CN109376908A - Prediction technique, system, storage medium and the equipment of logistics node health status - Google Patents
Prediction technique, system, storage medium and the equipment of logistics node health status Download PDFInfo
- Publication number
- CN109376908A CN109376908A CN201811123053.7A CN201811123053A CN109376908A CN 109376908 A CN109376908 A CN 109376908A CN 201811123053 A CN201811123053 A CN 201811123053A CN 109376908 A CN109376908 A CN 109376908A
- Authority
- CN
- China
- Prior art keywords
- logistics node
- profit
- health status
- addressee
- income
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- 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
- 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
- G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Quality & Reliability (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Game Theory and Decision Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The prediction technique of present invention offer logistics node health status, comprising: the potential factor of analyzing influence logistics node health status, to obtain the key factor for influencing logistics node health status;Wherein, the key factor includes: to send part income and retardation rate;Part income and retardation rate are sent in analyte stream site, to obtain the key feature of the logistics node of logout;The key feature includes: the gross profit of logistics node;A number of factors relevant to the gross profit of logistics node is analyzed, draws the revenue and expenditure system dynamics model of logistics node accordingly;Based on the revenue and expenditure system dynamics model, the regression analysis model for predicting logistics node logout possibility, and the prediction result by the solution of the regression analysis model as logistics node health status are established.The present invention facilitates the health status of look-ahead effluent stream site, consequently facilitating policymaker has found to avoid economic loss there are the logistics node of logout risk in time.
Description
Technical field
The present invention relates to logistics fields, are situated between more particularly to the prediction technique of logistics node health status, system, storage
Matter and equipment.
Background technique
With the continuous improvement of people's living standards, e-commerce industry and logistic industry are also grown rapidly.So
And express delivery site logout in 2017, the events such as go on strike, run away take place frequently.For express delivery network, one is joined the logout of site
Pressure can be increased to other site nearby, the serious express delivery delivery service that can also upset a region influences express company
Reputation.
In the circumstances for facing site logout risk, express company wants to identify the feature i.e. by logout site,
To carry out redemption in advance or to shunt scheme.But from service layer, the analysis of site health status inevitably has one
Fixed hysteresis quality.If judging how many days certain site does not sign for express mail with index is signed for addressee, or pass through reality
Ground visits to find whether the site has closed, then has had resulted in irremediable loss.
Summary of the invention
In view of the foregoing deficiencies of prior art, the purpose of the present invention is to provide the predictions of logistics node health status
Method, system, storage medium and equipment are difficult to the strong of look-ahead logistics node for solving loglstics enterprise in the prior art
The problem of health situation.
In order to achieve the above objects and other related objects, the present invention provides a kind of prediction side of logistics node health status
Method, comprising: the potential factor of analyzing influence logistics node health status, with obtain influence logistics node health status it is crucial because
Element;Wherein, the key factor includes: to send part income and retardation rate;Part income and retardation rate are sent in analyte stream site, to obtain
Obtain the key feature of the logistics node of logout;The key feature includes: the gross profit of logistics node;Analysis and logistics node
The relevant a number of factors of gross profit draws the revenue and expenditure system dynamics model of logistics node accordingly;It is dynamic based on the revenue and expenditure system
Mechanical model, establishes the regression analysis model for predicting logistics node logout possibility, and by the regression analysis model
Solve the prediction result as logistics node health status.
In one embodiment of the invention, the regression analysis model specifically:
Wherein, profit, billing.quantity, true.sign respectively indicate the gross profit of the logistics node, receive
Part amount and part amount is sent, p indicates the logout probability of the logistics node, β0、β1、β2Respectively constant, ε are random error.
In one embodiment of the invention, the mathematical model of the gross profit profit are as follows: gross profit profit=sends part sharp
Profit+addressee profit-delay loss;Wherein, described that part profit=send part is sent to take in IncomesPart is sent to pay costs;The addressee
Profit=addressee takes in IncomerAddressee pays Costr;The delay loss=E* postpones loss delay1-F* delay in one day
Lose delay2plus+H within loss delay2-G* delay in two days two days or more;Wherein, E, F, G, H are respectively constant.
It is described that part is sent to pay cost in one embodiment of the inventionsMathematical model are as follows: send part pay costs=A*
true.sign2+ B*true.sign, wherein A, B are respectively constant, and what true.sign indicated the logistics node sends part amount.
In one embodiment of the invention, the addressee takes in IncomerMathematical model are as follows: addressee take in Incomer=C*
billing.quantity2+ D*billing.quantity, wherein be respectively C, D, constant, billing.quantity is indicated
The addressee amount of the logistics node.
