sample data of the test second cart corresponding to the input residual discrete variable, +.>Is the first real-time sample data or the historical sample number of the test second cart corresponding to the input residual discrete variableAccording to the actual output value +.>Is a regularization parameter, min imize is a minimization function.
10. A second-hand car risk prediction system based on polynomial modeling, comprising:
the acquisition module is used for acquiring real-time data and historical data of the test second hand cart, and mapping the acquired real-time data and historical data of the test second hand cart into discrete variables respectively, wherein each type of real-time data or each type of historical data corresponds to one discrete variable;
the building module is used for building a polynomial objective function between a risk prediction value of the test second-hand vehicle and a plurality of discrete variables respectively mapped by the real-time data and the historical data based on a polynomial regression method;
the fitting module is used for fitting the established polynomial objective function aiming at the real-time sample data and the history sample data of the test second hand cart to obtain model parameters of the polynomial objective function;
And the output module is used for acquiring real-time data and historical data of the target second hand vehicle, inputting the acquired real-time data and the historical data of the target second hand vehicle into the polynomial objective function and outputting a risk prediction value of the target second hand vehicle.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310712379.8A CN116452008B (en) | 2023-06-16 | 2023-06-16 | Second-hand vehicle risk prediction method and system based on polynomial modeling |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310712379.8A CN116452008B (en) | 2023-06-16 | 2023-06-16 | Second-hand vehicle risk prediction method and system based on polynomial modeling |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116452008A true CN116452008A (en) | 2023-07-18 |
| CN116452008B CN116452008B (en) | 2023-08-29 |
Family
ID=87134116
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310712379.8A Active CN116452008B (en) | 2023-06-16 | 2023-06-16 | Second-hand vehicle risk prediction method and system based on polynomial modeling |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116452008B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117789373A (en) * | 2023-12-29 | 2024-03-29 | 山东四季车网络科技有限公司 | Shared carwash user data safety management method |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA3037941A1 (en) * | 2018-03-23 | 2019-09-23 | NthGen Software Inc. | Method and system for generating and using vehicle pricing models |
| CN110288369A (en) * | 2019-05-09 | 2019-09-27 | 北京善义善美科技有限公司 | A kind of used car is to calibration valence system and method |
| CN112446737A (en) * | 2020-12-03 | 2021-03-05 | 天津狮拓信息技术有限公司 | Method and system for evaluating and pricing residual value of second-hand commercial vehicle |
| CN113191805A (en) * | 2021-04-30 | 2021-07-30 | 平安科技（深圳）有限公司 | Vehicle owner replacement evaluation method, system, electronic equipment and storage medium |
| CN113532886A (en) * | 2021-08-18 | 2021-10-22 | 成都工业职业技术学院 | Second-hand vehicle detection system and detection method |
| CN115115414A (en) * | 2022-07-17 | 2022-09-27 | 扬州大学 | Second-hand car valuation method based on machine learning |
| WO2022267456A1 (en) * | 2021-06-24 | 2022-12-29 | 平安科技（深圳）有限公司 | Vehicle damage assessment method, apparatus and device, and storage medium |
| CN115630572A (en) * | 2022-10-18 | 2023-01-20 | 所托(杭州)汽车智能设备有限公司 | Vehicle maintenance pricing method, equipment and storage medium |
| CN115730964A (en) * | 2022-11-30 | 2023-03-03 | 启明信息技术股份有限公司 | Quality trend prediction based on product quality data analysis |
| CN116245365A (en) * | 2023-03-08 | 2023-06-09 | 北京易华录信息技术股份有限公司 | Vehicle risk assessment method and device, storage medium and electronic equipment |
-
2023
- 2023-06-16 CN CN202310712379.8A patent/CN116452008B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA3037941A1 (en) * | 2018-03-23 | 2019-09-23 | NthGen Software Inc. | Method and system for generating and using vehicle pricing models |
| CN110288369A (en) * | 2019-05-09 | 2019-09-27 | 北京善义善美科技有限公司 | A kind of used car is to calibration valence system and method |
