CN113379445A - Vehicle price prediction method and device - Google Patents
Vehicle price prediction method and device Download PDFInfo
- Publication number
- CN113379445A CN113379445A CN202110570772.9A CN202110570772A CN113379445A CN 113379445 A CN113379445 A CN 113379445A CN 202110570772 A CN202110570772 A CN 202110570772A CN 113379445 A CN113379445 A CN 113379445A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- sample
- parameter information
- scene
- price
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000012549 training Methods 0.000 claims abstract description 73
- 238000013528 artificial neural network Methods 0.000 claims description 16
- 238000004590 computer program Methods 0.000 claims description 16
- 238000007781 pre-processing Methods 0.000 claims description 15
- 239000013598 vector Substances 0.000 claims description 15
- 238000003860 storage Methods 0.000 claims description 14
- 238000012216 screening Methods 0.000 claims description 10
- 230000002159 abnormal effect Effects 0.000 claims description 6
- 238000013210 evaluation model Methods 0.000 description 22
- 238000010586 diagram Methods 0.000 description 15
- 238000012545 processing Methods 0.000 description 11
- 230000008569 process Effects 0.000 description 10
- 230000006870 function Effects 0.000 description 8
- 238000011156 evaluation Methods 0.000 description 7
- 230000008859 change Effects 0.000 description 6
- 239000003086 colorant Substances 0.000 description 6
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000005012 migration Effects 0.000 description 2
- 238000013508 migration Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 239000000446 fuel Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 230000002085 persistent effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0206—Price or cost determination based on market factors
-
- 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
- G06Q30/0283—Price estimation or determination
Landscapes
- Business, Economics & Management (AREA)
- Development Economics (AREA)
- Strategic Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Economics (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the application discloses a vehicle price prediction method and device, which are used for solving the problems of inaccurate price prediction and low efficiency of the conventional used vehicle. The method comprises the following steps: determining a scene to be traded of the target second-hand vehicle, and acquiring the designated parameter information of the target second-hand vehicle. Acquiring a first vehicle valuation model corresponding to a scene to be traded; the first vehicle valuation model is obtained by training based on sample vehicle parameter information of a second-hand vehicle of a first sample corresponding to a specific transaction scene and the second vehicle valuation model; the second vehicle valuation model is obtained by training based on sample vehicle parameter information of a second sample used vehicle corresponding to each transaction scene. And predicting the transaction price of the target second-hand vehicle in the scene to be transacted according to the designated parameter information and the first vehicle valuation model. The technical scheme effectively improves the accuracy and efficiency of the second-hand vehicle transaction price prediction, covers various transaction scenes, and realizes pertinence of second-hand vehicle price prediction in different transaction scenes.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a method and a device for predicting vehicle price.
Background
With the increase of the automobile holding capacity and the improvement of the life quality of people, the demand of the used-car trading circulation exists in the life cycle of the automobile, and the demand of the used-car price evaluation service follows. Whether the vehicle is received by a vehicle dealer, purchased by a user, or used for vehicle replacement, mortgage loan business and the like, the price of the transaction vehicle needs to be evaluated, and the evaluated price is close to the final transaction price as much as possible so as to ensure fair and fair transaction. For example, for car owners selling cars, the risk of being killed may be reduced based on the used car valuation; for the car dealer, the profit space maximization can be confirmed according to the difference between the purchase price and the retail price, the purchase and retail transaction can be completed quickly, and the fund and inventory risks are reduced. For another example, for related businesses such as financial loan on second-hand vehicles in banks, insurance company residual insurance business, replacement vehicle handling in host plants, and the like, second-hand vehicle valuation service is required to assist them in improving business efficiency and reducing cost.
In the prior art, two general second-hand car valuation methods are available: one is to perform offline detection of the vehicle by a professional vehicle evaluator or evaluation team and then give an evaluation price; most evaluators are only familiar with one or a plurality of brands of vehicle types, but the vehicles which are exchanged on the market are more, so for the vehicle types which are not familiar to evaluators, the evaluators often cannot evaluate the real price of the vehicle or the difference between the evaluated vehicle price and the real bargaining price is larger, so that a certain party of the transaction loses money or the transaction fails, and meanwhile, the efficiency of manually evaluating the vehicle is lower, the time consumption is longer, and the requirement of a huge second-hand vehicle trading market is difficult to meet. The other method is to estimate the price of the used vehicle through a simple statistical learning model or a machine learning model, and although the method can cover more vehicle types and quickly estimate the price of the used vehicle, the model is simple and cannot accurately depict the change trend of the vehicle price, so that the estimation precision of the vehicle price is low, and the method is difficult to apply in an actual transaction scene.
Therefore, an efficient and accurate used vehicle price evaluation model is needed to evaluate the transaction price of used vehicles.
Disclosure of Invention
The embodiment of the application aims to provide a vehicle price prediction method and a vehicle price prediction device, which are used for solving the problems of inaccurate price prediction and low prediction efficiency of the conventional used vehicle.
In order to solve the above technical problem, the embodiment of the present application is implemented as follows:
in one aspect, an embodiment of the present application provides a method for predicting a vehicle price, including:
determining a scene to be traded of a target second-hand vehicle, and acquiring specified parameter information of the target second-hand vehicle;
acquiring a first vehicle valuation model corresponding to the scene to be traded; the first vehicle valuation model is obtained by training based on sample vehicle parameter information of a second-hand vehicle of a first sample corresponding to a specific transaction scene and a second vehicle valuation model; the second vehicle valuation model is obtained by training based on sample vehicle parameter information of a second sample used vehicle corresponding to each transaction scene;
and predicting the transaction price of the target second-hand vehicle in the scene to be transacted according to the designated parameter information and the first vehicle valuation model.
In another aspect, an embodiment of the present application provides a vehicle price prediction apparatus, including:
the first determination module is used for determining a scene to be traded of a target second-hand vehicle and acquiring the designated parameter information of the target second-hand vehicle;
the first obtaining module is used for obtaining a first vehicle valuation model corresponding to the scene to be traded; the first vehicle valuation model is obtained by training based on sample vehicle parameter information of a second-hand vehicle of a first sample corresponding to a specific transaction scene and a second vehicle valuation model; the second vehicle valuation model is obtained by training based on sample vehicle parameter information of a second sample used vehicle corresponding to each transaction scene;
and the predicting module is used for predicting the transaction price of the target second-hand vehicle in the scene to be transacted according to the designated parameter information and the first vehicle valuation model.
In another aspect, an embodiment of the present application provides a vehicle price predicting device, including a processor and a memory electrically connected to the processor, where the memory stores a computer program, and the processor is configured to call and execute the computer program from the memory to implement: determining a scene to be traded of a target second-hand vehicle, and acquiring specified parameter information of the target second-hand vehicle;
acquiring a first vehicle valuation model corresponding to the scene to be traded; the first vehicle valuation model is obtained by training based on sample vehicle parameter information of a second-hand vehicle of a first sample corresponding to a specific transaction scene and a second vehicle valuation model; the second vehicle valuation model is obtained by training based on sample vehicle parameter information of a second sample used vehicle corresponding to each transaction scene;
and predicting the transaction price of the target second-hand vehicle in the scene to be transacted according to the designated parameter information and the first vehicle valuation model.
In another aspect, an embodiment of the present application provides a storage medium for storing a computer program, where the computer program is executed by a processor to implement the following processes: determining a scene to be traded of a target second-hand vehicle, and acquiring specified parameter information of the target second-hand vehicle;
acquiring a first vehicle valuation model corresponding to the scene to be traded; the first vehicle valuation model is obtained by training based on sample vehicle parameter information of a second-hand vehicle of a first sample corresponding to a specific transaction scene and a second vehicle valuation model; the second vehicle valuation model is obtained by training based on sample vehicle parameter information of a second sample used vehicle corresponding to each transaction scene;
and predicting the transaction price of the target second-hand vehicle in the scene to be transacted according to the designated parameter information and the first vehicle valuation model.
By adopting the technical scheme of the embodiment of the invention, the appointed parameter information of the target second-hand vehicle is obtained by determining the scene to be traded of the target second-hand vehicle; and then obtaining a first vehicle evaluation model corresponding to the to-be-traded scene, wherein the first vehicle evaluation model is obtained by training based on sample vehicle parameter information of the second-sample second-hand vehicle corresponding to the specific trading scene and a second vehicle evaluation model, and the second vehicle evaluation model is obtained by training based on sample vehicle related information of the second-sample second-hand vehicle corresponding to each trading scene. And then, predicting the transaction price of the target second-hand vehicle in the scene to be transacted according to the designated parameter information and the first vehicle valuation model. Therefore, according to the technical scheme, the second vehicle valuation model is trained on the basis of the sample vehicle parameter information of the second-hand vehicle corresponding to each transaction scene, the first vehicle valuation model corresponding to the specific transaction scene is trained on the basis of the second vehicle valuation model, so that the price prediction of the second-hand vehicle under the specific transaction scene can be carried out on the second-hand vehicle by using the vehicle valuation model corresponding to the specific transaction scene, and the price prediction of the second-hand vehicle is specifically carried out in the specific transaction scene, so that the prediction accuracy and the prediction efficiency of the transaction price of the second-hand vehicle are effectively improved. Moreover, by training the vehicle valuation models corresponding to various specific trading scenes, the vehicle valuation models can cover various trading scenes, and pertinence of second-hand vehicle price prediction in different trading scenes is achieved.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments described in the present application, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a schematic flow chart diagram of a method of predicting vehicle prices in accordance with an embodiment of the present invention;
FIG. 2 is a schematic flow chart diagram of a method of training a vehicle valuation model in accordance with an embodiment of the invention;
FIG. 3 is a schematic diagram of a model architecture of a vehicle valuation model in accordance with an embodiment of the invention;
FIG. 4 is a schematic block diagram of a vehicle price prediction apparatus according to an embodiment of the present invention;
fig. 5 is a schematic block diagram of a vehicle price prediction apparatus according to an embodiment of the present invention.
Detailed Description
The embodiment of the application provides a vehicle price prediction method and device, which are used for solving the problems of inaccurate price prediction and low prediction efficiency of the conventional used vehicle.
In order to make those skilled in the art better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
Fig. 1 is a schematic flow chart of a vehicle price prediction method according to an embodiment of the present invention, as shown in fig. 1, the method including:
s102, determining a scene to be traded of the target second-hand vehicle, and acquiring the designated parameter information of the target second-hand vehicle.
In the step, when acquiring the designated parameter information of the target second-hand vehicle, the vehicle parameter information of the target second-hand vehicle can be acquired firstly, wherein the vehicle parameter information comprises at least one item of vehicle attribute information, vehicle condition information, macroscopic economic data and vehicle industry index data; and further extracting vehicle parameter information matched with the sample target parameter information from the vehicle parameter information as specified parameter information. The sample target parameter information is information which is screened from a large amount of sample vehicle parameter information and is used for training a vehicle evaluation model, and the vehicle parameter information matched with the sample target parameter information is information which is screened from all vehicle parameter information and belongs to the same type of parameters as the sample target parameter information.
In practical application, the vehicle parameter information can be selected according to requirements, and the richer the information content of the vehicle parameter information is, the more accurate the prediction result of the transaction price of the target second-hand vehicle is. Preferably, the vehicle parameter information includes vehicle attribute information, vehicle condition information, macro-economic data, and vehicle industry index data.
The vehicle attribute information includes at least one of vehicle type information, vehicle color, vehicle operating properties (e.g., operating cars, non-operating cars), and the like. The vehicle type information comprises brand, vehicle series, annual amount, discharge capacity, emission standard, length, width and height of a vehicle body, wheel base, rotating speed, seat number, manufacturer guiding price, fuel type, gearbox type, vehicle series grade, country and the like, can be preset and stored as a vehicle type library, and can be directly obtained through the vehicle type library when the vehicle type information is required to be used.
The vehicle condition information may include at least one of a vehicle registration time, a vehicle registration location, a vehicle travel distance (e.g., kilometers traveled), a number of vehicle passing households, a vehicle condition rating, an accident rating, and the like.
The macro economic data may include GDP (Gross vehicle Product, total Domestic Product), house price, etc. data for the location of the target used-hand vehicle (e.g., country, province, or city). The vehicle industry index data may include statistical data for vehicle related information over a period of time, such as used vehicle sales, new vehicle sales, etc. for each city in the year.
And S104, acquiring a first vehicle valuation model corresponding to the scene to be traded.
The first vehicle valuation model is obtained by training based on sample vehicle parameter information of a second-hand vehicle corresponding to a specific trading scene and the second vehicle valuation model. And the second vehicle valuation model is obtained by training based on sample vehicle related information of the second sample used vehicle corresponding to each transaction scene. Based on the above, different transaction scenes respectively have corresponding second vehicle valuation models, such as a second vehicle valuation model corresponding to a retail scene of a vehicle dealer, a second vehicle valuation model corresponding to an on-line auction scene, a second vehicle valuation model corresponding to a purchase scene of the vehicle dealer, and the like.
