CN116136989A - Port cargo flow direction statistical method and system based on BP neural network algorithm - Google Patents

Port cargo flow direction statistical method and system based on BP neural network algorithm Download PDF

Info

Publication number
CN116136989A
CN116136989A CN202310395337.6A CN202310395337A CN116136989A CN 116136989 A CN116136989 A CN 116136989A CN 202310395337 A CN202310395337 A CN 202310395337A CN 116136989 A CN116136989 A CN 116136989A
Authority
CN
China
Prior art keywords
data
cargo flow
network model
neural network
sample
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
Application number
CN202310395337.6A
Other languages
Chinese (zh)
Other versions
CN116136989B (en
Inventor
赵衍维
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jiangsu Wurun United Shipping Internet Co ltd
Original Assignee
Jiangsu Wurun United Shipping Internet Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Jiangsu Wurun United Shipping Internet Co ltd filed Critical Jiangsu Wurun United Shipping Internet Co ltd
Priority to CN202310395337.6A priority Critical patent/CN116136989B/en
Publication of CN116136989A publication Critical patent/CN116136989A/en
Application granted granted Critical
Publication of CN116136989B publication Critical patent/CN116136989B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
    • G06Q10/083—Shipping
    • G06Q10/0838—Historical data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201—Market modelling; Market analysis; Collecting market data
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Strategic Management (AREA)
  • Development Economics (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Finance (AREA)
  • Accounting & Taxation (AREA)
  • General Physics & Mathematics (AREA)
  • Marketing (AREA)
  • Data Mining & Analysis (AREA)
  • General Business, Economics & Management (AREA)
  • Quality & Reliability (AREA)
  • Biophysics (AREA)
  • Operations Research (AREA)
  • Human Resources & Organizations (AREA)
  • Game Theory and Decision Science (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Tourism & Hospitality (AREA)
  • Computational Linguistics (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a port cargo flow direction statistical method and a port cargo flow direction statistical system based on BP neural network algorithm, which comprises the steps of firstly acquiring port cargo list information, inputting and forming a list data pool, and executing significant sample data and fuzzy sample data construction; then determining an initialization model A=A1 U.A2, and taking a higher confidence output result from the output results; performing neural network training; and finally, performing model test and output, and outputting a network model corresponding to each group of sample data in the inventory data pool as cargo flow direction information to realize statistics of cargo flow direction information. Through the scheme, the method and the device can extract the effective information related to the cargo flow direction from the port cargo list information; the method can realize the conventional derivation of simple data and significant parameters to obtain the cargo flow direction information, and also can dig out effective parameters from complex data as much as possible, so as to derive the cargo flow direction information, realize mutual complementation of the two and realize the cargo flow direction information statistics work efficiently.