| CN112446737A (en) * | 2020-12-03 | 2021-03-05 | 天津狮拓信息技术有限公司 | Method and system for evaluating and pricing residual value of second-hand commercial vehicle |
| CN113191805A (en) * | 2021-04-30 | 2021-07-30 | 平安科技（深圳）有限公司 | Vehicle owner replacement evaluation method, system, electronic equipment and storage medium |
| WO2022267456A1 (en) * | 2021-06-24 | 2022-12-29 | 平安科技（深圳）有限公司 | Vehicle damage assessment method, apparatus and device, and storage medium |
| CN113532886A (en) * | 2021-08-18 | 2021-10-22 | 成都工业职业技术学院 | Second-hand vehicle detection system and detection method |
| CN115115414A (en) * | 2022-07-17 | 2022-09-27 | 扬州大学 | Second-hand car valuation method based on machine learning |
| CN115630572A (en) * | 2022-10-18 | 2023-01-20 | 所托(杭州)汽车智能设备有限公司 | Vehicle maintenance pricing method, equipment and storage medium |
| CN115730964A (en) * | 2022-11-30 | 2023-03-03 | 启明信息技术股份有限公司 | Quality trend prediction based on product quality data analysis |
| CN116245365A (en) * | 2023-03-08 | 2023-06-09 | 北京易华录信息技术股份有限公司 | Vehicle risk assessment method and device, storage medium and electronic equipment |
Non-Patent Citations (2)
| Title |
|---|
| SHAN JIANG等: "Short-Term Demand Prediction Method for Online Car-Hailing Services Based on a Least Squares Support Vector Machine", 《ACCESS》, no. 7, pages 11882 - 11891, XP011707698, DOI: 10.1109/ACCESS.2019.2891825 * |
| 孙黛珍: "二手车交易价格预测分析 ——基于Stacking 融合模型方法", 《中国优秀硕士学位论文全文数据库（电子期刊） 工程科技Ⅱ辑》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117789373A (en) * | 2023-12-29 | 2024-03-29 | 山东四季车网络科技有限公司 | Shared carwash user data safety management method |
| CN117789373B (en) * | 2023-12-29 | 2024-05-31 | 山东四季车网络科技有限公司 | Shared carwash user data safety management method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116452008B (en) | 2023-08-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109641603B (en) | Abnormality detection device, abnormality detection method, and computer program | |
| CN116452008B (en) | Second-hand vehicle risk prediction method and system based on polynomial modeling | |
| CN108556682B (en) | Driving range prediction method, device and equipment | |
| CN112633532B (en) | Train wheel maintenance management system based on digital twin technology | |
| CN113705412B (en) | Track state detection method and device for multi-source data fusion based on deep learning | |
| CN116108717B (en) | Traffic transportation equipment operation prediction method and device based on digital twin | |
| CN113247041B (en) | Wheel set axle box device state monitoring method based on digital twin model | |
| CN106650157A (en) | Method, device and system for vehicle part fault probability estimation | |
| CN115398062A (en) | Method for generating learned model and road surface feature determination device | |
| CN115291515A (en) | Automatic driving simulation test system and evaluation method based on digital twinning | |
| Mercelis et al. | Towards detection of road weather conditions using large-scale vehicle fleets | |
| CN115034821A (en) | Vehicle estimation method and device, computer equipment and storage medium | |
| CN106780049B (en) | Vehicle financial credit granting method and device | |
| CN116358760B (en) | Method, device, equipment and storage medium for measuring load distribution of vehicle tyre | |
| CN112765877A (en) | Method and device for constructing PHM (hybrid construction model) general system of rolling stock | |
| CN114119256A (en) | UBI dangerous chemical vehicle driving behavior acquisition and analysis system and premium discount method | |
| CN115439263A (en) | Accurate vehicle insurance premium evaluation method based on automatic driving behavior assets | |
| CN112937587A (en) | Road feel simulation method based on K-Medoids and classification regression tree | |
| CN117974116B (en) | Vehicle fault auxiliary diagnosis system based on image recognition | |
| Sobrinho et al. | IoT and Big Data Analytics: Under-Rail Maintenance Management at Vitória–Minas Railway | |
| EP4389463A1 (en) | Method and device for determining tire wears | |
| CN116945907B (en) | New energy electric automobile mileage calculation method and system | |
| CN117807412B (en) | Cargo damage prediction system based on vehicle-mounted sensor and big data | |
| CN118243133A (en) | Driving path optimized navigation method and system based on road condition environment recognition | |
| CN114648414B (en) | Automobile tire insurance risk evaluation and control method and 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |