CN108053075A - A kind of scrap-car Forecasting Methodology and system - Google Patents
A kind of scrap-car Forecasting Methodology and system Download PDFInfo
- Publication number
- CN108053075A CN108053075A CN201711445897.9A CN201711445897A CN108053075A CN 108053075 A CN108053075 A CN 108053075A CN 201711445897 A CN201711445897 A CN 201711445897A CN 108053075 A CN108053075 A CN 108053075A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- preset time
- time period
- module
- scrap
- 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 21
- 238000012549 training Methods 0.000 claims abstract description 67
- 238000012360 testing method Methods 0.000 claims abstract description 54
- 238000012423 maintenance Methods 0.000 claims description 44
- 238000000605 extraction Methods 0.000 claims description 15
- 238000004422 calculation algorithm Methods 0.000 claims description 11
- 238000007637 random forest analysis Methods 0.000 claims description 11
- 238000007781 pre-processing Methods 0.000 claims description 10
- 239000000284 extract Substances 0.000 claims description 3
- 238000002203 pretreatment Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 230000008439 repair process Effects 0.000 claims description 2
- 206010068052 Mosaicism Diseases 0.000 abstract description 5
- 210000003765 sex chromosome Anatomy 0.000 abstract description 5
- 238000010801 machine learning Methods 0.000 abstract description 4
- 230000006870 function Effects 0.000 description 5
- 230000008901 benefit Effects 0.000 description 4
- 238000012935 Averaging Methods 0.000 description 1
- 241001269238 Data Species 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000001960 triggered effect Effects 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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Human Resources & Organizations (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Tourism & Hospitality (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- General Business, Economics & Management (AREA)
- Game Theory and Decision Science (AREA)
- Artificial Intelligence (AREA)
- Primary Health Care (AREA)
- Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Educational Administration (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Train Traffic Observation, Control, And Security (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses the Forecasting Methodologies and system of a kind of scrap-car, belong to intelligent transportation big data field.The described method includes：The characteristic information of each vehicle is extracted in the track data of each vehicle in the first preset time period；According to the characteristic information of each vehicle and default scrap-car standard generation training sample, and training sample is divided into training set and test set；Prediction model is trained according to training set, and prediction model is tested using test set；Use following scrap-car of the prediction model prediction after test.In the present invention, based on the historical trajectory data in each vehicle regular period, the training of prediction model is carried out so that the prediction model and prediction result finally obtained be all more scientific and reliability；It is combined simultaneously with machine learning, finds out operation rule of the vehicle before scrapping, solve hysteresis sex chromosome mosaicism of the prior art；For the manufacturer of each vehicle, marketing strategy also can be more accurately determined.
Description
Technical field
The present invention relates to intelligent transportation big data field more particularly to the Forecasting Methodologies and system of a kind of scrap-car.
Background technology
With the fast development of social economy, either private car or stream carrier vehicle all flies with exponential speed
Speed increases；Each vehicle manufacturers are also constantly producing and are selling various brands, the vehicle of various functions, all size.It is right
For each vehicle manufacturers, timely understanding the demand of car purchaser can make it produce more popular vehicle；Meanwhile
Scrap-car in the market can be timely predicted, so as to which demand of clearly marketing is also equal heavy for vehicle manufacturers
It will.And the Forecasting Methodology of existing scrap-car, typically only consider the off-duty number of days of vehicle, and set using subjective experience
It puts and scraps rule；However judge whether vehicle is scrapped by vehicle off-duty number of days, for vehicle manufacturers, often
Through losing optimal marketing opportunity；Meanwhile rule is scrapped using subjective experience setting, it is excessively subjective to also result in prediction result,
Without convincingness, and there are certain hysteresis sex chromosome mosaicisms.It can be seen that how accurate, the prediction scrap-car of science, to each vehicle
Manufacturer for, be still a urgent problem.
The content of the invention
To solve the deficiencies in the prior art, the present invention provides a kind of Forecasting Methodology and system of scrap-car.
On the one hand, the present invention provides a kind of Forecasting Methodology of scrap-car, including：
Step S1：The characteristic information of each vehicle is extracted in the track data of each vehicle in the first preset time period；
Step S2：According to the characteristic information of each vehicle and default scrap-car standard generation training sample, and by institute
It states training sample and is divided into training set and test set；
Step S3：Prediction model is trained according to the training set, and using the test set to the prediction model into
Row test；
Step S4：Use following scrap-car of the prediction model prediction after test.
Optionally, before the step S1, further include：The track data of each vehicle in first preset time period is carried out pre-
Processing；
Accordingly, the step S1, specially：The characteristic information of each vehicle is extracted in track data after the pre-treatment.
Optionally, the track data includes：A series of position data and travel speed；
The step S1, specifically includes：
First preset time period is divided into M the second preset time periods, and according to time order and function order by preceding n
Second preset time period is as the 3rd preset time period, wherein n<M；
According to the position data and travel speed contained in the track data of each vehicle, it is default each second to count each vehicle
Operation number of days, maintenance station in period stop number and stop duration；
When stopping number according to operation number of days of each vehicle in each second preset time period, maintenance station and stop
It is long, determine that operation number of days variation tendency of each vehicle in the 3rd preset time period, maintenance station are stopped number and averagely stopped
By duration.
According to operation number of days of each vehicle in each second preset time period and default stoppage in transit standard, each vehicle is determined
Stoppage in transit situation in each second preset time period.
Optionally, the step S2, specifically includes：
Step S2-1：According to stoppage in transit situation of each vehicle in each second preset time period and default scrap-car mark
Standard determines scrap-car and operation vehicle in each vehicle；
Step S2-2：By operation number of days variation tendency of each vehicle in scrap-car in the 3rd preset time period,
Maintenance station stops number and averagely stops duration establishes correspondence and as positive sample with corresponding vehicle, obtains positive sample collection
It closes；Operation number of days variation tendency of each vehicle in the 3rd preset time period in vehicle will be run, number is stopped in maintenance station
And averagely stop duration and establish correspondence and as negative sample with corresponding vehicle, obtain negative sample set；
Step S2-3：Positive sample is chosen in the positive sample set according to the first preset ratio, and in the negative sample
Negative sample is chosen in set, obtains the training sample of default quantity；
Step S2-4：The training sample of the default quantity is divided by training set and test according to the second preset ratio
Collection.
Optionally, the step S3, specifically includes：
Step S3-1：The training set is trained to obtain prediction model according to random forests algorithm；
Step S3-2：The prediction model is assessed according to Receiver operating curve and adjusts model parameter,
When the first area under Receiver operating curve meets predetermined threshold value, corresponding prediction model is exported；
Step S3-3：The prediction model of calculating output is on the test set under corresponding Receiver operating curve
Second area when the second area meets the predetermined threshold value, performs step S4；Otherwise return to step S3-1；
Accordingly, the step S4, specially：Using current time as deadline, count forward the described 3rd it is default when
Between in section the operation number of days variation tendency of each vehicle, maintenance station stop number and averagely stop duration, and using pre- after test
It surveys the scrap-car in model prediction future and scraps probability.
On the other hand, the present invention provides a kind of forecasting system of scrap-car, including：
Extraction module, for extracting the feature of each vehicle letter in the track data of each vehicle in the first preset time period
Breath；
Generation module, for the characteristic information of each vehicle extracted according to the extraction module and default scrap-car standard
Generate training sample；
Division module, the training sample for the generation module to be generated are divided into training set and test set；
Training module, the training set for being divided according to the division module train prediction model；
Test module, for the prediction model trained using the test set of division module division to the training module
It is tested；
Prediction module, for following scrap-car of the prediction model prediction after the test module is used to test.
Optionally, the system also includes：Preprocessing module；
The preprocessing module pre-processes for the track data to each vehicle in the first preset time period；
The extraction module, is specifically used for：Each vehicle is extracted in the pretreated track data of the preprocessing module
Characteristic information.
Optionally, the track data includes：A series of position data and travel speed；
The extraction module, is specifically used for：
First preset time period is divided into M the second preset time periods, and according to time order and function order by preceding n
Second preset time period is as the 3rd preset time period, wherein n<M；
According to the position data and travel speed contained in the track data of each vehicle, it is default each second to count each vehicle
Operation number of days, maintenance station in period stop number and stop duration；
When stopping number according to operation number of days of each vehicle in each second preset time period, maintenance station and stop
It is long, determine that operation number of days variation tendency of each vehicle in the 3rd preset time period, maintenance station are stopped number and averagely stopped
By duration.
According to operation number of days of each vehicle in each second preset time period and default stoppage in transit standard, each vehicle is determined
Stoppage in transit situation in each second preset time period.
Optionally, the generation module, specifically includes：Determination sub-module, sample generation submodule and selection submodule；
The determination sub-module, for according to each vehicle that the extraction module extracts in each second preset time period
Stoppage in transit situation and default scrap-car standard determine scrap-car and operation vehicle in each vehicle；
The sample generates submodule, for each vehicle in the scrap-car that determines the determination sub-module described the
Operation number of days variation tendency, maintenance station in three preset time periods stop number and averagely stop duration and established with corresponding vehicle
Correspondence is simultaneously used as positive sample, obtains positive sample set；Each vehicle in operation vehicle that the determination sub-module determines is existed
Operation number of days variation tendency, maintenance station in 3rd preset time period stop number and averagely stop duration and corresponding vehicle
It establishes correspondence and as negative sample, obtains negative sample set；
The selection submodule, for the positive sample collection obtained according to the first preset ratio in sample generation submodule
Positive sample is chosen in conjunction, and negative sample is chosen in the negative sample set obtained in sample generation submodule, obtains present count
The training sample of amount；
The division module, is specifically used for：The default quantity for being obtained the selection submodule according to the second preset ratio
Training sample be divided into training set and test set.
Optionally, the training module, specifically includes：Training submodule and assessment submodule；
The trained submodule, the training set for being divided according to random forests algorithm to the division module are trained
Obtain prediction model；
The assessment submodule, for the prediction mould obtained according to Receiver operating curve to the trained submodule
Type is assessed and adjusts model parameter, when the first area under Receiver operating curve meets predetermined threshold value, output
Corresponding prediction model；
The test module, is specifically used for：The prediction model of the assessment submodule output is calculated in the division module
Second area on the test set of division under corresponding Receiver operating curve, when the second area meets described preset
During threshold value, the prediction module is triggered；
The prediction module, is specifically used for：Using current time as deadline, the 3rd preset time period is counted forward
The operation number of days variation tendency of interior each vehicle, maintenance station stop number and averagely stop duration, and are surveyed using the test module
Prediction model after examination predicts following scrap-car and scraps probability.
The advantage of the invention is that：
In the present invention, based on the historical trajectory data in each vehicle regular period, the generation of prediction model is carried out,
The length and width of data has and significantly increases so that the prediction model and prediction result finally obtained it is all more scientific and
Reliability；It is combined simultaneously with machine learning, operation rule of the vehicle before scrapping is found out according to random forests algorithm, so as to
To prediction scrap-car and the prediction model of probability is scrapped, solves hysteresis sex chromosome mosaicism of the prior art；For each vehicle
For manufacturer, marketing strategy also can be more accurately determined.
Description of the drawings
By reading the detailed description of hereafter preferred embodiment, it is various other the advantages of and benefit it is common for this field
Technical staff will be apparent understanding.Attached drawing is only used for showing the purpose of preferred embodiment, and is not considered as to the present invention
Limitation.And throughout the drawings, the same reference numbers will be used to refer to the same parts.In the accompanying drawings：
Attached drawing 1 is a kind of scrap-car Forecasting Methodology flow chart provided by the invention；
Attached drawing 2 is a kind of scrap-car forecasting system module composition frame chart provided by the invention.
Specific embodiment
The illustrative embodiments of the disclosure are more fully described below with reference to accompanying drawings.Although this public affairs is shown in attached drawing
The illustrative embodiments opened, it being understood, however, that may be realized in various forms the disclosure without the reality that should be illustrated here
The mode of applying is limited.It is to be able to be best understood from the disclosure on the contrary, providing these embodiments, and can be by this public affairs
The scope opened completely is communicated to those skilled in the art.
Embodiment one
According to the embodiment of the present invention, a kind of Forecasting Methodology of scrap-car is provided, as shown in Figure 1, including：
Step 101：The characteristic information of each vehicle is extracted in the track data of each vehicle in the first preset time period；
Preferably, in the present invention, each vehicle is equipped with mobile unit, and mobile unit is every prefixed time interval (such as 30
Second) report the track data of corresponding vehicle；Wherein, track data includes but not limited to：A series of position data and traveling speed
Degree；
In the present embodiment, before step 101, further include：To the track data of each vehicle in the first preset time period into
Row pretreatment；
Specifically, the track data of each vehicle in the first preset time period is proceeded as follows：
Wrong data in filter footprint data；
Correct the position data deviated in track data；
The benefit corrected according to time dimension in track data passes data；
Wherein, the wrong data in filter footprint data, such as location information mistake, travel speed mistake；Correct track
The position data of offset is modified by the position data deviated in data specifically by the methods of Fourier filtering；
Accordingly, step 101, it is specially：The characteristic information of each vehicle is extracted in track data after the pre-treatment.
Further, in the present embodiment, step 101 specifically includes：
First preset time period is divided into M the second preset time periods, and according to time order and function order by preceding n second
Preset time period is as the 3rd preset time period, wherein n<M；
According to the position data and travel speed contained in the track data of each vehicle, it is default each second to count each vehicle
Operation number of days, maintenance station in period stop number and stop duration；
Number is stopped according to operation number of days of each vehicle in each second preset time period, maintenance station and stops duration, really
Operation number of days variation tendency of the fixed each vehicle in the 3rd preset time period, maintenance station stop number and averagely stop duration.
According to operation number of days of each vehicle in each second preset time period and default stoppage in transit standard, determine each vehicle each
Stoppage in transit situation in second preset time period.
Wherein, the first preset time period, the second preset time period, the 3rd preset time end can voluntarily be set according to demand
It is fixed, for example, in the present embodiment, using the first preset time period as in January, 2016 in June, 2017, the second preset time period
First preset time period is then divided into 12 by Shi Changwei mono- month, the 3rd preset time period to illustrate exemplified by first 12 months
A month, the 3rd preset time period was in January, 2016 in December, 2016.
Wherein, according to the position data and travel speed contained in the track data of each vehicle, each vehicle is counted each
Operation number of days, maintenance station in two preset time periods stop number and stop duration, are specially：
According to the position data, travel speed and corresponding time contained in the track data of each vehicle, arbitrary phase is determined
The distance between adjacent two position datas calculate each vehicle in daily distance travelled, when each vehicle exists according to definite distance
Daily distance travelled is more than preset travel fare register, using corresponding one day service time as vehicle, counts each vehicle and exists
Operation number of days in each second preset time period；Wherein, preset travel mileage is preferably 5 kms；
The position data contained in the track data of each vehicle is matched with the position data of each maintenance station, and general
At the corresponding maintenance station of successful position data, stop the conduct that duration is more than preset duration and once really stop, statistics is each
It stops number and stops duration in maintenance station of the vehicle in each second preset time period；Wherein, preset duration, when being preferably 2 small.
Wherein, when stopping number according to operation number of days of each vehicle in each second preset time period, maintenance station and stop
It is long, when determining that operation number of days variation tendency of each vehicle in the 3rd preset time period, maintenance station are stopped number and averagely stopped
It is long, be specially：Using least square method to each vehicle in each second preset time period for being included in the 3rd preset time period
Operation number of days is calculated, and obtains operation number of days variation tendency of each vehicle in the 3rd preset time period；It is and default to the 3rd
Each vehicle maintenance station in each second preset time period included in period is stopped number and is summed it up, and it is pre- the 3rd to obtain each vehicle
If number is stopped in the maintenance station in the period；And each vehicle in each second preset time period to being included in the 3rd preset time period
Duration averaging is stopped in maintenance station, obtains maintenance station of each vehicle in the 3rd preset time period and averagely stops duration；
Wherein, stoppage in transit standard is preset, it is preferable that be specially in the present embodiment：When vehicle is in the second preset time period
When operation number of days is zero, it is denoted as vehicle and stops transport in second preset time period；Accordingly, it is default each second according to each vehicle
Operation number of days and default stoppage in transit standard in period determine stoppage in transit situation of each vehicle in each second preset time period, tool
Body is：Judge whether operation number of days of each vehicle in each second preset time period is zero, and be zero corresponding by operation number of days
Second preset time period is denoted as corresponding vehicle and stops transport.
Step 102：According to the characteristic information of each vehicle and default scrap-car standard generation training sample, and by generation
Training sample is divided into training set and test set；
Preferably, in the present embodiment, presetting scrap-car standard is：The last one contained in 3rd preset time period
Second preset time period is not stopped transport, and the second preset time that is adjacent, continuous, presetting quantity after the 3rd preset time period
It stops transport in section；Wherein, default quantity is preferably 6；
For example, in the present embodiment, in December, 2016 is not stopped transport, and continuous 6 months after in December, 2016
The vehicle that (i.e. in January, 2017 to June) stops transport is as scrap-car.
In the present embodiment, step 102 specifically includes：
Step 102-1：According to stoppage in transit situation of each vehicle in each second preset time period and default scrap-car standard,
Determine the scrap-car and operation vehicle in each vehicle；
Step 102-2：By operation number of days variation tendency of each vehicle in scrap-car in the 3rd preset time period, repair
It stands to stop number and averagely stop duration and establishes correspondence and as positive sample with corresponding vehicle, obtain positive sample set；
Operation number of days variation tendency of each vehicle in the 3rd preset time period in vehicle will be run, number is stopped in maintenance station and averagely stops
Correspondence is established with corresponding vehicle and as negative sample, obtain negative sample set by duration；
Step 102-3：Positive sample is chosen in positive sample set according to the first preset ratio, and is selected in negative sample set
Negative sample is taken, obtains the training sample of default quantity；
Preferably, in the present embodiment, the first preset ratio is 1：1, it is 10000 to preset quantity.
Step 102-4：The training sample of default quantity is divided by training set and test set according to the second preset ratio.
Preferably, in the present embodiment, the second preset ratio is 7：3.
It may be noted that ground, scrap-car standard is preset, it can according to demand and actual conditions sets itself.
Step 103：Prediction model is trained according to the training set, and using the test set to the prediction model into
Row test；
In the present embodiment, step 103, specifically include：
Step 103-1：Training set is trained according to random forests algorithm to obtain prediction model；
Specifically, using Python instruments, the random forests algorithm in scikit-learn machine learning storehouse is called to instruction
Practice collection to be trained to obtain prediction model.
Step 103-2：According to Receiver operating curve (Receiver OperatingCharacteristic
Curve, abbreviation ROC curve) prediction model is assessed and adjusts model parameter, when under Receiver operating curve
When one area meets predetermined threshold value, corresponding prediction model is exported；
Wherein, the area under Receiver operating curve, also known as AUC, specifically, when the first AUC meets default threshold
During value, corresponding prediction model is exported；
Wherein, predetermined threshold value is preferably 0.8.
Step 103-3：Calculate the of the prediction model of output on test set under corresponding Receiver operating curve
Two areas when second area meets the predetermined threshold value, perform step 104；Otherwise return to step 103-1；
In the present invention, prediction model is trained using random forests algorithm, it is made to have more interpretation, and is solved existing
Sex chromosome mosaicism is lagged in technology；Prediction model is tested by test set simultaneously, has further ensured the standard of prediction model
True property.
Step 104：Use following scrap-car of the prediction model prediction after test.
Specifically, using current time as deadline, the 3rd preset time period Nei Geche is counted forward according to preceding method
Operation number of days variation tendency, maintenance station stopped number and averagely stops duration, and predicted using the prediction model after test
Following scrap-car and scrap probability.
Embodiment two
According to the embodiment of the present invention, a kind of forecasting system of scrap-car is provided, as shown in Fig. 2, including：
Extraction module 201, for extracting the feature of each vehicle in the track data of each vehicle in the first preset time period
Information；
Generation module 202, for the characteristic information of each vehicle extracted according to extraction module 201 and default scrap-car mark
Quasi- generation training sample；
Division module 203, the training sample for generation module 202 to be generated are divided into training set and test set；
Training module 204, the training set for being divided according to division module 203 train prediction model；
Test module 205, for using division module 203 divide test set to the prediction model that training module is trained into
Row test；
Prediction module 206, for following scrap-car of the prediction model prediction after test module 205 is used to test.
According to the embodiment of the present invention, which further includes：Preprocessing module；
Preprocessing module pre-processes for the track data to each vehicle in the first preset time period；
Accordingly, extraction module 201 are specifically used for：Each vehicle is extracted in the pretreated track data of preprocessing module
Characteristic information.
According to the embodiment of the present invention, track data includes：A series of position data and travel speed；
Accordingly, extraction module 201 are specifically used for：
First preset time period is divided into M the second preset time periods, and according to time order and function order by preceding n
Second preset time period is as the 3rd preset time period, wherein n<M；
According to the position data and travel speed contained in the track data of each vehicle, it is default each second to count each vehicle
Operation number of days, maintenance station in period stop number and stop duration；
When stopping number according to operation number of days of each vehicle in each second preset time period, maintenance station and stop
It is long, determine that operation number of days variation tendency of each vehicle in the 3rd preset time period, maintenance station are stopped number and averagely stopped
By duration.
According to operation number of days of each vehicle in each second preset time period and default stoppage in transit standard, each vehicle is determined
Stoppage in transit situation in each second preset time period.
In the present embodiment, the first preset time period, the second preset time period, the 3rd preset time end can be according to need
Sets itself is sought, for example, in the present embodiment, being preset by January, 2016 in June, 2017, second of the first preset time period
Period when a length of one month, the 3rd preset time period to illustrate exemplified by first 12 months, then by the first preset time period
It is divided into 12 months, the 3rd preset time period is in January, 2016 in December, 2016.
According to the embodiment of the present invention, generation module 202 specifically include：Determination sub-module, sample generation submodule and
Submodule is chosen, wherein：
Determination sub-module, for the stoppage in transit according to each vehicle that extraction module 201 extracts in each second preset time period
Situation and default scrap-car standard determine scrap-car and operation vehicle in each vehicle；
Wherein, presetting scrap-car standard is preferably：Contain in 3rd preset time period the last one second it is default when
Between section do not stop transport, and stop transport after the 3rd preset time period in the second preset time period of adjacent, continuous, default quantity；
Wherein, default quantity is preferably 6；
For example, in the present embodiment, in December, 2016 is not stopped transport, and continuous 6 months after in December, 2016
The vehicle that (i.e. in January, 2017 to June) stops transport is as scrap-car.
Sample generates submodule, for each vehicle in the scrap-car that determines determination sub-module in the 3rd preset time period
In operation number of days variation tendency, maintenance station stop number and averagely stop duration establish correspondence with corresponding vehicle and make
For positive sample, positive sample set is obtained；Each vehicle is in the 3rd preset time period in the operation vehicle that determination sub-module is determined
Operation number of days variation tendency, maintenance station stop number and averagely stop duration establish correspondence and conduct with corresponding vehicle
Negative sample obtains negative sample set；
Submodule is chosen, for being chosen according to the first preset ratio in the obtained positive sample set of sample generation submodule
Positive sample, and negative sample is chosen in the negative sample set obtained in sample generation submodule, obtain the training sample of default quantity；
Accordingly, division module 203 are specifically used for：The present count for being obtained selection submodule according to the second preset ratio
The training sample of amount is divided into training set and test set.
Wherein, the first preset ratio is preferably 1：1；Second preset ratio is preferably 7：3.
According to the embodiment of the present invention, training module 204 specifically include：Training submodule and assessment submodule,
In：
Training submodule, the training set for being divided according to random forests algorithm to division module 203 are trained to obtain
Prediction model；
In the present embodiment, training submodule, is specifically used for：Using Python instruments, scikit-learn machines are called
The training set that random forests algorithm in learning database divides division module 203 is trained to obtain prediction model.
Submodule is assessed, the prediction model for being obtained according to Receiver operating curve to training submodule is commented
Estimate and adjust model parameter, when the first area under Receiver operating curve meets predetermined threshold value, output is corresponding pre-
Survey model；Wherein, predetermined threshold value is preferably 0.8；
Test module 205, is specifically used for：Calculate what the prediction model that assessment submodule exports was divided in division module 203
Second area on test set under corresponding Receiver operating curve, when second area meets predetermined threshold value, triggering is pre-
Survey module 206；
Accordingly, prediction module 206 are specifically used for：Using current time as deadline, the 3rd preset time is counted forward
The operation number of days variation tendency of each vehicle, maintenance station are stopped number and averagely stop duration, and surveyed using test module 205 in section
Prediction model after examination predicts following scrap-car and scraps probability.
In the present invention, based on the historical trajectory data in each vehicle regular period, the generation of prediction model is carried out,
The length and width of data has and significantly increases so that the prediction model and prediction result finally obtained it is all more scientific and
Reliability；It is combined simultaneously with machine learning, operation rule of the vehicle before scrapping is found out according to random forests algorithm, so as to
To prediction scrap-car and the prediction model of probability is scrapped, solves hysteresis sex chromosome mosaicism of the prior art；For each vehicle
For manufacturer, marketing strategy can be more accurately determined.
The foregoing is only a preferred embodiment of the present invention, but protection scope of the present invention be not limited thereto,
Any one skilled in the art in the technical scope disclosed by the present invention, the change or replacement that can be readily occurred in,
It should be covered by the protection scope of the present invention.Therefore, protection scope of the present invention should be with the protection model of the claim
Subject to enclosing.
Claims (10)
1. a kind of Forecasting Methodology of scrap-car, which is characterized in that including：
Step S1：The characteristic information of each vehicle is extracted in the track data of each vehicle in the first preset time period；
Step S2：According to the characteristic information of each vehicle and default scrap-car standard generation training sample, and by the instruction
Practice sample and be divided into training set and test set；
Step S3：Prediction model is trained according to the training set, and the prediction model is surveyed using the test set
Examination；
Step S4：Use following scrap-car of the prediction model prediction after test.
2. according to the method described in claim 1, it is characterized in that, before the step S1, further include：To the first preset time
The track data of each vehicle is pre-processed in section；
The step S1, specially：The characteristic information of each vehicle is extracted in track data after the pre-treatment.
3. according to the method described in claim 1, it is characterized in that, the track data includes：A series of position data and
Travel speed；
The step S1, specifically includes：
First preset time period is divided into M the second preset time periods, and according to time order and function order by preceding n second
Preset time period is as the 3rd preset time period, wherein n<M；
According to the position data and travel speed contained in the track data of each vehicle, each vehicle is counted in each second preset time
Operation number of days, maintenance station in section stop number and stop duration；
Number is stopped according to operation number of days of each vehicle in each second preset time period, maintenance station and stops duration, really
When operation number of days variation tendency of the fixed each vehicle in the 3rd preset time period, maintenance station are stopped number and are averagely stopped
It is long.
According to operation number of days of each vehicle in each second preset time period and default stoppage in transit standard, determine each vehicle each
Stoppage in transit situation in second preset time period.
4. according to the method described in claim 3, it is characterized in that, the step S2, specifically includes：
Step S2-1：According to stoppage in transit situation of each vehicle in each second preset time period and default scrap-car standard,
Determine the scrap-car and operation vehicle in each vehicle；
Step S2-2：By operation number of days variation tendency of each vehicle in scrap-car in the 3rd preset time period, repair
It stands to stop number and averagely stop duration and establishes correspondence and as positive sample with corresponding vehicle, obtain positive sample set；
Operation number of days variation tendency of each vehicle in the 3rd preset time period in vehicle will be run, number is stopped in maintenance station and flat
It stops duration and establishes correspondence and as negative sample with corresponding vehicle, obtain negative sample set；
Step S2-3：Positive sample is chosen in the positive sample set according to the first preset ratio, and in the negative sample set
Middle selection negative sample obtains the training sample of default quantity；
Step S2-4：The training sample of the default quantity is divided by training set and test set according to the second preset ratio.
5. according to the method described in claim 4, it is characterized in that, the step S3, specifically includes：
Step S3-1：The training set is trained to obtain prediction model according to random forests algorithm；
Step S3-2：The prediction model is assessed according to Receiver operating curve and adjusts model parameter, when by
When the first area under examination person's performance curve meets predetermined threshold value, corresponding prediction model is exported；
Step S3-3：Calculate second of the prediction model exported on the test set under corresponding Receiver operating curve
Area when the second area meets the predetermined threshold value, performs step S4；Otherwise return to step S3-1；
The step S4, specially：Using current time as deadline, each vehicle in the 3rd preset time period is counted forward
Operation number of days variation tendency, maintenance station stop number and averagely stop duration, and using the prediction model prediction after test not
Come scrap-car and scrap probability.
6. a kind of forecasting system of scrap-car, which is characterized in that including：
Extraction module, for extracting the characteristic information of each vehicle in the track data of each vehicle in the first preset time period；
Generation module, for the characteristic information of each vehicle extracted according to the extraction module and the generation of default scrap-car standard
Training sample；
Division module, the training sample for the generation module to be generated are divided into training set and test set；
Training module, the training set for being divided according to the division module train prediction model；
Test module, for being carried out using the test set of division module division to the prediction model that the training module is trained
Test；
Prediction module, for following scrap-car of the prediction model prediction after the test module is used to test.
7. system according to claim 1, which is characterized in that further include：Preprocessing module；
The preprocessing module pre-processes for the track data to each vehicle in the first preset time period；
The extraction module, is specifically used for：The spy of each vehicle is extracted in the pretreated track data of the preprocessing module
Reference ceases.
8. system according to claim 7, which is characterized in that the track data includes：A series of position data and
Travel speed；
The extraction module, is specifically used for：
First preset time period is divided into M the second preset time periods, and according to time order and function order by preceding n second
Preset time period is as the 3rd preset time period, wherein n<M；
According to the position data and travel speed contained in the track data of each vehicle, each vehicle is counted in each second preset time
Operation number of days, maintenance station in section stop number and stop duration；
Number is stopped according to operation number of days of each vehicle in each second preset time period, maintenance station and stops duration, really
When operation number of days variation tendency of the fixed each vehicle in the 3rd preset time period, maintenance station are stopped number and are averagely stopped
It is long.
According to operation number of days of each vehicle in each second preset time period and default stoppage in transit standard, determine each vehicle each
Stoppage in transit situation in second preset time period.
9. system according to claim 8, which is characterized in that the generation module specifically includes：Determination sub-module, sample
This generation submodule and selection submodule；
The determination sub-module, for the stoppage in transit according to each vehicle that the extraction module extracts in each second preset time period
Situation and default scrap-car standard determine scrap-car and operation vehicle in each vehicle；
The sample generates submodule, pre- the described 3rd for each vehicle in the scrap-car that determines the determination sub-module
If the operation number of days variation tendency, maintenance station in the period stop number and averagely stop duration is corresponding with the foundation of corresponding vehicle
Relation is simultaneously used as positive sample, obtains positive sample set；Each vehicle is described in the operation vehicle that the determination sub-module is determined
Operation number of days variation tendency, maintenance station in 3rd preset time period stop number and averagely stop duration and built with corresponding vehicle
Vertical correspondence is simultaneously used as negative sample, obtains negative sample set；
The selection submodule, for according to the first preset ratio in the obtained positive sample set of sample generation submodule
Positive sample is chosen, and negative sample is chosen in the negative sample set obtained in sample generation submodule, obtains default quantity
Training sample；
The division module, is specifically used for：According to the instruction for the default quantity that the second preset ratio obtains the selection submodule
Practice sample and be divided into training set and test set.
10. system according to claim 9, which is characterized in that the training module specifically includes：Training submodule and
Assess submodule；
The trained submodule, the training set for being divided according to random forests algorithm to the division module are trained to obtain
Prediction model；
The assessment submodule, for according to Receiver operating curve to the prediction model that the trained submodule obtains into
Row is assessed and adjusts model parameter, and when the first area under Receiver operating curve meets predetermined threshold value, output corresponds to
Prediction model；
The test module, is specifically used for：The prediction model for calculating the assessment submodule output is divided in the division module
Test set on second area under corresponding Receiver operating curve, when the second area meets the predetermined threshold value
When, trigger the prediction module；
The prediction module, is specifically used for：Using current time as deadline, count each in the 3rd preset time period forward
The operation number of days variation tendency of vehicle, maintenance station stop number and averagely stop duration, and after use test module test
The following scrap-car of prediction model prediction and scrap probability.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711445897.9A CN108053075B (en) | 2017-12-27 | 2017-12-27 | Scrapped vehicle prediction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711445897.9A CN108053075B (en) | 2017-12-27 | 2017-12-27 | Scrapped vehicle prediction method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108053075A true CN108053075A (en) | 2018-05-18 |
| CN108053075B CN108053075B (en) | 2021-03-26 |
Family
ID=62128185
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711445897.9A Active CN108053075B (en) | 2017-12-27 | 2017-12-27 | Scrapped vehicle prediction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108053075B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108810526A (en) * | 2018-06-12 | 2018-11-13 | 广东惠禾科技发展有限公司 | Camera intelligent fault prediction technique, device and camera management server |
| CN109102149A (en) * | 2018-06-06 | 2018-12-28 | 北京市燃气集团有限责任公司 | A kind of prediction technique of city gas buried pipeline third party breakage in installation risk |
| CN109345136A (en) * | 2018-10-22 | 2019-02-15 | 中山大学 | Vehicle survival curve model optimization method based on China's motor vehicle resignation system |
| CN109635965A (en) * | 2018-12-24 | 2019-04-16 | 成都四方伟业软件股份有限公司 | Bus scraps decision-making technique, device and readable storage medium storing program for executing |
| CN111144485A (en) * | 2019-12-26 | 2020-05-12 | 北京中交兴路车联网科技有限公司 | Vehicle accident judgment method and system based on xgboost classification algorithm |
| CN111860954A (en) * | 2020-06-18 | 2020-10-30 | 上海钧正网络科技有限公司 | Vehicle loss of contact prediction method and device, computer equipment and storage medium |
| CN111861073A (en) * | 2019-04-30 | 2020-10-30 | 北京嘀嘀无限科技发展有限公司 | Operation strategy prediction method and device, electronic equipment and readable storage medium |
| CN112529104A (en) * | 2020-12-23 | 2021-03-19 | 东软睿驰汽车技术（沈阳）有限公司 | Vehicle fault prediction model generation method, fault prediction method and device |
| CN113570469A (en) * | 2021-07-27 | 2021-10-29 | 上海交通大学 | Intelligent vehicle change prediction method for vehicle insurance users |
| CN114912675A (en) * | 2022-05-06 | 2022-08-16 | 杉数科技(北京)有限公司 | Prediction method and device for scrappage of aero-engine blade |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105389974A (en) * | 2015-11-19 | 2016-03-09 | 深圳市赛格导航科技股份有限公司 | Vehicle tracking method and system based on vehicle historical driving data |
| CN105868865A (en) * | 2016-04-13 | 2016-08-17 | 华北电力大学 | Electric vehicle parc prediction method based on multivariate linear regression method and proportional substitution method |
| CN106127312A (en) * | 2016-06-21 | 2016-11-16 | 中车青岛四方机车车辆股份有限公司 | The life cycle management history information management method of EMUs and system |
| CN107025345A (en) * | 2017-03-31 | 2017-08-08 | 柳州顺联科技有限公司 | A kind of Forecasting Methodology of engineering machinery vehicle failure time |
-
2017
- 2017-12-27 CN CN201711445897.9A patent/CN108053075B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105389974A (en) * | 2015-11-19 | 2016-03-09 | 深圳市赛格导航科技股份有限公司 | Vehicle tracking method and system based on vehicle historical driving data |
| CN105868865A (en) * | 2016-04-13 | 2016-08-17 | 华北电力大学 | Electric vehicle parc prediction method based on multivariate linear regression method and proportional substitution method |
| CN106127312A (en) * | 2016-06-21 | 2016-11-16 | 中车青岛四方机车车辆股份有限公司 | The life cycle management history information management method of EMUs and system |
| CN107025345A (en) * | 2017-03-31 | 2017-08-08 | 柳州顺联科技有限公司 | A kind of Forecasting Methodology of engineering machinery vehicle failure time |
Non-Patent Citations (1)
| Title |
|---|
| 赵鹏等: "基于灰色***理论的汽车报废量预测", 《工业技术经济》 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109102149A (en) * | 2018-06-06 | 2018-12-28 | 北京市燃气集团有限责任公司 | A kind of prediction technique of city gas buried pipeline third party breakage in installation risk |
| CN108810526A (en) * | 2018-06-12 | 2018-11-13 | 广东惠禾科技发展有限公司 | Camera intelligent fault prediction technique, device and camera management server |
| CN109345136A (en) * | 2018-10-22 | 2019-02-15 | 中山大学 | Vehicle survival curve model optimization method based on China's motor vehicle resignation system |
| WO2020082683A1 (en) * | 2018-10-22 | 2020-04-30 | 中山大学 | Chinese motor vehicle registration system-based method for optimizing vehicle survival curve model |
| CN109635965A (en) * | 2018-12-24 | 2019-04-16 | 成都四方伟业软件股份有限公司 | Bus scraps decision-making technique, device and readable storage medium storing program for executing |
| CN111861073B (en) * | 2019-04-30 | 2024-05-03 | 北京嘀嘀无限科技发展有限公司 | Operation policy prediction method and device, electronic equipment and readable storage medium |
| CN111861073A (en) * | 2019-04-30 | 2020-10-30 | 北京嘀嘀无限科技发展有限公司 | Operation strategy prediction method and device, electronic equipment and readable storage medium |
| CN111144485A (en) * | 2019-12-26 | 2020-05-12 | 北京中交兴路车联网科技有限公司 | Vehicle accident judgment method and system based on xgboost classification algorithm |
| CN111860954A (en) * | 2020-06-18 | 2020-10-30 | 上海钧正网络科技有限公司 | Vehicle loss of contact prediction method and device, computer equipment and storage medium |
| CN111860954B (en) * | 2020-06-18 | 2024-07-05 | 上海钧正网络科技有限公司 | Vehicle loss-of-link prediction method, device, computer equipment and storage medium |
| CN112529104A (en) * | 2020-12-23 | 2021-03-19 | 东软睿驰汽车技术（沈阳）有限公司 | Vehicle fault prediction model generation method, fault prediction method and device |
| CN113570469A (en) * | 2021-07-27 | 2021-10-29 | 上海交通大学 | Intelligent vehicle change prediction method for vehicle insurance users |
| CN113570469B (en) * | 2021-07-27 | 2024-05-28 | 上海交通大学 | Intelligent vehicle change prediction method for vehicle insurance user |
| CN114912675A (en) * | 2022-05-06 | 2022-08-16 | 杉数科技(北京)有限公司 | Prediction method and device for scrappage of aero-engine blade |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108053075B (en) | 2021-03-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108053075A (en) | A kind of scrap-car Forecasting Methodology and system | |
| CN104899405B (en) | Data predication method and system and alarm method and system | |
| CN104134349B (en) | A kind of public transport road conditions disposal system based on traffic multisource data fusion and method | |
| CN101504717B (en) | Characteristic area positioning method, car body color depth and color recognition method | |
| CN104933860B (en) | Bus traffic congestion delay time at stop Forecasting Methodology based on gps data | |
| CN102881162B (en) | Data processing and fusion method for large-scale traffic information | |
| CN108345666A (en) | A kind of vehicle abnormality track-detecting method based on time-space isolated point | |
| CN102646332A (en) | Traffic state estimation device and method based on data fusion | |
| US20220097557A1 (en) | Information provision service system for electric vehicle using vehicle sensors | |
| CN109243173A (en) | Track of vehicle analysis method and system based on road high definition bayonet data | |
| CN106845768A (en) | Bus hourage model building method based on survival analysis parameter distribution | |
| CN107563566A (en) | A kind of run time interval prediction method between bus station based on SVMs | |
| CN109887292B (en) | Vehicle type identification method and system | |
| CN103606272B (en) | A kind of bus rapid transit arrival time Forecasting Methodology based on the volume of the flow of passengers | |
| CN106649709A (en) | Vehicle stopping behavior mode prediction and assessment method based on data mining | |
| CN108710637A (en) | Taxi exception track real-time detection method based on time-space relationship | |
| CN115691120A (en) | Congestion identification method and system based on highway running water data | |
| CN106157609A (en) | The monitoring method of a kind of block status and device | |
| CN101794510A (en) | Data processing method and device of floating cars | |
| CN109493449A (en) | A kind of lorry loading method for estimating state based on lorry GPS track data and high speed transaction data | |
| CN108197977A (en) | A kind of recommendation method and device of vehicle brand | |
| CN106127407A (en) | Aircraft stroke scoring method based on multi-sensor information fusion and scoring system | |
| CN107993441A (en) | A kind of lorry often runs away the Forecasting Methodology and device of line | |
| Li et al. | Diagnosing and interpolating loop detector data errors with probe vehicle data | |
| CN109727474A (en) | A kind of bus based on fused data precise recognition method out of the station |
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 |