And S106, predicting the transaction price of the target second-hand vehicle in the scene to be transacted according to the designated parameter information and the first vehicle evaluation model.
By adopting the technical scheme of the embodiment of the invention, the appointed parameter information of the target second-hand vehicle is obtained by determining the scene to be traded of the target second-hand vehicle; and then obtaining a first vehicle evaluation model corresponding to the to-be-traded scene, wherein the first vehicle evaluation model is obtained by training based on sample vehicle parameter information of the second-sample second-hand vehicle corresponding to the specific trading scene and a second vehicle evaluation model, and the second vehicle evaluation model is obtained by training based on sample vehicle related information of the second-sample second-hand vehicle corresponding to each trading scene. And then, predicting the transaction price of the target second-hand vehicle in the scene to be transacted according to the designated parameter information and the first vehicle valuation model. Therefore, according to the technical scheme, the second vehicle valuation model is trained on the basis of the sample vehicle parameter information of the second-hand vehicle corresponding to each transaction scene, the first vehicle valuation model corresponding to the specific transaction scene is trained on the basis of the second vehicle valuation model, so that the price prediction of the second-hand vehicle under the specific transaction scene can be carried out on the second-hand vehicle by using the vehicle valuation model corresponding to the specific transaction scene, and the price prediction of the second-hand vehicle is specifically carried out in the specific transaction scene, so that the prediction accuracy and the prediction efficiency of the transaction price of the second-hand vehicle are effectively improved. Moreover, by training the vehicle valuation models corresponding to various specific trading scenes, the vehicle valuation models can cover various trading scenes, and pertinence of second-hand vehicle price prediction in different trading scenes is achieved.
In one embodiment, before determining the second vehicle valuation model corresponding to each trading scenario, the second vehicle valuation model is trained in advance. The manner of training the second vehicle valuation model is described in detail below.
In one embodiment, the training mode of the first vehicle valuation model can include steps S202-S206 as shown in FIG. 2:
s202, sample vehicle parameter information of the second sample used vehicle corresponding to the plurality of transaction scenes is obtained.
The transaction scene includes a retail scene of a car dealer, an online auction scene, a purchase scene of the car dealer and the like. The sample vehicle parameter information includes at least one of vehicle transaction information, vehicle attribute information, vehicle condition information, macro-economic data, vehicle industry index data. In practical application, the sample vehicle parameter information can be selected according to requirements, the richer the information content of the sample vehicle parameter information is, the higher the accuracy of the trained vehicle evaluation model is. Preferably, the sample vehicle parameter information includes vehicle transaction information, vehicle attribute information, vehicle condition information, macro-economic data, and vehicle industry index data for the sample used vehicle.
The vehicle transaction information may include at least one of a transaction scene, a transaction area, transaction time, a transaction price, a new vehicle instruction price corresponding to the transaction time (e.g., a new vehicle instruction price in the current month of the transaction), second-hand vehicle branding data corresponding to the transaction time (e.g., second-hand vehicle sales in provinces and cities in the current month of the transaction), and the like. The specific contents of the vehicle attribute information, the vehicle condition information, the macro economic data, and the vehicle industry index data have been described in the above embodiments, and are not repeated here.
S204, determining sample target parameter information for training a second vehicle valuation model according to the sample vehicle parameter information of the second sample second-hand vehicle; and determining the price residual value rate of the second sample used vehicle according to the transaction price of the second sample used vehicle and the corresponding new vehicle guide price.
The sample target parameter information used for training the second vehicle valuation model is parameter information which meets a preset relevant condition with the price residual value rate, and the preset relevant condition comprises at least one of the following items: the contribution degree to the price residual value rate is positioned at the top N, and the contribution degree is higher than a preset contribution threshold value; n is a positive integer.
In one embodiment, in determining the sample target parameter information based on the sample vehicle parameter information, an intermediate model may be trained based on all sample vehicle parameter information of a second sample secondary vehicle, the all sample vehicle parameter information of the second sample secondary vehicle being used as input data and the price residual rate of the second sample secondary vehicle being used as output data during the training. After the intermediate model is trained, the sample vehicle parameter information of the second sample used vehicle is input into the intermediate model, so that the intermediate model outputs the contribution degree of each sample vehicle parameter information to the price residual value rate, the sample vehicle parameter information is sorted according to the contribution degree, and the sample target parameter information meeting the preset relevant condition is screened out according to the sorting result.
In this embodiment, after the sample vehicle parameter information of the second sample second-hand vehicle is input into the intermediate model, the output data of the intermediate model includes the price residual value rate of the second sample second-hand vehicle and the contribution degree of each sample vehicle parameter information to the price residual value rate.
The sample target parameter information determined according to the sample vehicle parameter information may be the sample vehicle parameter information itself, or may be derived information indirectly determined by the sample vehicle parameter information, such as the vehicle age, the average travel distance (e.g., the average number of miles traveled per month), and the like. The sample target parameter information is not uniform, but can be adjusted according to the change of various factors (such as policy, market change, used vehicle inventory, and the like).
The contribution degree corresponding to the sample vehicle parameter information can represent the influence degree of the sample vehicle parameter information on the second-hand vehicle transaction price (or the price residual value rate). In one embodiment, the contribution degree corresponding to the sample vehicle parameter information may be set in advance according to the influence degree of the sample vehicle parameter information on the second-hand vehicle transaction price (or price residual value rate), for example, if the influence degree of the vehicle driving distance on the second-hand vehicle transaction price (or price residual value rate) is high, the contribution degree corresponding to the vehicle driving distance may be set to 100%; the influence degree of the vehicle color on the used vehicle trading price (or the price residual value rate) is second highest, and the contribution degree corresponding to the vehicle color can be set to be 80%; the influence degree of the length, the width and the height of the vehicle body on the second-hand vehicle trading price (or the price residual value rate) is low, and the contribution degree corresponding to the vehicle color can be set to be 20 percent; and so on. Of course, the sample vehicle parameter information is not a constant one, and may be adjusted according to the change of various factors (such as policy, market change, used vehicle inventory, etc.).
The new car guide price corresponding to the sample second-hand vehicle is the new car guide price corresponding to the trading time of the sample second-hand vehicle (for example, the sample second-hand vehicle is in the current trading month). The trading price of the sample second-hand vehicle can be divided by the new vehicle guide price corresponding to the sample second-hand vehicle, namely the price residual rate of the sample second-hand vehicle, and the price residual rate is the target value of model training.
And S206, training a second vehicle valuation model by taking the sample target parameter information as input data and taking the price residual value rate as output data.
In one embodiment, in consideration of the possible situations of missing values, abnormal information and the like in the sample vehicle parameter information, the sample vehicle parameter information meeting the preset condition may be screened out after the sample vehicle parameter information of the second sample used vehicle is acquired (i.e., S202). Wherein the preset condition may include at least one of: the missing rate of the sample vehicle parameter information is greater than or equal to a preset threshold value, and the sample vehicle parameter information belongs to preset abnormal information.
The proportion of the missing quantity in the certain type of sample vehicle parameter information to the total quantity is the missing rate of the sample vehicle parameter information. For example, sample vehicle parameter information of 100 sample secondary vehicles is obtained, wherein the vehicle color of 10 sample secondary vehicles is unknown, and the missing rate of the sample vehicle parameter information "vehicle color" is 10/100, namely 10%.
The preset abnormality information may be empirically set, and for example, the vehicle age exceeds 16 years, and the number of kilometers traveled exceeds 60 kilometers.
In this embodiment, for the sample vehicle parameter information whose loss rate is smaller than the preset threshold, whether to screen out the sample vehicle parameter information may be determined according to the contribution degree of the sample vehicle parameter information to the price residual value rate. If the contribution degree of the sample vehicle parameter information to the price residual value rate is higher than or equal to a second preset threshold value, the importance degree of the sample vehicle parameter information is higher, and the missing sample vehicle parameter information can be filled. Optionally, the evaluation values corresponding to the sample vehicle parameter information that is not missing in the similar sample vehicle parameter information (the evaluation values include the average value and/or the value with the highest occurrence frequency) may be determined, and then the missing sample vehicle parameter information is filled with the evaluation values.
For example, for the vehicle colors of 1000 sample second-hand vehicles, wherein 100 vehicle colors are missing, the vehicle color with the highest occurrence number in the remaining 900 vehicle colors that are not missing can be determined, and then the missing 100 vehicle colors can be filled with the vehicle color with the highest occurrence number. For another example, in the sample vehicle parameter information of 1000 sample second-hand vehicles, 900 sample second-hand vehicles have a running kilometer number, that is, the running kilometer number of 100 sample second-hand vehicles is missing, and in this case, the average value of the running kilometer numbers of the 900 sample second-hand vehicles that are not missing may be calculated, and the missing 100 running kilometers numbers may be filled in with the average value.
Optionally, tag information may be added to the filled sample vehicle parameter information, where the tag information is used to indicate that the sample vehicle parameter information is filled with the evaluation value, and is not directly acquired, so as to provide more sufficient and accurate sample data for model training.
In one embodiment, after determining the price residual rate of the sample used vehicle by performing S204, the price residual rate may be filtered.
One way to screen out the price residual rate is to screen out the price residual rate belonging to the abnormal information in the price residual rate, and in general, since the price residual rate is the ratio of the transaction price of the second-hand vehicle and the guide price of the new vehicle, the ratio should be less than or equal to 1. If the price residual rate is greater than 1, the price residual rate can be considered as belonging to the abnormal information.
The second way to filter out the remaining price rate is to filter out unreasonable remaining price rate, which includes the remaining price rate not matching with the parameter information(s) of the second-hand vehicle. For example, as the age and mileage of a second-hand vehicle increase, the price residual rate of the second-hand vehicle decreases, and the price residual rate falls within a certain fixed interval for a vehicle with a fixed age or mileage. Based on this, according to the experience of the second-hand vehicle industry, the price residual rate intervals (including the upper limit and the lower limit) corresponding to different vehicle ages and driving mileage numbers can be preset according to different vehicle brands and vehicle series, and when the price residual rate is screened, the price residual rate exceeding the price residual rate intervals (such as being lower than the lower limit or higher than the upper limit) corresponding to the vehicle ages and the driving mileage numbers can be screened.
Of course, the trading price intervals corresponding to different vehicle ages and driving mileage can be preset, so that when the price residual rate is screened out, the trading price of the second-hand vehicle can be determined according to the price residual rate, and the price residual rate corresponding to the unreasonable trading price is screened out when the trading price exceeding the trading price interval corresponding to the vehicle ages and the driving mileage is considered to be unreasonable trading price.
In one embodiment, when the sample target parameter information of the second sample used vehicle is used as the input data, and the price residual rate is used as the output data to train the second vehicle valuation model (i.e. S206), the following steps are specifically performed:
step one, preprocessing the sample target parameter information of the second sample used vehicle to obtain characteristic data corresponding to each sample target parameter information.
In this step, the preprocessing may include data encoding, so that the encoded sample target parameter information is processed into data that can be recognized by the model; the preprocessing can also comprise data standardization processing to eliminate the dimensional influence of variables in the model training process, and meanwhile, the values of the features are concentrated near 0, so that the calculated amount is reduced; the preprocessing can also comprise the step of performing binning processing (namely segmentation processing) on the continuous variable, such as binning a plurality of vehicle ages into 1-year vehicle age, 2-year vehicle age and other sections, so as to enhance the robustness of model training and reduce the risk of overfitting of the model; the preprocessing may also include functionally transforming the variables to more balance the transformed feature distributions, thereby increasing the convergence speed of the model training. For different types of sample target parameter information, the most suitable preprocessing method is preferred, and several preprocessing methods matched with the sample target parameter information are listed below.
For continuous variables such as the age, the historical kilometers of driving, the length of the vehicle body and the like, in order to enable the model to learn the nonlinear characteristics, the preprocessing further comprises the step of carrying out nonlinear conversion on the continuous variables, such as the conversion modes of log, e ^ x and the like.
For discrete variables (e.g., text-like data that the model cannot directly recognize), the preprocessing may be to encode the discrete variables. For example, a small number of discrete variables such as a city on which a used vehicle is listed and vehicle colors are thermally encoded, assuming that the vehicle colors include 4 types of red, black, white and blue, a "1" indicates that the vehicle belongs to the color, a "0" indicates that the vehicle does not belong to the color, and the encoding is performed in the order of "red-black-white-blue", the encoding may be "0010" for a white vehicle, and "0001" for a blue vehicle. For another example, for the vehicle model name and other discrete variables with a large number of values, a word nesting preprocessing mode is preferably selected, and word nesting is a method for converting texts into digital vectors for low-dimensional dense representation of the texts, and meanwhile, the similarity relation among the texts can be kept, namely, the distances between the digital vectors of the similar texts are close. Specifically, all vehicle type names are segmented, for each word, 3 segmented words on the left side and the right side of the word are used as input, the word is used as output to train a neural network model, after the training is finished, a connecting parameter vector of the neural network model is low-density dense representation of each segmented word, then vectors corresponding to each segmented word are added and averaged, and vector representation of the vehicle type names is obtained, namely the vehicle type names are mapped into digital vectors from characters, and the vectors can well represent the relationship among the vehicle type names.
Of course, the several encoding methods listed here are merely illustrative examples, and in practical applications, any existing encoding method may be selected to encode the discrete variable so that the encoded information can be identified by the model.
And secondly, splicing the characteristic data corresponding to the target parameter information of each sample, taking the first characteristic vector obtained after splicing as input data, and taking the price residual value rate as output data to carry out model training to obtain a basic valuation model.
And thirdly, splicing the feature data corresponding to the sample target parameter information respectively corresponding to each trading scene by taking each trading scene as a splicing dimension, taking the second feature vector obtained after splicing as input data of the basic evaluation model, and optimizing the basic evaluation model by taking the price residual rate as output data to obtain a second vehicle evaluation model.
If the trading scenes are the retail scene of the car dealer, the on-line auction scene and the purchase scene of the car dealer, each trading scene corresponds to a large amount of sample target parameter information, namely, all the sample target parameter information is divided by taking the trading scenes as dimensions, and then the sample target parameter information corresponding to each trading scene can be obtained.
In this embodiment, the vehicle valuation model is trained by using sample target parameter information with a high contribution degree to the price residual rate, and feature data corresponding to the sample target parameter information is spliced for a dimension based on the trading scene, so that the second vehicle valuation model obtained after model optimization can fully learn features related to the trading scene, including fully learning the correlation between the trading scene and the price residual rate of the second-hand vehicle, thereby facilitating more accurate training of the first vehicle valuation model corresponding to a specific trading scene, and improving the prediction accuracy of the first vehicle valuation model on the trading price of the second-hand vehicle.
In one embodiment, after training the second vehicle valuation model, the first vehicle valuation model can be further trained by:
first, a neural network including a multi-layer network structure is constructed, and the second vehicle evaluation model is taken as the bottommost network structure of the neural network.
Considering the complexity, efficiency and over-fitting of model training, optionally, a neural network comprising a 3-layer network structure is constructed. That is, the second vehicle evaluation model is used as the bottom layer network structure, the bottom layer network structure is consistent with the network structure of the second vehicle evaluation model, a 2-layer network structure is reconstructed on the bottom layer network structure, and parameters of the 2-layer network structure are initialized randomly.
Secondly, regarding any specific trading scene in the multiple trading scenes, taking the sample vehicle parameter information of the second-hand vehicle corresponding to the specific trading scene as input data of a neural network, namely as input data of a second vehicle valuation model of the bottom layer network structure, and taking the price residual value rate of the second-hand vehicle as output data of the neural network for model training to obtain the first vehicle valuation model corresponding to the specific trading scene. The sample vehicle parameter information of the second-hand sample vehicle corresponding to the specific transaction scene can be screened from the sample vehicle parameter information corresponding to all transaction scenes, and the sample vehicle parameter information of the second-hand sample vehicle which conducts transaction in the specific transaction scene can be additionally collected.
Taking a retail scenario as an example, after the neural network is constructed, the second vehicle valuation model is taken as the bottom layer network structure, and the parameters of other layer network structures are initialized randomly. And taking the sample vehicle parameter information of the second-hand vehicle corresponding to the retail scene as input data of the neural network, and taking the price residual rate of the second-hand vehicle as output data to perform model training, so as to obtain a first vehicle valuation model corresponding to the retail scene.
In this embodiment, because the pre-trained second vehicle valuation model has fully learned the correlation between the trading scenario and the price residual rate of the second-hand vehicle, the second vehicle valuation model is migrated to a new neural network, and the model is trained by using the sample vehicle parameter information corresponding to the specific trading scenario as training data, so that the trained first vehicle valuation model can more fully learn the features related to the specific trading scenario on the basis of the second vehicle valuation model, and simultaneously weaken the related features of other trading scenarios except the specific trading scenario, thereby making the price prediction result of the first vehicle valuation model for the vehicle in the specific trading scenario more accurate and more targeted.
Fig. 3 is a schematic diagram of a model structure of a vehicle valuation model according to an embodiment of the invention, and as shown in fig. 3, a transaction scenario includes 3 scenarios of purchasing, retail sale, and auction.
On the left side of the arrow labeled "parameter migration", the model structure of the second vehicle valuation model is shown, wherein the sample vehicle parameter information of the sample second-hand vehicle corresponding to the purchasing, retail and auction scenes, respectively, is taken as a whole and is used as input data of the model structure, and the price residual rate of the sample second-hand vehicle is used as output data "output" of the model structure.
On the right side of the arrow labeled "parameter migration", the model structure of the first vehicle valuation model is shown, which only schematically shows a 2-tier network structure, and in practical applications, different numbers of layers of network structures can be constructed based on considerations on convergence speed, fitting degree, etc. of model training. And directly transferring the model structure of the second vehicle valuation model to the model structure of the first vehicle valuation model to serve as the bottom layer network structure of the model structure. The parameters of other layer network structure (i.e. "hidden layer") can be initialized randomly, and the parameters of the "hidden layer" are optimized continuously by using the error between the predicted price residual value rate and the actual price residual value rate of the model as a loss function in the training process of the model. The smaller the error value, the higher the accuracy of the trained model.
The training process of the first vehicle valuation model corresponding to a specific trading scenario is described in detail above. Based on this, when the trading price of the target second-hand vehicle is predicted, the designated parameter information of the target second-hand vehicle (i.e. the vehicle parameter information which is screened from the vehicle parameter information of the target second-hand vehicle and is matched with the sample target parameter information) can be preprocessed, and each feature data obtained after preprocessing is spliced into a feature vector, and then the feature vector is input into the first vehicle valuation model corresponding to the specific trading scene, so that the first vehicle valuation model outputs the trading price of the target second-hand vehicle in the specific trading scene.
The preprocessing of the target used vehicle can comprise data coding, standardization processing, box separation processing, function transformation, word nesting and the like. Various processing methods have been described in detail in the above embodiments, and are not described herein again.
As can be seen from the foregoing embodiment, the output data of the first vehicle valuation model for predicting the transaction price of the target second-hand vehicle is the price residual rate of the target second-hand vehicle, so that when S106 is executed, by inputting the specified parameter information of the target second-hand vehicle into the first vehicle valuation model corresponding to the scene to be traded, the directly obtained data is the price residual rate corresponding to the target second-hand vehicle. And then the trading price of the target second-hand vehicle in the scene to be traded can be determined according to the price residual value rate and the new vehicle guide price corresponding to the target second-hand vehicle.
In summary, particular embodiments of the present subject matter have been described. Other embodiments are within the scope of the following claims. In some cases, the actions recited in the claims can be performed in a different order and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may be advantageous.
Based on the same idea, the vehicle price prediction method provided in the embodiment of the present application further provides a vehicle price prediction device.
Fig. 4 is a schematic block diagram of a vehicle price prediction apparatus according to an embodiment of the present invention, as shown in fig. 4, including:
the first determining module 410 is used for determining a scene to be traded of a target second-hand vehicle and acquiring designated parameter information of the target second-hand vehicle;
a first obtaining module 420, configured to obtain a first vehicle valuation model corresponding to the scene to be traded; the first vehicle valuation model is obtained by training based on sample vehicle parameter information of a second-hand vehicle of a first sample corresponding to a specific transaction scene and a second vehicle valuation model; the second vehicle valuation model is obtained by training based on sample vehicle parameter information of a second sample used vehicle corresponding to each transaction scene;
and the predicting module 430 is configured to predict the transaction price of the target second-hand vehicle in the scene to be transacted according to the specified parameter information and the first vehicle valuation model.
In one embodiment, the apparatus further comprises:
the second acquisition module is used for acquiring sample vehicle parameter information of a second sample used vehicle corresponding to a plurality of transaction scenes before determining a scene to be transacted of a target used vehicle; the sample vehicle parameter information comprises at least one item of vehicle transaction information, vehicle attribute information, vehicle condition information, macroscopic economic data and vehicle industry index data;
the second determination module is used for determining sample target parameter information used for training the second vehicle valuation model according to the sample vehicle parameter information of the second sample second-hand vehicle; determining the price residual value rate of the second sample used vehicle according to the transaction price of the second sample used vehicle and the corresponding new vehicle guide price;
and the first training module is used for training the second vehicle valuation model by taking the sample target parameter information as input data and taking the price residual value rate as output data.
In one embodiment, the second determining module comprises:
the first screening unit is used for screening out at least one sample target parameter information which meets a preset relevant condition with the price residual value rate from the sample vehicle parameter information of the second sample used vehicle;
wherein the preset relevant condition comprises at least one of the following: the contribution degree to the price residual value rate is located at the top N, and the contribution degree is higher than a preset contribution threshold value; and N is a positive integer.
In one embodiment, the screening unit is further configured to:
taking the sample vehicle parameter information of the second sample secondary vehicle as input data, and taking the price residual value rate of the second sample secondary vehicle as output data to train an intermediate model;
inputting the sample vehicle parameter information of the second sample used vehicle into the intermediate model, so that the intermediate model outputs the contribution degree of each sample vehicle parameter information to the price residual value rate;
and sorting the vehicle parameter information of each sample according to the contribution degree, and screening the target parameter information of the sample according to a sorting result.
In one embodiment, the first training module comprises:
the preprocessing unit is used for preprocessing the sample target parameter information of the second sample used vehicle to obtain characteristic data corresponding to each sample target parameter information;
the model training unit is used for splicing the characteristic data corresponding to the target parameter information of each sample, using the spliced first characteristic vector as input data, and using the price residual value rate as output data to perform model training to obtain a basic valuation model;
and the model optimization unit is used for splicing the feature data corresponding to each transaction scene by taking each transaction scene as a splicing dimension, using a second feature vector obtained after splicing as input data of the basic valuation model, and using the price residual rate as output data to optimize the basic valuation model so as to obtain the second vehicle valuation model.
In one embodiment, the apparatus further comprises:
a building module for building a neural network comprising a multi-layer network structure after training the second vehicle valuation model, and taking the second vehicle valuation model as a bottom-layer network structure of the neural network;
and the second training module is used for taking the sample vehicle parameter information of the second sample vehicle corresponding to the specific trading scene as the input data of the neural network and taking the price residual rate of the second sample vehicle as the output data of the neural network for model training to obtain the first vehicle valuation model corresponding to the specific trading scene.
In one embodiment, the first obtaining module 420 includes:
the acquisition unit is used for acquiring vehicle parameter information of the target second-hand vehicle; the vehicle parameter information comprises at least one item of vehicle attribute information, vehicle condition information, macro economic data and vehicle industry index data;
an extraction unit configured to extract the specified parameter information that matches the sample target parameter information from the vehicle parameter information.
In one embodiment, the apparatus further comprises:
the second screening unit is used for screening the sample vehicle parameter information meeting preset conditions after the sample vehicle parameter information of the second sample used vehicle corresponding to the plurality of transaction scenes is obtained;
wherein the preset condition comprises at least one of the following: the missing rate of the sample vehicle parameter information is greater than or equal to a preset threshold value, and the sample vehicle parameter information belongs to preset abnormal information.
It should be understood by those skilled in the art that the vehicle price predicting device in fig. 4 can be used to implement the vehicle price predicting method described above, wherein the detailed description thereof should be similar to that of the method described above, and in order to avoid complexity, the detailed description thereof is omitted.
By adopting the device provided by the embodiment of the invention, the appointed parameter information of the target second-hand vehicle is obtained by determining the scene to be traded of the target second-hand vehicle; and then obtaining a first vehicle evaluation model corresponding to the to-be-traded scene, wherein the first vehicle evaluation model is obtained by training based on sample vehicle parameter information of the second-sample second-hand vehicle corresponding to the specific trading scene and a second vehicle evaluation model, and the second vehicle evaluation model is obtained by training based on sample vehicle related information of the second-sample second-hand vehicle corresponding to each trading scene. And then, predicting the transaction price of the target second-hand vehicle in the scene to be transacted according to the designated parameter information and the first vehicle valuation model. Therefore, the device trains the second vehicle valuation model based on the sample vehicle parameter information of the sample second-hand vehicle corresponding to each transaction scene, trains the first vehicle valuation model corresponding to the specific transaction scene based on the second vehicle valuation model, can perform price prediction on the second-hand vehicle under the specific transaction scene by using the vehicle valuation model corresponding to the specific transaction scene, and specifically predicts the price of the second-hand vehicle in the specific transaction scene, so that the prediction accuracy and the prediction efficiency of the transaction price of the second-hand vehicle are effectively improved. Moreover, by training the vehicle valuation models corresponding to various specific trading scenes, the vehicle valuation models can cover various trading scenes, and pertinence of second-hand vehicle price prediction in different trading scenes is achieved.
Based on the same idea, the embodiment of the present application further provides a device for predicting the price of a vehicle, as shown in fig. 5. The vehicle price prediction device may have a large difference due to different configurations or performances, and may include one or more processors 501 and a memory 502, and the memory 502 may store one or more stored applications or data. Memory 502 may be, among other things, transient or persistent storage. The application stored in memory 502 may include one or more modules (not shown), each of which may include a series of computer-executable instructions in a device for predicting vehicle prices. Still further, the processor 501 may be configured to communicate with the memory 502 to execute a series of computer-executable instructions in the memory 502 on a vehicle price prediction device. The vehicle price prediction apparatus may also include one or more power supplies 503, one or more wired or wireless network interfaces 504, one or more input-output interfaces 505, and one or more keyboards 506.
In particular, in this embodiment, the apparatus for predicting a price of a vehicle comprises a memory, and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs may comprise one or more modules, and each module may comprise a series of computer-executable instructions in the apparatus for predicting a price of a vehicle, and the one or more programs configured to be executed by one or more processors comprise computer-executable instructions for:
determining a scene to be traded of a target second-hand vehicle, and acquiring specified parameter information of the target second-hand vehicle;
acquiring a first vehicle valuation model corresponding to the scene to be traded; the first vehicle valuation model is obtained by training based on sample vehicle parameter information of a second-hand vehicle of a first sample corresponding to a specific transaction scene and a second vehicle valuation model; the second vehicle valuation model is obtained by training based on sample vehicle parameter information of a second sample used vehicle corresponding to each transaction scene;
and predicting the transaction price of the target second-hand vehicle in the scene to be transacted according to the designated parameter information and the first vehicle valuation model.
The embodiment of the present application further provides a storage medium, where the storage medium stores one or more computer programs, where the one or more computer programs include instructions, and when the instructions are executed by an electronic device including multiple application programs, the electronic device can execute each process of the above-mentioned vehicle price prediction method embodiment, and can achieve the same technical effect, and in order to avoid repetition, details are not repeated here.
The systems, devices, modules or units illustrated in the above embodiments may be implemented by a computer chip or an entity, or by a product with certain functions. One typical implementation device is a computer. In particular, the computer may be, for example, a personal computer, a laptop computer, a cellular telephone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.
For convenience of description, the above devices are described as being divided into various units by function, and are described separately. Of course, the functionality of the units may be implemented in one or more software and/or hardware when implementing the present application.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
The memory may include forms of volatile memory in a computer readable medium, Random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.
Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.
The above description is only an example of the present application and is not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.
Claims (11)
1. A method for predicting a price of a vehicle, comprising:
determining a scene to be traded of a target second-hand vehicle, and acquiring specified parameter information of the target second-hand vehicle;
acquiring a first vehicle valuation model corresponding to the scene to be traded; the first vehicle valuation model is obtained by training based on sample vehicle parameter information of a second-hand vehicle of a first sample corresponding to a specific transaction scene and a second vehicle valuation model; the second vehicle valuation model is obtained by training based on sample vehicle parameter information of a second sample used vehicle corresponding to each transaction scene;
and predicting the transaction price of the target second-hand vehicle in the scene to be transacted according to the designated parameter information and the first vehicle valuation model.
2. The method of claim 1, wherein prior to determining the scene of the target second-hand vehicle to be traded, further comprising:
obtaining sample vehicle parameter information of the second sample used vehicles corresponding to the plurality of transaction scenes; the sample vehicle parameter information comprises at least one item of vehicle transaction information, vehicle attribute information, vehicle condition information, macroscopic economic data and vehicle industry index data;
determining sample target parameter information for training the second vehicle valuation model according to the sample vehicle parameter information of the second sample used vehicle; determining the price residual value rate of the second sample used vehicle according to the transaction price of the second sample used vehicle and the corresponding new vehicle guide price;
and training the second vehicle valuation model by taking the sample target parameter information as input data and the price residual value rate as output data.
3. The method of claim 2, wherein determining sample target parameter information for training the second vehicle valuation model based on the sample vehicle parameter information for the second sample used vehicle comprises:
screening out at least one sample target parameter information which meets a preset relevant condition with the price residual value rate from the sample vehicle parameter information of the second sample used vehicle;
wherein the preset relevant condition comprises at least one of the following: the contribution degree to the price residual value rate is located at the top N, and the contribution degree is higher than a preset contribution threshold value; and N is a positive integer.
4. The method according to claim 3, wherein the screening out at least one sample target parameter information satisfying a preset correlation condition with the price residual value rate from the sample vehicle parameter information of the second sample used vehicle comprises:
taking the sample vehicle parameter information of the second sample secondary vehicle as input data, and taking the price residual value rate of the second sample secondary vehicle as output data to train an intermediate model;
inputting the sample vehicle parameter information of the second sample used vehicle into the intermediate model, so that the intermediate model outputs the contribution degree of each sample vehicle parameter information to the price residual value rate;
and sorting the vehicle parameter information of each sample according to the contribution degree, and screening the target parameter information of the sample according to a sorting result.
5. The method of claim 3, wherein training the second vehicle valuation model using the sample target parameter information as input data and the price residual rate as output data comprises:
preprocessing the sample target parameter information of the second sample used vehicle to obtain characteristic data corresponding to each sample target parameter information;
splicing the characteristic data corresponding to the target parameter information of each sample, taking the first characteristic vector obtained after splicing as input data, and taking the price residual value rate as output data to carry out model training to obtain a basic valuation model;
and splicing the characteristic data corresponding to each transaction scene by taking each transaction scene as a splicing dimension, taking a second characteristic vector obtained after splicing as input data of the basic valuation model, and taking the price residual value rate as output data to optimize the basic valuation model to obtain the second vehicle valuation model.
6. The method of claim 2, wherein after training the second vehicle valuation model, further comprising:
constructing a neural network comprising a multi-layer network structure, and taking the second vehicle valuation model as the bottom layer network structure of the neural network;
according to any specific trading scenario in the trading scenarios, sample vehicle parameter information of the second-hand sample vehicle corresponding to the specific trading scenario is used as input data of the neural network, and the price residual rate of the second-hand sample vehicle is used as output data of the neural network for model training to obtain the first vehicle valuation model corresponding to the specific trading scenario.
7. The method of claim 2, wherein the obtaining of the designated parameter information of the target used vehicle comprises:
acquiring vehicle parameter information of the target second-hand vehicle; the vehicle parameter information comprises at least one item of vehicle attribute information, vehicle condition information, macro economic data and vehicle industry index data;
and extracting the specified parameter information matched with the sample target parameter information from the vehicle parameter information.
8. The method of claim 2, wherein after obtaining sample vehicle parameter information for the second sample used vehicle corresponding to the plurality of transaction scenarios, further comprising:
screening out the sample vehicle parameter information meeting preset conditions;
wherein the preset condition comprises at least one of the following: the missing rate of the sample vehicle parameter information is greater than or equal to a preset threshold value, and the sample vehicle parameter information belongs to preset abnormal information.