Description

Port cargo flow direction statistical method and system based on BP neural network algorithm
Technical Field
The application belongs to the technical field of computers, and particularly relates to a port cargo flow direction statistical method and system based on a BP neural network algorithm.
Background
The port cargo flow direction refers to effective records from the port cargo source to the destination, has strong correlation with the operating condition of port enterprises, cargo demand, port throughput, port cargo operation efficiency and the like, and is especially commonly used for measuring the economic traffic from the port cargo source to the destination, a large number of data records can provide analysis basis for economic condition analysis, economic policy formulation and development prospect prediction for statistical bodies, and the port cargo flow direction is an economic analysis basic index which is frequently and more important to use. The main ports at home and abroad can record basic information such as weight, value, time limit, freight, cargo owner, agent, transfer mark, insurance policy and the like for the transported cargo, and the port cargo flow direction can be usually obtained by mining based on the recorded basic information, for example, a simple cargo flow direction record is formed based on two factors { cargo source place, cargo destination place }.
In practical use, the following difficulties are generally encountered in statistics of port cargo flow direction: (1) The recorded basic information standards of the cargo transportation of the ports are different in each country/region and port, the basic information registered by one or some ports is unnecessary optional information for other ports, such as a plurality of small ports which do not perform standardization, so that the shortage of statistical data items is easily caused, and the effective data mining is relatively difficult; (2) Although simple deductions can be performed from the recorded relevant basic data to obtain the required project information, for the case of complex deductions and massive data, the relation rule between the basic data is hidden, and it is difficult to intuitively obtain effective and accurate project information based on experience and simple analysis. For example, patent CN109711773a in the prior art discloses a method for counting the flow of container cargo flow, which uses a clustering algorithm to cluster standard attribute files, selects a cluster where the standard attribute files are located, classifies all transportation documents in the cluster based on rules, and finally calculates the container traffic and output information between transportation nodes.
Disclosure of Invention
Aiming at the problems existing in the prior art, the invention provides a port cargo flow direction statistical method and a port cargo flow direction statistical system based on a BP neural network algorithm.
In order to achieve the above purpose, the present application provides a port cargo flow direction statistical method based on a BP neural network algorithm, the method comprising:
step S110, acquiring port cargo list information, inputting and forming a list data pool, and executing significant sample data and fuzzy sample data construction;
step S120, determining an initialization model A=A1 U.A2, wherein A1 is a significant parameter sub-bp network model, A2 is a fuzzy parameter sub-bp network model, and a symbol U represents a higher confidence output result in the output results; performing neural network training;
and step S130, model testing and outputting, namely outputting a network model corresponding to each group of sample data in the inventory data pool as cargo flow information, and realizing statistics of the cargo flow information.
In one embodiment, in the step of constructing the salient sample data and the blurred sample data, the method further includes step S1101 of vectorizing the manifest data:
constructing a mapping table of the list data to a vector space; after acquiring port bill information, inputting and forming a bill data pool, randomly extracting bill data according to a preset number M from the bill data pool; mapping each set of the randomly extracted manifest data to a vector space zm= { am1, am2, … … amp } based on a mapping table; wherein n is the basic information quantity of each group of samples, 0< m < M;
step S1102 is further included, wherein a significant correlation parameter Zm1 of the cargo flow direction of the vector space Zm is extracted, and the manifest cargo flow direction Dm of each manifest data of the randomly extracted manifest data is calibrated based on a preset algorithm for the significant correlation parameter; if the calibration is successful, merging the calibration successful inventory cargo flow direction Dm based on the significance related parameter Zm1 to be used as data Tm1= { Zm1, dm } in the first training data set, removing the significance related parameter Zm1 based on Zm to obtain a fuzzy related parameter Zm2, merging the calibration successful inventory cargo flow direction Dm to be used as data Tm2= { Zm2, dm } in the second training data set, if the calibration is unsuccessful, throwing the randomly extracted inventory data back to the inventory data pool, simultaneously re-acquiring inventory data with the same quantity as the thrown back quantity, and repeating the calibration step until the calibrated quantity reaches the preset quantity M.
In one embodiment, the step S120 further includes: normalizing the confidence coefficient of the output results of the A1 and A2 models to a (0, 1) space based on the balance coefficient A1 scale、 a2 scale And carrying out coefficient weighting on the normalized confidence coefficient.
In one embodiment, the performing neural network training specifically includes:
s1201 parameter initializing step, including: initializing A1 model parameters: determining a first input layer node number n1 according to a Zm1 dimension in the Tm1 sequence, determining a first hidden layer node number k1 according to the first input layer node number n1, and determining a first output layer node number o1; initializing A2 model parameters: determining a second input layer node number n2 according to the Zm2 dimension in the Tm2 sequence, determining a second hidden layer node number k2 according to the second input layer node number n2, and determining a second output layer node number o2; respectively initializing connection weights and bias thresholds of two sub-bp network models, and selecting an activation function;
s1202 training step: inputting each sample in the sample sequence based on the Tm1 to A1, inputting each sample in the sample sequence based on the Tm2 to A2, and calculating each node function forward to output; calculating a loss function, solving a bias derivative of each weight value for the loss function, optimizing a reverse transmission process based on a gradient descent method, updating each parameter value of A1 and A2, and repeating the training steps until a preset training target is reached.
In one embodiment, the model testing and outputting step specifically includes:
and acquiring a test sample from the inventory data pool, inputting the test sample into the trained Bp neural network model A after training, and inputting inventory data in the inventory data pool into the trained Bp neural network model to obtain network model output as cargo flow information if the model output meets the expectation, so as to realize statistics of the cargo flow information.