In order to achieve the above objects and other related objects, the present invention provides a kind of prediction system of logistics node health status
System, comprising: the first analysis module influences logistics net for the potential factor of analyzing influence logistics node health status to obtain
The key factor of point health status;Wherein, the key factor includes: to send part income and retardation rate;Second analysis module, is used for
Part income and retardation rate are sent in analyte stream site, to obtain the key feature of the logistics node of logout;The key feature packet
It includes: the gross profit of logistics node;Third analysis module, for analyzing a number of factors relevant to the gross profit of logistics node;It is imitative
True modeling module draws the revenue and expenditure system dynamics mould of logistics node for the analysis result based on the third analysis module
Type;Regression analysis module is established for being based on the revenue and expenditure system dynamics model for predicting logistics node logout possibility
Regression analysis model, and by prediction result of the solution as logistics node health status of the regression analysis model.
In one embodiment of the invention, the regression analysis model specifically:
Wherein, profit, billing.quantity, true.sign respectively indicate the gross profit of the logistics node, receive
Part amount and part amount is sent, p indicates the logout probability of the logistics node, β0、β1、β2Respectively constant, ε are random error.
In one embodiment of the invention, the mathematical model of the gross profit profit are as follows: gross profit profit=sends part sharp
Profit+addressee profit-delay loss;Wherein, described that part profit=send part is sent to take in IncomesPart is sent to pay costs;The addressee
Profit=addressee takes in IncomerAddressee pays Costr;The delay loss=E* postpones loss delay1-F* delay in one day
Lose delay2plus+H within loss delay2-G* delay in two days two days or more;Wherein, E, F, G, H are respectively constant.
It is described that part is sent to pay cost in one embodiment of the inventionsMathematical model are as follows: send part pay costs=A*
true.sign2+ B*true.sign, wherein A, B are respectively constant, and what true.sign indicated the logistics node sends part amount.
In one embodiment of the invention, the addressee takes in IncomerMathematical model are as follows: addressee take in Incomer=C*
billing.quantity2+ D*billing.quantity, wherein be respectively C, D, constant, billing.quantity is indicated
The addressee amount of the logistics node.
In order to achieve the above objects and other related objects, the present invention provides a kind of storage medium, wherein being stored with computer
Program when the computer program is by processor load and execution, realizes the pre- of any logistics node health status as above
Survey method.
In order to achieve the above objects and other related objects, the present invention provides a kind of electronic equipment, comprising: processor and deposits
Reservoir;Wherein, the memory is for storing computer program;The processor is used for computer program described in load and execution,
So that the electronic equipment executes the prediction technique of as above any logistics node health status.
As described above, prediction technique, system, storage medium and the equipment of logistics node health status of the invention, have
Below the utility model has the advantages that facilitating the health status of look-ahead effluent stream site, moved back consequently facilitating policymaker has found to exist in time
The logistics node of net risk, avoids economic loss.
Detailed description of the invention
Fig. 1 is shown as the flow diagram of the prediction technique of the logistics node health status in one embodiment of the invention.
Fig. 2A is shown as the dynamics flow diagram of the revenue and expenditure system of the single logistics node in one embodiment of the invention.
Fig. 2 B is shown as the regression analysis model schematic diagram in one embodiment of the invention.
Fig. 3 is shown as the module diagram of the forecasting system of the logistics node health status in one embodiment of the invention.
Fig. 4 is shown as the structural schematic diagram of the electronic equipment in one embodiment of the invention.
Specific embodiment
Illustrate embodiments of the present invention below by way of specific specific example, those skilled in the art can be by this specification
Other advantages and efficacy of the present invention can be easily understood for disclosed content.The present invention can also pass through in addition different specific realities
The mode of applying is embodied or practiced, the various details in this specification can also based on different viewpoints and application, without departing from
Various modifications or alterations are carried out under spirit of the invention.It should be noted that in the absence of conflict, following embodiment and implementation
Feature in example can be combined with each other.
It should be noted that illustrating the basic structure that only the invention is illustrated in a schematic way provided in following embodiment
Think, only shown in schema then with related component in the present invention rather than component count, shape and size when according to actual implementation
Draw, when actual implementation kenel, quantity and the ratio of each component can arbitrarily change for one kind, and its assembly layout kenel
It is likely more complexity.
The present invention is directed to predict the health status of logistics node, thus found in time there are the logistics node of logout risk,
There is provided reference for business decision, avoid enterprise by logistics node logout and caused by economic loss.
Refering to fig. 1, the present embodiment proposes a kind of prediction technique of logistics node health status.So-called logistics node is exactly
The node for referring to logistics network specifically includes that storage and transportation warehouse, circulation warehouse, transfer house etc..The prediction technique of the present embodiment is main
The following steps are included:
The potential factor of S11, analyzing influence logistics node health status, to obtain the pass for influencing logistics node health status
Key factor;Wherein, the key factor includes: to send part income and retardation rate.
In this present embodiment, using exploratory Principal Component Analysis to a certain logistics node health status latent factor of influence
Correlation analysis, to explore the major influence factors for influencing the logistics node health status.
1 data structure table of table
Table 1 summarizes the most attributes and index of logistics node, first in order to excavate the association in table 1 between data
First, dimensionality reduction is carried out to index using EFA method.EFA (Exploratory Factor Analysis, factor analysis exploratory)
The purpose of method is that will have the aggregation of variable of complex relationship for a few core factor, to achieve the purpose that dimensionality reduction.Assuming that
The value of one logistics node is made of N number of first class index, this N number of first class index is made of a series of two-level index again, level-one
Index can take 1 middle finger target major class of table, such as: index, income index are signed for, two-level index then includes in such as expenditure target
Weighing expenditure, transfer expenditure, send part pay etc..Then, using Principal Component Analysis, matrix rotation is carried out using varimax
Turn, specifically: varimax matrix rotation makes each factor loading (also make aij, i.e. i-th of variable by coordinate transform
With the related coefficient of j-th of common factor) sum of variance it is maximum, be exactly for popular: any one of (a) variable is at one
There is high contribution rate in the factor, and the load on other factors is almost 0;Any one of (b) factor only has on a small number of variables
Top load, and the load on other variables is almost 0.Table 2 shows that all points of indexs can explain 70% or more variance, drop
Dimension did not lose multi information.
Table 2 explains population variance table
It makes explanations below to table 2: judging the number of principal component (or factor), table 2 using Principal Component Analysis first
The first row of middle composition is the number of composition, and second is classified as the characteristic value of composition, and characteristic value is bigger, and the importance of the composition is got over
It is high.Table 2 shows the principal component that characteristic value is greater than 1.First unit in " initial characteristic values " indicates that the composition explains always
The 26.937% of variance, the first seven composition is accumulative to explain the variance of entire data set 70.792%.Factor number is being determined
Afterwards, second step we need to extract common factor, the common factor not rotated is generally extracted using main shaft iterative method.It " extracts flat
The result of side and loading " part is close with the data meaning of front.Third step usage factor rotates to enhance explanation degree, is compared
Relatively satisfactory main gene.Factor rotation seeks to make the square value of factor loading in Factor load-matrix to 0 and 1 both direction point
Change, keeps big load bigger, small load is smaller.Original variable population variance is still explained after the rotation of first four common factor
53.994%, information amount lost is few, but has redistributed variance contribution of each index in each principal component, so that index exists
Load on some common factor compares concentration, is convenient for better distribution factor.
From data analysis find out and initial data in site total expenditure (only with parent company interaction expenditure) with weighing branch
Out, transfer, which is paid, pays with part is sent in the same dimension;Send part income and real label amount degree of correlation highest;Factor correlation is sent in delay with charge free
Property is very high.
Following data conversion is carried out as a result:
1) it sends part expenditure as addressee total expenditure using expenditure of weighing+transfer is paid+, is denoted as Costr;
2) with weight
With the delay number of days n sum of products as delay outbox severity, it is denoted as delay_weighted (weighted delay), k
For the integer more than or equal to 3.The value range of weight w be 0 < w < 1 and and for 1, be one to the severity of various late days
Kind distribution, and k is bigger, the weight of delay 1 day and 2 days is smaller, and delay 2 days or more weights are bigger.Assuming that k=4, then postpone 1
Its weight w=1/ (1+2+3+4)=1/10 postpones 2 days weight w=2/ (1+2+3+4)=1/5, delay two days or more w=1-1/
10-1/5=7/10.
S12, analyte stream site send part take in and retardation rate, to obtain the key feature of the logistics node of logout;Institute
State the gross profit that key feature includes: logistics node.
In this present embodiment, part income and retardation rate is sent to be moved back to excavate by what clustering methodology analyzed the logistics node
Net the feature of site.
Hereinafter, income and retardation rate to estimate roughly excavate the feature of logistics logout site.
Under the premise of not calculating human cost and sending expense (assessment for not influencing enterprise's relative profit) with charge free, each logistics
It is average addressee income that the gross profit profit=of site, which sends part income+addressee amount * R, R,.Assuming that R=10, by delay_
Sample is divided into 2*2=4 class by weighted (weighted delay) field and profit field clustering, and the results are shown in Table 3.
3 K-means cluster result of table (being input with delay_weighted and profit field)
It should be added that KMeans clustering algorithm main thought is to defining K value and K initial classes cluster central point
In the case where, each point (data record) is assigned in class cluster representated by nearest class cluster central point, all the points distribution
After finishing, the central point (being averaged) of such cluster is recalculated according to all the points in a class cluster, then iteration again
The step of being allocated a little and updating class cluster central point, until class cluster central point varies less, or reaches specified iteration
Number.The determination of K value can be by selecting a suitable class cluster index, such as mean radius, in the K value such as model from 2 to 7
Enclose operation for several times kmeans to obtain the change curve of class cluster index.If the downward trend of class cluster index is most when K value 4
Fastly, then the correct value of K is 4.