9. A vehicle price prediction apparatus characterized by comprising:
the first determination module is used for determining a scene to be traded of a target second-hand vehicle and acquiring the designated parameter information of the target second-hand vehicle;
the first obtaining module is used for obtaining a first vehicle valuation model corresponding to the scene to be traded; the first vehicle valuation model is obtained by training based on sample vehicle parameter information of a second-hand vehicle of a first sample corresponding to a specific transaction scene and a second vehicle valuation model; the second vehicle valuation model is obtained by training based on sample vehicle parameter information of a second sample used vehicle corresponding to each transaction scene;
and the predicting module is used for predicting the transaction price of the target second-hand vehicle in the scene to be transacted according to the designated parameter information and the first vehicle valuation model.
10. An apparatus for predicting vehicle prices, comprising a processor and a memory electrically connected to the processor, the memory storing a computer program, the processor being configured to invoke and execute the computer program from the memory to implement:
determining a scene to be traded of a target second-hand vehicle, and acquiring specified parameter information of the target second-hand vehicle;
acquiring a first vehicle valuation model corresponding to the scene to be traded; the first vehicle valuation model is obtained by training based on sample vehicle parameter information of a second-hand vehicle of a first sample corresponding to a specific transaction scene and a second vehicle valuation model; the second vehicle valuation model is obtained by training based on sample vehicle parameter information of a second sample used vehicle corresponding to each transaction scene;
and predicting the transaction price of the target second-hand vehicle in the scene to be transacted according to the designated parameter information and the first vehicle valuation model.
11. A storage medium for storing a computer program which, when executed by a processor, implements the following:
determining a scene to be traded of a target second-hand vehicle, and acquiring specified parameter information of the target second-hand vehicle;
acquiring a first vehicle valuation model corresponding to the scene to be traded; the first vehicle valuation model is obtained by training based on sample vehicle parameter information of a second-hand vehicle of a first sample corresponding to a specific transaction scene and a second vehicle valuation model; the second vehicle valuation model is obtained by training based on sample vehicle parameter information of a second sample used vehicle corresponding to each transaction scene;
and predicting the transaction price of the target second-hand vehicle in the scene to be transacted according to the designated parameter information and the first vehicle valuation model.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110570772.9A CN113379445A (en) | 2021-05-25 | 2021-05-25 | Vehicle price prediction method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110570772.9A CN113379445A (en) | 2021-05-25 | 2021-05-25 | Vehicle price prediction method and device |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113379445A true CN113379445A (en) | 2021-09-10 |
Family
ID=77571832
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110570772.9A Pending CN113379445A (en) | 2021-05-25 | 2021-05-25 | Vehicle price prediction method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113379445A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114331573A (en) * | 2022-03-15 | 2022-04-12 | 蜗牛货车网(山东)电子商务有限公司 | Vehicle residual value evaluation method based on big data and trading platform |
CN115600942A (en) * | 2022-12-15 | 2023-01-13 | 中汽传媒(天津)有限公司(Cn) | Automobile part transaction management method and system |
CN115984002A (en) * | 2023-02-22 | 2023-04-18 | 上海信宝博通电子商务有限公司 | Data processing method and device for vehicle transaction management |
CN116883029A (en) * | 2023-07-13 | 2023-10-13 | 上海信宝博通电子商务有限公司 | Object estimation method, device, equipment and medium |
-
2021
- 2021-05-25 CN CN202110570772.9A patent/CN113379445A/en active Pending
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114331573A (en) * | 2022-03-15 | 2022-04-12 | 蜗牛货车网(山东)电子商务有限公司 | Vehicle residual value evaluation method based on big data and trading platform |
CN115600942A (en) * | 2022-12-15 | 2023-01-13 | 中汽传媒(天津)有限公司(Cn) | Automobile part transaction management method and system |
CN115600942B (en) * | 2022-12-15 | 2023-03-31 | 中汽传媒(天津)有限公司 | Automobile part transaction management method and system |
CN115984002A (en) * | 2023-02-22 | 2023-04-18 | 上海信宝博通电子商务有限公司 | Data processing method and device for vehicle transaction management |
CN115984002B (en) * | 2023-02-22 | 2024-01-16 | 上海信宝博通电子商务有限公司 | Data processing method and device for vehicle transaction management |
CN116883029A (en) * | 2023-07-13 | 2023-10-13 | 上海信宝博通电子商务有限公司 | Object estimation method, device, equipment and medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113379445A (en) | Vehicle price prediction method and device | |
Olson | Perspective: the green innovation value chain: a tool for evaluating the diffusion prospects of green products | |
Kumar et al. | Artificial intelligence in insurance sector | |
Grize et al. | Machine learning applications in nonlife insurance | |
CN113159837A (en) | Vehicle price evaluation method and device | |
CN111582538A (en) | Community value prediction method and system based on graph neural network | |
CN111047193A (en) | Enterprise credit scoring model generation algorithm based on credit big data label | |
CN113674040B (en) | Vehicle quotation method, computer device and computer-readable storage medium | |
CN111105311A (en) | Dynamic credit rating method and device for bond body | |
Chaudhary et al. | The function of driver categorisation in the ride-hailing industry: A study on on-demand transport | |
WO2023225529A2 (en) | Predictive systems and processes for product attribute research and development | |
CN117094781A (en) | Intelligent vehicle insurance pricing and claim settlement processing method and system | |
Vinokurov et al. | Estimating the economic effects of reducing non-tariff barriers in the EEU | |
CN114529400A (en) | Consumption loan preauthorization evaluation method, device and medium | |
CN111383060A (en) | Vehicle price determination method and device, electronic equipment and storage medium | |
Yuping et al. | New methods of customer segmentation and individual credit evaluation based on machine learning | |
CN111667307B (en) | Method and device for predicting financial product sales volume | |
CN113408627A (en) | Target object determination method and device and server | |
Cho et al. | Efficiency comparison of public bike-sharing repositioning strategies based on predicted demand patterns | |
Pinheiro et al. | Introduction to Statistical and Machine Learning Methods for Data Science | |
Bauer et al. | Corporate income tax challenges arising from digitalised business models | |
CN114741592A (en) | Product recommendation method, device and medium based on multi-model fusion | |
CN110659920B (en) | Method and device for recommending service object | |
CN114219084B (en) | Sales visit display counterfeiting identification method and device in fast moving industry | |
CN110428281A (en) | Combine the method and apparatus for determining peers amount for a variety of related products |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210910 |
|
RJ01 | Rejection of invention patent application after publication |