The application also provides a port cargo flow direction statistical system based on BP neural network algorithm, comprising:
the sample construction module is used for acquiring port cargo list information, inputting and forming a list data pool, and executing construction of remarkable sample data and fuzzy sample data;
the neural network training module is used for determining an initialization model A=A1U A2, wherein A1 is a significant parameter sub-bp network model, A2 is a fuzzy parameter sub-bp network model, and a symbol U represents a higher confidence output result in the output results; performing neural network training;
and the test and output module is used for model test and output, outputting a network model corresponding to each group of sample data in the list data pool as cargo flow direction information, and realizing statistics of cargo flow direction information.
In an embodiment, the sample construction module further includes a manifest data vectorization module, configured to perform manifest data vectorization: constructing a mapping table of the list data to a vector space; after acquiring port bill information, inputting and forming a bill data pool, randomly extracting bill data according to a preset number M from the bill data pool; mapping each set of the randomly extracted manifest data to a vector space zm= { am1, am2, … … amp } based on a mapping table; wherein n is the basic information quantity of each group of samples, 0< m < M; the sample construction module further comprises a significant sample data and fuzzy sample data construction module, which is used for extracting a significant correlation parameter Zm1 of the cargo flow direction of the vector space Zm, and calibrating the manifest cargo flow direction Dm of each manifest data of the randomly extracted manifest data based on a preset algorithm; if the calibration is successful, merging the calibration successful inventory cargo flow direction Dm based on the significance related parameter Zm1 to be used as data Tm1= { Zm1, dm } in the first training data set, removing the significance related parameter Zm1 based on Zm to obtain a fuzzy related parameter Zm2, merging the calibration successful inventory cargo flow direction Dm to be used as data Tm2= { Zm2, dm } in the second training data set, if the calibration is unsuccessful, throwing the randomly extracted inventory data back to the inventory data pool, simultaneously re-acquiring inventory data with the same quantity as the thrown back quantity, and repeating the calibration step until the calibrated quantity reaches the preset quantity M.
In one embodiment, the neural network training module is further configured to normalize the confidence levels of the output results of the A1 and A2 models to a (0, 1) space based on the balance coefficient A1 scale、 a2 scale And carrying out coefficient weighting on the normalized confidence coefficient.
In one embodiment, the neural network training module further comprises:
the parameter initialization module is used for initializing A1 model parameters: determining a first input layer node number n1 according to a Zm1 dimension in the Tm1 sequence, determining a first hidden layer node number k1 according to the first input layer node number n1, and determining a first output layer node number o1; initializing A2 model parameters: determining a second input layer node number n2 according to the Zm2 dimension in the Tm2 sequence, determining a second hidden layer node number k2 according to the second input layer node number n2, and determining a second output layer node number o2; respectively initializing connection weights and bias thresholds of two sub-bp network models, and selecting an activation function;
the training module is used for inputting each sample in the sample sequence based on the Tm1 to A1, inputting each sample in the sample sequence based on the Tm2 to A2, and calculating each node function output in the forward direction; calculating a loss function, solving a bias derivative of each weight value for the loss function, optimizing a reverse transmission process based on a gradient descent method, updating each parameter value of A1 and A2, and repeating the training steps until a preset training target is reached.
In one embodiment, the test and output module is further to:
and acquiring a test sample from the inventory data pool, inputting the test sample into the trained Bp neural network model A after training, and if the model output accords with the expectation, inputting the inventory data in the inventory data pool into the trained Bp neural network model A to obtain the network model output as the cargo flow information, so as to realize the statistics of the cargo flow information.
The invention provides a port cargo flow direction statistical method and a port cargo flow direction statistical system based on BP neural network algorithm, which comprise the steps of acquiring port cargo list information, inputting and forming a list data pool, and executing significant sample data and fuzzy sample data construction; determining an initialization model A=A1 U.A2, and executing neural network training; and finally, testing and outputting based on the trained model, and outputting the network model corresponding to each group of sample data in the inventory data pool as cargo flow information to realize statistics of cargo flow information. The method or the system can at least achieve the following effects: 1, carrying out big data standardization processing on port bill information, and mining effective information related to the flow direction of the cargoes; the trained neural network model not only can deduce the cargo flow information from parameters obviously related to the cargo flow information, but also can deduce the cargo flow information based on fuzzy parameters which are not obviously related to the cargo flow information, so that the cargo flow information can be obtained by conventional deduction of simple data and obvious parameters for various port cargo list information, and effective parameters can be extracted from complex data as much as possible, so that the cargo flow information is deduced, the cargo flow information and the cargo flow information are mutually complementary, and the cargo flow information statistics work is effectively realized.
Drawings
Fig. 1 is a flowchart of a port cargo flow statistical method based on a BP neural network algorithm according to an embodiment of the present invention;
fig. 2 is a diagram of a port cargo flow direction statistical system based on a BP neural network algorithm according to an embodiment of the present invention;
fig. 3 is a hardware configuration diagram of an electronic device according to an embodiment of the present application.
Detailed Description
The present application will be described in detail with reference to the embodiments shown in the drawings. The embodiments are not intended to be limiting and structural, methodological, or functional changes made by those of ordinary skill in the art in light of the embodiments are intended to be included within the scope of the present application.