Table 3 is that data are carried out with resulting result after kmeans cluster.Each point has been assigned in a class cluster,
Therefore their profit and weighted delay summation can be read directly.Logout site is the data having had, according to existing logout
Data find the logout ratio highest in third class.From table 3 it is observed that last line can be seen according to the result of the 3rd column
The gross profit of third class and total delay are all minimum out namely logistics logout site concentrates on class 3, and it is low to belong to retardation rate,
Low-income one kind attempts that region factor is added, so that it may find the concentration region of logistics logout sample.
Furthermore it is also possible to be clustered by express delivery handling capacity.The site express delivery handling capacity of different geographical may difference ---
Big city is few into having more, and the intensive place of electric business is more into going out less;Great Qu is accurate only to the analysis of site region.It is average daily with site
It receives, part amount is sent to cluster, Hierarchical Clustering suggests k-means cluster.By average daily income and part amount field is sent to carry out Kmeans cluster, side
Method is same as above, it is therefore an objective to influence of the express delivery handling capacity to logout is verified, and handling capacity is also one of the measurement factor of site income, by
In handling capacity by regional impact, hereafter to provide foundation for province information as dummy variable (dummy variable).
S13, analysis a number of factors relevant to the gross profit of logistics node, the revenue and expenditure system for drawing logistics node accordingly are dynamic
Mechanical model
In this present embodiment, based on the missing of truthful data, the logistics node is estimated using system dynamics flow graph
Receipts and expenditures, the logout possibility of logistics node described in analogue simulation.
As can be seen that profit, which is, influences the major reason of logistics node logout behavior, therefore in last point of analysis
The receipts and expenditures for surrounding logistics node are modeled in next analysis.
Fig. 2A be related to article receiving and sending information and miscellaneous receipt to logistics node and expenditure estimate after the list drawn
The revenue and expenditure system dynamics flow graph of a logistics node.In order to attempt the angle from receipts and expenditures (with reference to logistics node location)
The relationship of data with existing and logistics logout possibility is found out, as shown in Figure 2 B, does following regression analysis.
S14, it is based on the revenue and expenditure system dynamics model, establishes the recurrence point for predicting logistics node logout possibility
Analyse model, and the prediction result by the solution of the regression analysis model as logistics node health status.
1) classification panel data regression
Data processing explanation: (1) province information is converted into dummy variable;(2) for class imbalance (class-
) the considerations of problem imbalance (refer to the very big situation of training examples number difference different classes of in classification task, as adopt
In sample data, there is the logout site sample of data information to only have 6), logistics logout dot data is reconstructed into 5000+
A sample, and non-logout sample constitute matched sample.It is specific: " over-sampling " (oversampling) method to be used, by 6 objects
It flows the duplication of logout site sample repeatedly, slight error is subject to for management data (referring to receipts and expenditures data), while reconstructing volume
Logistics logout site of the code as " new ";(3) profit field is reconstructed, the non-logout sample profit of unified definition is 10000, fixed
Adopted logout sample is 0.By site reward structure, have:
Profit=Income-cost-delay.punishment
=Incomer+Incomes-costr-costs-delay.punishment
IncomerFor addressee income, i.e., direct income acquired by express delivery is received from client there, since each site is charged
Difference, usable 2SLS fit match value;IncomesTo send part to be taken in;costrFor addressee expenditure, i.e., and organize taking for interaction
With, comprising: weighing take, transfer charges, send part take etc.;costsTo send part to be paid, i.e., artificial and vehicle expense etc. may be set to send
The quadratic function form of part amount.Due to Incomes、costrIt is known, it is assumed that addressee income is the weighting function of all addressee attributes,
The quadratic function form of real label amount is used using green hand's number of packages, Jingdone district number of packages etc. as the instrumental variable singly measured, outbox cost is opened, i.e.,
costs=A*true.sign2+B*true.sign+C
Delay sends part punishment using weighting number of days weighting summation form:
Delay.punishiment=a*delay1+b*delay2+c*delay2plus
Mode is returned using two-stage-OLS, using point_code and date as the specific variable of panel data and time
Node sets are returned using province as classified variable.Due to only having the logout net in Guangdong and Hunan in sample data
Point information, so only exporting the regression result in Guangdong and Hunan, regression model formula is as follows:
Regression result is as shown in table 4, and model is examined by F, and the side R is 0.55 after model (1) adjustment, and model (2) is 0.67.
The side R is the coefficient of determination evaluated the fitting effect of recurrence.The side R is equal to regression sum of square ratio shared in total sum of squares
Rate reflects the construable ratio of regression equation institute in total variation of dependent variable y.Since the increase of the number of arguments will affect
The size of the coefficient of determination side R.When increasing independent variable, residual sum of squares (RSS) can be made to reduce, so that the side R be made to become larger.To avoid increasing
Independent variable and over-evaluate the side R, statistician propose with the number (k) of sample size (n) and independent variable go adjustment R.The side R is simultaneously after adjustment
Consider the influence of sample size n and the number of arguments k, this make adjustment after the side's R perseverance be less than the side R, and adjust after the side R value not
1 can be become closer to due to the increase of the number of arguments in returning.
The side's R calculating formula is as follows:
Inexplicable part in fit equation is subtracted with 1.
The side's R calculating formula is as follows after adjustment:
4 regression result of table
Numerical value in 4 bracket of table is t statistic, i.e., test coefficient whether the statistic significantly different from zero.
As can be drawn from Table 4 to draw a conclusion:
(1) excessive to send part amount that carry out band pressure to site, rather than the source of juice;
(2) it sends part amount to show the feature of invertedU curve really, i.e., after part amount should being sent to reach certain number, can lead instead
Cause the decline of site profit, wherein the bearing capacity of Hunan site is lower than Guangdong site;
(3) from the point of view of delay the case where sending part, slight delay sends part to be conducive to the operation of site, only excessive delay instead
It is possible to lead to the jumbo decline of profit.
2) Multiple Non Linear Regression
Regression analysis is only capable of finding out the influence relationship between variable, can not as the foundation for judging site business circumstance,
The order of magnitude also simply indicates that relativeness, cannot function as decision references, so the side of Multiple Non Linear Regression should further be used
Method carries out the fitting of site profit, and then predicts the truth of each site.
It proposes with drag:
Profit=send part profit+addressee profit-delay loss
=Incomes-(A*true.sign2+B*true.sign)
+(C*billing.quantity2+D*billing.quantity)-Costr
-E*delay1-F*delay2-G*delay2plus+H
Wherein, A, B, C, D, E, F, G, H are parameter to be estimated, and obtain site profit according to the calculation method of hereinbefore profit,
After sample data classification by geographical area, training set 70%, test set 30%, set seed 10101 are randomly selected.
The method of training set Multiple Non Linear Regression is fitted, obtained model is surveyed in test set
This process is repeated several times to find out as gap is calculated with the corresponding difference such as mean square deviation of match value using actual data value in examination
So that calculating each the smallest parameter to be estimated of gap.By fitting result it can be concluded that the area Guang Dong great site is being sent
Part profit is sent to be maximized when part amount true.sign=1350, Hunan is 887, then sends decreasing returns with charge free more than this quantity;
The area Guang Dong great opens single income with single amount monotonic increase is opened, and average every one express delivery income of receipts is 3.48 yuan;And the net in the area Hu Nan great
From the point of view of point is according to the quadratic function of rough estimate, addressee amount is more than that 125 sites can just maintain addressee profit.
So far, the hypothesis previously proposed is demonstrated above --- profit is to influence the fundamental factor of site logout behavior,
Therefore the following prediction that logistics node logout behavior will be carried out using logit model constructs field out_service, definition
Logout sample value is 1, and non-logout sample value is 0;Still training set/test set partitioned nodes of previous section are used, list is opened in addition
It is as follows to establish logit model as control variable for amount/worksheet processing amount:
Wherein, profit value is using the above obtained predicted value of Multiple Non Linear Regression.
Prediction result explanation: (1) from the point of view of prediction result, recall ratio is very high --- there is a logout site totally 6 of numerical value, one
5 are predicted altogether;(2) reconstructed sample is the sample due to class-imbalance problem reformulation, 6. reconstruct mode is shown in annotation,
Because these samples are constructed according to logout sample, thus it is 1 that health estimation value is accrued;(3) health status is predicted
Value is 1, and the sample that actual value is 0 is that prediction has logout behavior, practical still in the sample of business --- this part sample
Due to logout site similarity highest, profit estimated value is also closest, thus progression risk is very high, be worth improve concern
Degree.
Realize that all or part of the steps of above-mentioned each method embodiment can be by the relevant hardware of computer program come complete
At.Based on this understanding, the present invention also provides a kind of computer program products, including one or more computer instructions.Institute
Stating computer instruction may be stored in a computer readable storage medium.The computer readable storage medium can be computer
Any usable medium that can be stored either includes the data such as one or more usable mediums integrated server, data center
Store equipment.The usable medium can be magnetic medium (such as: floppy disk, hard disk, tape), optical medium (such as: DVD) or half
Conductive medium (such as: solid state hard disk Solid State Disk (SSD)).
Refering to Fig. 3, the present embodiment provides a kind of forecasting systems 30 of logistics node health status, carry as a software
In electronic equipment, to execute the prediction technique of logistics node health status described in preceding method embodiment at runtime.By
It is similar thus no longer thin to same technology in the technical principle and the technical principle of preceding method embodiment of this system embodiment
Section does repeatability and repeats.
The forecasting system of the logistics node health status of the present embodiment specifically includes: the analysis of the first analysis module 31, second
Module 32, third analysis module 33, simulation modeling module 34, regression analysis module 35.First analysis module 31 is for analyzing shadow
The potential factor of logistics node health status is rung, to obtain the key factor for influencing logistics node health status;Second analysis mould
Block 32 sends part income and retardation rate for analyte stream site, to obtain the key feature of the logistics node of logout;Third point
Analysis module 33 is for analyzing a number of factors relevant to the gross profit of logistics node;Simulation modeling module 34 is used for based on described the
The analysis result of three analysis modules draws the revenue and expenditure system dynamics model of logistics node;Regression analysis module 35 is used to be based on institute
Revenue and expenditure system dynamics model is stated, establishes the regression analysis model for predicting logistics node logout possibility, and described will return
Return prediction result of the solution of analysis model as logistics node health status.
It will be appreciated by those skilled in the art that the division of the modules in Fig. 3 embodiment is only a kind of logic function
Division, can completely or partially be integrated on one or more physical entities in actual implementation.And these modules can be whole
It is realized, can also be all realized in the form of hardware by way of processing element calls with software, it can be logical with part of module
Crossing processing element calls the form of software to realize that part of module passes through formal implementation of hardware.For example, simulation modeling module 34 can
Think the processing element individually set up, also can integrate and realized in some chip, in addition it is also possible to the shape of program code
Formula is stored in memory, and is called by some processing element and executed the function of simulation modeling module 34.The reality of other modules
It is now similar therewith.Processing element described here can be a kind of integrated circuit, the processing capacity with signal.In the process of realization
In, each step of the above method or the above modules can by the integrated logic circuit of the hardware in processor elements or
The instruction of software form is completed.
Refering to Fig. 4, the present embodiment provides a kind of electronic equipment, electronic equipment can be with desktop computer, portable computer, intelligent hand
The equipment such as machine.Detailed, electronic equipment, which includes at least, passes through what bus 41 connected: memory 42, processor 43, wherein storage
Device 42 is for storing computer program, and processor 43 is used to execute the computer program of the storage of memory 42, to execute aforementioned side
All or part of the steps in method embodiment.
System bus mentioned above can be Peripheral Component Interconnect standard (Peripheral Pomponent
Interconnect, abbreviation PCI) bus or expanding the industrial standard structure (Extended Industry Standard
Architecture, abbreviation EISA) bus etc..The system bus can be divided into address bus, data/address bus, control bus etc..
Only to be indicated with a thick line in figure, it is not intended that an only bus or a type of bus convenient for indicating.Communication connects
Mouth is for realizing the communication between database access device and other equipment (such as client, read-write library and read-only library).Storage
Device may include random access memory (Random Access Memory, abbreviation RAM), it is also possible to further include non-volatile deposit
Reservoir (non-volatile memory), for example, at least a magnetic disk storage.
Above-mentioned processor can be general processor, including central processing unit (Central Processing Unit,
Abbreviation CPU), network processing unit (Network Processor, abbreviation NP) etc.;It can also be digital signal processor
(Digital Signal Processing, abbreviation DSP), specific integrated circuit (Application Specific
Integrated Circuit, abbreviation ASIC), field programmable gate array (Field-Programmable Gate Array,
Abbreviation FPGA) either other programmable logic device, discrete gate or transistor logic, discrete hardware components.
In conclusion prediction technique, system, storage medium and the equipment of logistics node health status of the invention, can mention
Before predict the health status of logistics node, consequently facilitating policymaker has found to avoid there are the logistics node of logout risk in time
Economic loss.So the present invention effectively overcomes various shortcoming in the prior art and has high industrial utilization value.
Traditional view thinks, express delivery site logout is personnel caused by the profit as caused by price competition is meagre and treatment
It is lost, is that site profitability is insufficient after all.The present invention has found " profit using principal component analysis and clustering method
It is to lead to site logout behavior main cause ", by using emulation mode and regression analysis (regression analysis and Multiple Non Linear Regression)
Data are analyzed, classification by geographical area has obtained the quantitative relation between correlated variables and site profit, and passes through gained model
It is predicted, prediction result precision is good.By means of the present invention it can be concluded that
1) send part amount and site profit may inverted u-shaped relationship, the excessive rising for sending part task both to result in retardation rate,
Also operated pressure is brought to site;
2) part site profit is meagre, and addressee can just generate profit growth after reaching a certain level;
3) method for using regression analysis and Multiple Non Linear Regression sends part amount/open the development law singly measured in conjunction with site,
It largely may be implemented to manage the accurate prediction deteriorated to site;Simultaneously during actual prediction, discovery fitting benefit
Moisten 4000 or so, send part amount/open single site logout probability highest of the amount lower than 200.
The above-described embodiments merely illustrate the principles and effects of the present invention, and is not intended to limit the present invention.It is any ripe
The personage for knowing this technology all without departing from the spirit and scope of the present invention, carries out modifications and changes to above-described embodiment.Cause
This, institute is complete without departing from the spirit and technical ideas disclosed in the present invention by those of ordinary skill in the art such as
At all equivalent modifications or change, should be covered by the claims of the present invention.