The terms "first," "second," "third," "fourth" and the like in the description and in the claims of this application and in the above-described figures, if any, are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that embodiments of the present application described herein may be capable of operation in sequences other than those illustrated or described herein, for example. Furthermore, the terms "comprises," "comprising," and "includes" and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed or inherent to such process, method, article, or apparatus.
The port cargo flow statistical method based on the BP neural network algorithm can be applied to artificial intelligence. Artificial intelligence (Artificial Intelligence, AI) is the theory, method, technique and application system that uses a digital computer or a machine controlled by a digital computer to simulate, extend and extend human intelligence, sense the environment, acquire knowledge and use the knowledge to obtain optimal results. In other words, artificial intelligence is an integrated technology of computer science that attempts to understand the essence of intelligence and to produce a new intelligent machine that can react in a similar way to human intelligence. Artificial intelligence, i.e. research on design principles and implementation methods of various intelligent machines, enables the machines to have functions of sensing, reasoning and decision.
Artificial intelligence (Artificial Intelligence, AI): the system is a theory, a method, a technology and an application system which simulate, extend and extend human intelligence by using a digital computer or a machine controlled by the digital computer, sense environment, acquire knowledge and acquire an optimal result by using the knowledge.
It should be noted that artificial intelligence is an integrated technology of computer science that attempts to understand the essence of intelligence and to produce a new intelligent machine that can react in a similar way to human intelligence. Artificial intelligence, i.e. research on design principles and implementation methods of various intelligent machines, enables the machines to have functions of sensing, reasoning and decision.
In addition, the artificial intelligence technology is a comprehensive discipline, and relates to the technology with wide fields, namely the technology with a hardware level and the technology with a software level. Artificial intelligence infrastructure technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation/interaction systems, mechatronics, and the like. The artificial intelligence software technology mainly comprises a computer vision technology, a voice processing technology, a natural language processing technology, machine learning/deep learning and other directions.
With the research and advancement of artificial intelligence technology, artificial intelligence technology has been developed for research and application in a variety of fields; for example, common smart homes, smart wearable devices, virtual assistants, smart speakers, smart marketing, unmanned, automatic driving, unmanned, robotic, smart medical and smart customer service, and the like; with the development of technology, artificial intelligence technology will find application in more fields and will develop more and more important value. In the embodiment of the invention, the application of artificial intelligence in deep learning, in particular to the statistics of port cargo flow by using BP neural network will be described.
Embodiment one: as shown in fig. 1, the present invention provides a port cargo flow statistical method based on BP neural network algorithm, which includes:
step S110, acquiring port cargo list information, inputting and forming a list data pool, and executing significant sample data and fuzzy sample data construction.
The step of constructing the salient sample data and the blurred sample data further includes a step S1101 of vectorizing the manifest data: constructing a mapping table of the list data to a vector space; after acquiring port bill information, inputting and forming a bill data pool, randomly extracting bill data according to a preset number M from the bill data pool; mapping each set of the randomly extracted manifest data to a vector space zm= { am1, am2, … … amp } based on a mapping table; where n is the number of basic information per set of samples, 0< m.
It will be appreciated that each set of samples (data) herein includes one salient sample data and one ambiguous sample data.
The port bill information content is in various formats, and the graphic and text information cannot be directly processed in the mathematical model, so that in the preprocessing process, a mapping table from the port bill information content to a vector space is set, and the bill contents in different formats are unified into a processable format; furthermore, the mapped data can be subjected to data alignment and normalization processing, so that the data dispersion is reduced, and the data processing application is facilitated.
In the step of constructing the salient sample data and the fuzzy sample data, the method further includes step S1102 of extracting a salient correlation parameter Zm1 of the cargo flow direction of the vector space Zm, and calibrating a manifest cargo flow direction Dm of each manifest data of the randomly extracted manifest data based on a preset algorithm for the salient correlation parameter; if the calibration is successful, merging the calibrated list cargo flow Dm based on the significance related parameter Zm1 to be used as data Tm1= { Zm1, dm } in the first training data set, removing the significance related parameter Zm1 based on Zm to obtain a fuzzy related parameter Zm2, and merging the calibrated list cargo flow Dm to be used as data Tm2= { Zm2, dm } in the second training data set. If the calibration is unsuccessful, the randomly extracted list data is returned to the list data pool, list data with the same number as the returned list data are obtained again, and the calibration step is repeated until the calibrated number reaches the preset number M. Finally, normalizing the confidence coefficient of the output result of the A1 and A2 models to a (0, 1) space and based on the balance coefficient A1 scale、 a2 scale And carrying out coefficient weighting on the normalized confidence coefficient.
The sample vector contains various parameters of irregularity including a saliency-related parameter, a blur-related parameter, and a combination of the two. In this embodiment, the sample vector space is divided into significant correlation parameters and fuzzy correlation parameters. The saliency related parameters can be, for example, a delivery site, a target country, a place of origin, a cargo receiving party and the like, port cargo flow direction information can be deduced relatively simply based on the saliency related parameters, the saliency related parameters are combined with corresponding deduction results to be used as training parameters, a first sub-bp network model can be obtained through training, the network model is high in training efficiency, the iteration number is small, and the result confidence is high.
However, not all sample vectors are or contain significant correlation parameters, and port statistics of non-standardized operations will typically only contain fuzzy correlation parameters, and it is often difficult to obtain port cargo flow information based on the fuzzy correlation parameters under conventional empirical derivation. For example, the type of cargo, the transportation route, the coordinates, the owner, the logistics cost, the customs process, etc. in the freight information, the correlation between the parameter information and the port cargo flow information is relatively fuzzy. Although the target information can not be directly obtained from the target information or the deviation between the target information obtained through simple deduction and the actual situation is larger, the parameter information or the combination of the parameter information can show the hidden probability law under the BP neural network model provided by the application, so that in the case, the fuzzy correlation parameters are obtained except the obvious correlation parameters, and the successfully calibrated manifest cargo flow is combined and used as the training parameters in the second training data set, and the second sub-BP network model can be obtained after the training is completed.
Step S120, determining an initialization model a=a1 u_a2, and performing neural network training based on the significant sample data and the fuzzy sample data to obtain a trained BP neural network model.
Wherein A1 is a significant parameter sub-bp network model, A2 is a fuzzy parameter sub-bp network model, and a symbol U represents a higher confidence output result in the output results.
Further, the performing neural network training specifically includes:
s1201 parameter initializing step, including: initializing A1 model parameters: determining a first input layer node number n1 according to a Zm1 dimension in the Tm1 sequence, determining a first hidden layer node number k1 according to the first input layer node number n1, and determining a first output layer node number o1; initializing A2 model parameters: determining a second input layer node number n2 according to the Zm2 dimension in the Tm2 sequence, determining a second hidden layer node number k2 according to the second input layer node number n2, and determining a second output layer node number o2; respectively initializing connection weights and bias thresholds of two sub-bp network models, and selecting an activation function;
s1202 training step: inputting each sample in the sample sequence based on the Tm1 to A1, inputting each sample in the sample sequence based on the Tm2 to A2, and calculating each node function forward to output; calculating a loss function, solving a bias derivative of each weight value for the loss function, optimizing a reverse transmission process based on a gradient descent method, updating each parameter value of A1 and A2, and repeating the training steps until a preset training target is reached.
It may be understood that the preset training target herein may be, for example, that the number of training times reaches a threshold of iteration number of the significant parameter sub-bp network model and the fuzzy parameter sub-bp network model, or until the accuracy of the output results of the significant parameter sub-bp network model or the fuzzy parameter sub-bp network model reaches a predetermined threshold, which is not limited in this application.
Based on the BP neural network model provided by the embodiment, for various port bill information, the conventional derivation of simple data and significant parameters can be realized to obtain the flow information of the cargo, and the effective parameters can be extracted from complex data as much as possible, so that the flow information of the cargo is derived, and the two are mutually and complementarily verified.
And step S130, model testing and outputting, namely outputting a network model corresponding to each group of sample data in the inventory data pool as cargo flow information, and realizing statistics of the cargo flow information.
And acquiring a test sample from the inventory data pool, inputting the test sample into the trained Bp neural network model A after training, and inputting inventory data in the inventory data pool into the trained Bp neural network model to obtain network model output as cargo flow information if the model output meets the expectation, so as to realize statistics of the cargo flow information.
Embodiment two: as shown in fig. 2, the present invention further provides a port cargo flow direction statistical system based on a BP neural network algorithm, which includes:
the sample construction module is used for acquiring port cargo list information, inputting and forming a list data pool, and executing construction of remarkable sample data and fuzzy sample data;
the neural network training module is used for determining an initialization model A=A1U A2, wherein A1 is a significant parameter sub-bp network model, A2 is a fuzzy parameter sub-bp network model, and a symbol U represents a higher confidence output result in the output results; performing neural network training;
and the test and output module is used for model test and output, outputting a network model corresponding to each group of sample data in the list data pool as cargo flow direction information, and realizing statistics of cargo flow direction information.
Further, the sample construction module further includes a manifest data vectorization module, configured to perform manifest data vectorization: constructing a mapping table of the list data to a vector space; after acquiring port bill information, inputting and forming a bill data pool, randomly extracting bill data according to a preset number M from the bill data pool; mapping each set of the randomly extracted manifest data to a vector space zm= { am1, am2, … … amp } based on a mapping table; where n is the number of basic information per set of samples, 0< m.
Further, the sample construction module further includes a significant sample data and fuzzy sample data construction module, configured to extract a significant correlation parameter Zm1 of the cargo flow direction of the vector space Zm, and calibrate a manifest cargo flow Dm of each manifest data of the randomly extracted manifest data for the significant correlation parameter based on a preset algorithm; if the calibration is successful, merging the calibration successful inventory cargo flow direction Dm based on the significance related parameter Zm1 to be used as data Tm1= { Zm1, dm } in the first training data set, removing the significance related parameter Zm1 based on Zm to obtain a fuzzy related parameter Zm2, merging the calibration successful inventory cargo flow direction Dm to be used as data Tm2= { Zm2, dm } in the second training data set, if the calibration is unsuccessful, throwing the randomly extracted inventory data back to the inventory data pool, simultaneously re-acquiring inventory data with the same quantity as the thrown back quantity, and repeating the calibration step until the calibrated quantity reaches the preset quantity M.
Further, the neural network training module is further configured to normalize the confidence levels of the output results of the A1 and A2 models to a (0, 1) space based on the balance coefficient A1 scale、 a2 scale And carrying out coefficient weighting on the normalized confidence coefficient.
Further, the neural network training module further includes:
the parameter initialization module is used for initializing A1 model parameters: determining a first input layer node number n1 according to a Zm1 dimension in the Tm1 sequence, determining a first hidden layer node number k1 according to the first input layer node number n1, and determining a first output layer node number o1; initializing A2 model parameters: determining a second input layer node number n2 according to the Zm2 dimension in the Tm2 sequence, determining a second hidden layer node number k2 according to the second input layer node number n2, and determining a second output layer node number o2; respectively initializing connection weights and bias thresholds of two sub-bp network models, and selecting an activation function;
the training module is used for inputting each sample in the sample sequence based on the Tm1 to A1, inputting each sample in the sample sequence based on the Tm2 to A2, and calculating each node function output in the forward direction; calculating a loss function, solving a bias derivative of each weight value for the loss function, optimizing a reverse transmission process based on a gradient descent method, updating each parameter value of A1 and A2, and repeating the training steps until a preset training target is reached.
Further, the test and output module is further configured to:
and acquiring a test sample from the inventory data pool, inputting the test sample into the trained BP neural network model A, and if the model output accords with the expectation, inputting inventory data in the inventory data pool into the trained BP neural network model A to obtain network model output as cargo flow information, so as to realize statistics of the cargo flow information.
It can be seen that the present application can achieve at least the following effects by employing the above method or system: 1, carrying out big data standardization processing on port bill information, and mining effective information related to the flow direction of the cargoes; the trained neural network model not only can deduce the cargo flow information from parameters obviously related to the cargo flow information, but also can deduce the cargo flow information based on fuzzy parameters which are not obviously related to the cargo flow information, so that the cargo flow information can be obtained by conventional deduction of simple data and obvious parameters for various port cargo list information, and effective parameters can be extracted from complex data as much as possible, so that the cargo flow information is deduced, the cargo flow information and the cargo flow information are mutually complementary, and the cargo flow information statistics work is effectively realized.
Fig. 3 also shows a hardware configuration diagram of the electronic device according to the embodiment of the present specification. As shown in fig. 3, the electronic device 30 may include at least one processor 31, a memory 32 (e.g., a non-volatile memory), a memory 33, and a communication interface 34, and the at least one processor 31, the memory 32, the memory 33, and the communication interface 34 are connected together via an internal bus 35. The at least one processor 31 executes at least one computer readable instruction stored or encoded in the memory 32.
It should be understood that the computer-executable instructions stored in the memory 32, when executed, cause the at least one processor 31 to perform the various operations and functions described above in connection with fig. 1 in various embodiments of the present description.
In embodiments of the present description, electronic device 30 may include, but is not limited to: personal computers, server computers, workstations, desktop computers, laptop computers, notebook computers, mobile electronic devices, smart phones, tablet computers, cellular phones, personal Digital Assistants (PDAs), handsets, messaging devices, wearable electronic devices, consumer electronic devices, and the like.
According to one embodiment, a program product, such as a machine-readable medium, is provided. The machine-readable medium may have instructions (i.e., the elements described above implemented in software) that, when executed by a machine, cause the machine to perform the various operations and functions described above in connection with fig. 1 in various embodiments of the specification. In particular, a system or apparatus provided with a readable storage medium having stored thereon software program code implementing the functions of any of the above embodiments may be provided, and a computer or processor of the system or apparatus may be caused to read out and execute instructions stored in the readable storage medium.
In this case, the program code itself read from the readable medium may implement the functions of any of the above embodiments, and thus the machine-readable code and the readable storage medium storing the machine-readable code form part of the present specification.
Examples of readable storage media include floppy disks, hard disks, magneto-optical disks, optical disks (e.g., CD-ROMs, CD-R, CD-RWs, DVD-ROMs, DVD-RAMs, DVD-RWs), magnetic tapes, nonvolatile memory cards, and ROMs. Alternatively, the program code may be downloaded from a server computer or cloud by a communications network.
It will be appreciated by those skilled in the art that various changes and modifications can be made to the embodiments disclosed above without departing from the spirit of the invention. Accordingly, the scope of protection of this specification should be limited by the attached claims.
It should be noted that not all the steps and units in the above flowcharts and the system configuration diagrams are necessary, and some steps or units may be omitted according to actual needs. The order of execution of the steps is not fixed and may be determined as desired. The apparatus structures described in the above embodiments may be physical structures or logical structures, that is, some units may be implemented by the same physical client, or some units may be implemented by multiple physical clients, or may be implemented jointly by some components in multiple independent devices.
In the above embodiments, the hardware units or modules may be implemented mechanically or electrically. For example, a hardware unit, module or processor may include permanently dedicated circuitry or logic (e.g., a dedicated processor, FPGA or ASIC) to perform the corresponding operations. The hardware unit or processor may also include programmable logic or circuitry (e.g., a general purpose processor or other programmable processor) that may be temporarily configured by software to perform the corresponding operations. The particular implementation (mechanical, or dedicated permanent, or temporarily set) may be determined based on cost and time considerations.
The detailed description set forth above in connection with the appended drawings describes exemplary embodiments, but does not represent all embodiments that may be implemented or fall within the scope of the claims. The term "exemplary" used throughout this specification means "serving as an example, instance, or illustration," and does not mean "preferred" or "advantageous over other embodiments. The detailed description includes specific details for the purpose of providing an understanding of the described technology. However, the techniques may be practiced without these specific details. In some instances, well-known structures and devices are shown in block diagram form in order to avoid obscuring the concepts of the described embodiments.
The previous description of the disclosure is provided to enable any person skilled in the art to make or use the disclosure. Various modifications to the disclosure will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other variations without departing from the scope of the disclosure. Thus, the disclosure is not limited to the examples and designs described herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (10)