Claims (12)
1. a kind of prediction technique of logistics node health status, which comprises the steps of:
The potential factor of analyzing influence logistics node health status, to obtain the key factor for influencing logistics node health status;
Wherein, the key factor includes: to send part income and retardation rate;
Part income and retardation rate are sent in analyte stream site, to obtain the key feature of the logistics node of logout;It is described crucial special
Sign includes: the gross profit of logistics node;
A number of factors relevant to the gross profit of logistics node is analyzed, draws the revenue and expenditure system dynamics mould of logistics node accordingly
Type;
Based on the revenue and expenditure system dynamics model, the regression analysis model for predicting logistics node logout possibility is established,
And the prediction result by the solution of the regression analysis model as logistics node health status.
2. the method according to claim 1, wherein the regression analysis model specifically:
Wherein, profit, billing.quantity, true.sign respectively indicate the gross profit of the logistics node, addressee
It measures and sends part amount, p indicates the logout probability of the logistics node, β0、β1、β2Respectively constant, ε are random error.
3. the method according to claim 1, wherein the mathematical model of the gross profit profit are as follows:
Gross profit profit=sends part profit+addressee profit-delay loss;Wherein,
It is described that part profit=send part is sent to take in IncomesPart is sent to pay costs;
The addressee profit=addressee takes in IncomerAddressee pays Costr;
The delay loss=E* postpones one day loss delay1-F* and postpones loss delay2-G* delay in two days two days or more to damage
Lose delay2plus+H;
Wherein, E, F, G, H are respectively constant.
4. according to the method described in claim 3, it is characterized in that, described send part to pay costsMathematical model are as follows:
Part is sent to pay costs=A*true.sign2+ B*true.sign, wherein A, B are respectively constant, and true.sign indicates institute
That states logistics node sends part amount.
5. according to the method described in claim 3, it is characterized in that, the addressee takes in IncomerMathematical model are as follows:
Addressee takes in Incomer=C*billing.quantity2+ D*billing.quantity, wherein be respectively C, D, normal
Number, billing.quantity indicate the addressee amount of the logistics node.
6. a kind of forecasting system of logistics node health status characterized by comprising
First analysis module, it is strong to obtain influence logistics node for the potential factor of analyzing influence logistics node health status
The key factor of health situation;Wherein, the key factor includes: to send part income and retardation rate;
Second analysis module sends part income and retardation rate for analyte stream site, to obtain the pass of the logistics node of logout
Key feature;The key feature includes: the gross profit of logistics node;
Third analysis module, for analyzing a number of factors relevant to the gross profit of logistics node;
Simulation modeling module draws the revenue and expenditure system dynamic of logistics node for the analysis result based on the third analysis module
Learn model;
Regression analysis module is established for being based on the revenue and expenditure system dynamics model for predicting that logistics node logout may
Property regression analysis model, and by prediction result of the solution as logistics node health status of the regression analysis model.
7. system according to claim 6, which is characterized in that the regression analysis model specifically:
Wherein, profit, billing.quantity, true.sign respectively indicate the gross profit of the logistics node, addressee
It measures and sends part amount, p indicates the logout probability of the logistics node, β0、β1、β2Respectively constant, ε are random error.
8. system according to claim 6, which is characterized in that the mathematical model of the gross profit profit are as follows:
Gross profit profit=sends part profit+addressee profit-delay loss;Wherein,
It is described that part profit=send part is sent to take in IncomesPart is sent to pay costs;
The addressee profit=addressee takes in IncomerAddressee pays Costr;
The delay loss=E* postpones one day loss delay1-F* and postpones loss delay2-G* delay in two days two days or more to damage
Lose delay2plus+ piece;
Wherein, E, F, G, H are respectively constant.
9. system according to claim 8, which is characterized in that described that part is sent to pay costsMathematical model are as follows:
Part is sent to pay costs=A*true.sign2+ B*true.sign, wherein A, B are respectively constant, and true.sign indicates institute
That states logistics node sends part amount.
10. system according to claim 8, which is characterized in that the addressee takes in IncomerMathematical model are as follows:
Addressee takes in Incomer=C*billing.quantity2+ D*billing.quantity, wherein be respectively C, D, normal
Number, billing.quantity indicate the addressee amount of the logistics node.
11. a kind of storage medium, wherein being stored with computer program, which is characterized in that the computer program is added by processor
When carrying execution, the prediction technique of the logistics node health status as described in any in claim 1 to 5 is realized.