1. A port cargo flow direction statistical method based on a BP neural network algorithm, the method comprising:
step S110, acquiring port cargo list information, inputting and forming a list data pool, and executing significant sample data and fuzzy sample data construction;
step S120, determining an initialization model A=A1U A2, and performing neural network training based on the significant sample data and the fuzzy sample data to obtain a trained BP neural network model, wherein A1 is a significant parameter sub-BP network model, A2 is a fuzzy parameter sub-BP network model, and a symbol U represents a higher confidence output result in the output result;
and step S130, model testing and outputting, namely outputting a network model corresponding to each group of sample data in the list data pool as cargo flow information, and realizing statistics of the cargo flow information.
2. The port cargo flow direction statistical method based on the BP neural network algorithm according to claim 1, wherein in the significant sample data and fuzzy sample data constructing step, further comprising step S1101 of vectorizing the inventory data:
constructing a mapping table of the list data to a vector space; randomly extracting inventory data from the inventory data pool by a predetermined number M; mapping each set of manifest data of the random extraction to a vector space zm= { am1, am2, … … amp } based on the mapping table; wherein n is the basic information quantity of each group of samples, 0< m < M;
step S1102 is further included, wherein a significant correlation parameter Zm1 of the cargo flow direction of the vector space Zm is extracted, and the manifest cargo flow direction Dm of each manifest data of the randomly extracted manifest data is calibrated based on a preset algorithm for the significant correlation parameter; if the calibration is successful, merging the calibration successful inventory cargo flow direction Dm based on the significance related parameter Zm1 to be used as data Tm1= { Zm1, dm } in the first training data set, removing the significance related parameter Zm1 based on Zm to obtain a fuzzy related parameter Zm2, merging the calibration successful inventory cargo flow direction Dm to be used as data Tm2= { Zm2, dm } in the second training data set, if the calibration is unsuccessful, throwing the randomly extracted inventory data back to the inventory data pool, simultaneously re-acquiring inventory data with the same quantity as the thrown back quantity, and repeating the calibration step until the calibrated quantity reaches the preset quantity M.
3. The port cargo flow statistical method based on the BP neural network algorithm of claim 1, wherein the step S120 further comprises: normalizing the confidence coefficient of the output results of the obvious parameter sub-bp network model and the fuzzy parameter sub-bp network model to a (0, 1) space based on the balance coefficient a1 scale、 a2 scale For normalizationThe confidence levels after that are coefficient weighted.
4. The port cargo flow statistical method based on the BP neural network algorithm according to claim 2, wherein the performing the neural network training specifically comprises:
s1201 parameter initializing step, including: initializing a salient parameter sub-bp network model parameter: determining a first input layer node number n1 according to a Zm1 dimension in the Tm1 sequence, determining a first hidden layer node number k1 according to the first input layer node number n1, and determining a first output layer node number o1; initializing parameters of a fuzzy parameter sub-bp network model: determining a second input layer node number n2 according to the Zm2 dimension in the Tm2 sequence, determining a second hidden layer node number k2 according to the second input layer node number n2, and determining a second output layer node number o2; respectively initializing connection weights and bias thresholds of two sub-bp network models, and selecting an activation function;
s1202 training step: inputting each sample in the Tm1 sample sequence to a significant parameter sub-bp network model, inputting each sample in the Tm2 sample sequence to a fuzzy parameter sub-bp network model, and forward calculating each node function output; calculating a loss function, solving partial derivatives of all weights for the loss function, optimizing a reverse transmission process based on a gradient descent method, and updating all parameter values of a significant parameter sub-bp network model and a fuzzy parameter sub-bp network model; repeating the training steps until reaching the preset training target.
5. The port cargo flow statistical method based on the BP neural network algorithm according to claim 1, wherein the model testing and outputting step specifically comprises:
and acquiring a test sample from the inventory data pool, inputting the test sample into the trained BP neural network model, and inputting inventory data in the inventory data pool into the BP neural network model to obtain network model output as cargo flow information if the model output meets the expectation, so as to realize statistics of the cargo flow information.
6. The port cargo flow direction statistical system based on BP neural network algorithm is characterized by comprising:
the sample construction module is used for acquiring port cargo list information, inputting and forming a list data pool, and executing construction of remarkable sample data and fuzzy sample data;
the neural network training module is used for determining an initialization model A=A1U A2 and executing neural network training based on the significant sample data and the fuzzy sample data to obtain a trained BP neural network model, wherein A1 is a significant parameter sub-BP network model, A2 is a fuzzy parameter sub-BP network model, and a symbol U represents a higher confidence output result in the output result;
and the test and output module is used for model test and output, outputting a network model corresponding to each group of sample data in the list data pool as cargo flow direction information, and realizing statistics of cargo flow direction information.
7. The port cargo flow statistical system based on the BP neural network algorithm of claim 6, wherein the sample construction module further comprises a manifest data vectorization module for performing manifest data vectorization: constructing a mapping table of the list data to a vector space; randomly extracting inventory data from the inventory data pool by a predetermined number M; mapping each set of manifest data of the random extraction to a vector space zm= { am1, am2, … … amp } based on the mapping table; wherein n is the basic information quantity of each group of samples, 0< m < M;
the sample construction module further comprises a significant sample data and fuzzy sample data construction module, which is used for extracting a significant correlation parameter Zm1 of the cargo flow direction of the vector space Zm, and calibrating the manifest cargo flow direction Dm of each manifest data of the randomly extracted manifest data based on a preset algorithm; if the calibration is successful, merging the calibration successful inventory cargo flow direction Dm based on the significance related parameter Zm1 to be used as data Tm1= { Zm1, dm } in the first training data set, removing the significance related parameter Zm1 based on Zm to obtain a fuzzy related parameter Zm2, merging the calibration successful inventory cargo flow direction Dm to be used as data Tm2= { Zm2, dm } in the second training data set, if the calibration is unsuccessful, throwing the randomly extracted inventory data back to the inventory data pool, simultaneously re-acquiring inventory data with the same quantity as the thrown back quantity, and repeating the calibration step until the calibrated quantity reaches the preset quantity M.
8. The port cargo flow statistical system based on BP neural network algorithm as claimed in claim 6, wherein the neural network training module is further configured to normalize confidence levels of output results of the salient parameter sub-BP network model and the fuzzy parameter sub-BP network model to a (0, 1) space based on the balance coefficient a1 scale、 a2 scale And carrying out coefficient weighting on the normalized confidence coefficient.
9. The port cargo flow statistical system based on the BP neural network algorithm of claim 7, wherein the neural network training module further comprises:
the parameter initialization module is used for initializing the parameters of the obvious parameter sub-bp network model: determining a first input layer node number n1 according to a Zm1 dimension in the Tm1 sequence, determining a first hidden layer node number k1 according to the first input layer node number n1, and determining a first output layer node number o1; initializing parameters of a fuzzy parameter sub-bp network model: determining a second input layer node number n2 according to the Zm2 dimension in the Tm2 sequence, determining a second hidden layer node number k2 according to the second input layer node number n2, and determining a second output layer node number o2; respectively initializing connection weights and bias thresholds of two sub-bp network models, and selecting an activation function;
the training module is used for inputting each sample in the Tm1 sample sequence to the significant parameter sub-bp network model, inputting each sample in the Tm2 sample sequence to the fuzzy parameter sub-bp network model, and calculating each node function forward to output; calculating a loss function, solving partial derivatives of all weights for the loss function, optimizing a reverse transmission process based on a gradient descent method, and updating all parameter values of a significant parameter sub-bp network model and a fuzzy parameter sub-bp network model; repeating the training steps until reaching the preset training target.
10. The port cargo flow statistical system based on the BP neural network algorithm of claim 6, wherein the test and output module is further configured to:
and acquiring a test sample from the list data pool, inputting the test sample into the trained BP neural network model, and if the model output accords with the expectation, inputting the list data in the list data pool into the trained BP neural network model to obtain the network model output as the cargo flow information, so as to realize the statistics of the cargo flow information.
CN202310395337.6A 2023-04-14 2023-04-14 Port cargo flow direction statistical method and system based on BP neural network algorithm Active CN116136989B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310395337.6A CN116136989B (en) 2023-04-14 2023-04-14 Port cargo flow direction statistical method and system based on BP neural network algorithm