12. a kind of electronic equipment characterized by comprising processor and memory;Wherein,
The memory is for storing computer program;
The processor is for computer program described in load and execution, so that the electronic equipment is executed as in claim 1 to 5
The prediction technique of any logistics node health status.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811123053.7A CN109376908A (en) | 2018-09-26 | 2018-09-26 | Prediction technique, system, storage medium and the equipment of logistics node health status |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811123053.7A CN109376908A (en) | 2018-09-26 | 2018-09-26 | Prediction technique, system, storage medium and the equipment of logistics node health status |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109376908A true CN109376908A (en) | 2019-02-22 |
Family
ID=65402557
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811123053.7A Pending CN109376908A (en) | 2018-09-26 | 2018-09-26 | Prediction technique, system, storage medium and the equipment of logistics node health status |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109376908A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113011666A (en) * | 2021-03-29 | 2021-06-22 | 上海寻梦信息技术有限公司 | Aging prediction model training method, aging prediction method and related equipment |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104156836A (en) * | 2013-05-13 | 2014-11-19 | 阿里巴巴集团控股有限公司 | Logistic network load query method and system |
CN106156966A (en) * | 2015-04-03 | 2016-11-23 | 阿里巴巴集团控股有限公司 | Logistics monitoring method and equipment |
DE102017009959A1 (en) * | 2017-10-26 | 2018-07-19 | Daimler Ag | Method for optimizing a logistics process |
-
2018
- 2018-09-26 CN CN201811123053.7A patent/CN109376908A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104156836A (en) * | 2013-05-13 | 2014-11-19 | 阿里巴巴集团控股有限公司 | Logistic network load query method and system |
CN106156966A (en) * | 2015-04-03 | 2016-11-23 | 阿里巴巴集团控股有限公司 | Logistics monitoring method and equipment |
DE102017009959A1 (en) * | 2017-10-26 | 2018-07-19 | Daimler Ag | Method for optimizing a logistics process |
Non-Patent Citations (4)
Title |
---|
BIN YE 等: "Research on the Express Delivery Delay Prediction Based on Neural Network in the Background of Big Data", 《HTTPS://DOI.ORG/10.2991/EMIM-16.2016.294》 * |
孙东川 等: "系统动力学建模与回归分析在现代物流业中的组合应用", 《太原理工大学学报》 * |
岳俊: "基于主成分回归的企业物流成本多元线性预测模型研究", 《价值工程》 * |
钱枫林 等: "基于主成分聚类分析的快递物流发展水平综合评价", 《物流科技》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113011666A (en) * | 2021-03-29 | 2021-06-22 | 上海寻梦信息技术有限公司 | Aging prediction model training method, aging prediction method and related equipment |
CN113011666B (en) * | 2021-03-29 | 2024-04-26 | 上海寻梦信息技术有限公司 | Training method of aging prediction model, aging prediction method and related equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Bussmann et al. | Explainable machine learning in credit risk management | |
Bai et al. | Banking credit worthiness: Evaluating the complex relationships | |
Jiang et al. | A prediction-driven mixture cure model and its application in credit scoring | |
Alfaro-Navarro et al. | A fully automated adjustment of ensemble methods in machine learning for modeling complex real estate systems | |
Sentas et al. | Categorical missing data imputation for software cost estimation by multinomial logistic regression | |
WO2023040494A1 (en) | Resource recommendation method, and multi-target fusion model training method and apparatus | |
He et al. | LSTM with particle Swam optimization for sales forecasting | |
Hayden et al. | Statistical methods to develop rating models | |
Yiping et al. | An improved multi-view collaborative fuzzy C-means clustering algorithm and its application in overseas oil and gas exploration | |
Li et al. | Multiattribute fuzzy decision evaluation approach and its application in enterprise competitiveness evaluation | |
Zhang | Research on precision marketing based on consumer portrait from the perspective of machine learning | |
CN113450004A (en) | Power credit report generation method and device, electronic equipment and readable storage medium | |
CN109376908A (en) | Prediction technique, system, storage medium and the equipment of logistics node health status | |
Wu et al. | The BP neural network with adam optimizer for predicting audit opinions of listed companies. | |
CN115841345A (en) | Cross-border big data intelligent analysis method, system and storage medium | |
Sun et al. | Prediction of Information Talent Demand Based on the Grayscale Prediction Model and the BP Neural Network | |
Wenjing | BP neural network-based evaluation method for enterprise comprehensive performance | |
Bhat et al. | A multivariate hurdle count data model with an endogenous multiple discrete–continuous selection system | |
Xia et al. | Analysis and prediction of telecom customer churn based on machine learning | |
Guo et al. | Statistical decision research of long-term deposit subscription in banks based on decision tree | |
Sun et al. | Short-term stock price forecasting based on an svd-lstm model | |
Mo et al. | Dynamic spatiotemporal correlation coefficient based on adaptive weight | |
He et al. | Effectiveness of price limit on stock market network: A time-migrated DCCA approach | |
Li | Design of Customer Churn Early Warning System Based on Mobile Communication Technology Based on Data Mining | |
Tu | Analysis and prediction method of student behavior mining based on campus big data |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
TA01 | Transfer of patent application right | ||
TA01 | Transfer of patent application right |
Effective date of registration: 20220422 Address after: Room 1005, No. 707, Zhangyang Road, China (Shanghai) pilot Free Trade Zone, Pudong New Area, Shanghai Applicant after: Fei Xiang Applicant after: Feiliu (Guangzhou) Information Technology Co.,Ltd. Address before: Room 1005, No. 707, Zhangyang Road, China (Shanghai) pilot Free Trade Zone, Pudong New Area, Shanghai, 200120 Applicant before: Fei Xiang |
|
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190222 |