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310395337.6A CN116136989B (en) 2023-04-14 2023-04-14 Port cargo flow direction statistical method and system based on BP neural network algorithm

Publications (2)

Publication Number Publication Date
CN116136989A true CN116136989A (en) 2023-05-19
CN116136989B CN116136989B (en) 2023-06-27

Family

ID=86334601

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310395337.6A Active CN116136989B (en) 2023-04-14 2023-04-14 Port cargo flow direction statistical method and system based on BP neural network algorithm

Country Status (1)

Country Link
CN (1) CN116136989B (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111008870A (en) * 2019-12-09 2020-04-14 金陵科技学院 Regional logistics demand prediction method based on PCA-BP neural network model
CN115049334A (en) * 2022-06-16 2022-09-13 四川省交通勘察设计研究院有限公司 Statistical analysis method, equipment and storage medium for waterway transportation demand
CN115409446A (en) * 2022-08-30 2022-11-29 阿里健康科技(中国)有限公司 Flow direction completion method, device and system, model training method and electronic equipment

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111008870A (en) * 2019-12-09 2020-04-14 金陵科技学院 Regional logistics demand prediction method based on PCA-BP neural network model
CN115049334A (en) * 2022-06-16 2022-09-13 四川省交通勘察设计研究院有限公司 Statistical analysis method, equipment and storage medium for waterway transportation demand
CN115409446A (en) * 2022-08-30 2022-11-29 阿里健康科技(中国)有限公司 Flow direction completion method, device and system, model training method and electronic equipment

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
夏天俊: "基于自适应神经网络的BDI预测研究", 中国优秀硕士学位论文全文数据库, no. 5 *
江志娟;: "基于BP神经网络的港口吞吐量预测模型――钦州港的实证研究", 物流科技, no. 01 *
王晓强,张爱玲,曲永岗: "基于神经网络的集装箱运量预测", 河海大学学报(自然科学版), no. 04 *

Also Published As

Publication number Publication date
CN116136989B (en) 2023-06-27

Similar Documents

Publication Publication Date Title
US11741693B2 (en) System and method for semi-supervised conditional generative modeling using adversarial networks
TWI703503B (en) Risk transaction identification method, device, server and storage medium
AU2021218110B2 (en) Learning from distributed data
CN104699772B (en) A kind of big data file classification method based on cloud computing
CN109658033B (en) Method, system, device and storage medium for calculating similarity of goods source route
CN106296195A (en) A kind of Risk Identification Method and device
CN107563757A (en) The method and device of data risk control
CN113298152B (en) Model training method, device, terminal equipment and computer readable storage medium
CN115658955B (en) Cross-media retrieval and model training method, device, equipment and menu retrieval system
CN110457677A (en) Entity-relationship recognition method and device, storage medium, computer equipment
CN108537270A (en) Image labeling method, terminal device and storage medium based on multi-tag study
CN111415167B (en) Network fraud transaction detection method and device, computer storage medium and terminal
CN110765276A (en) Entity alignment method and device in knowledge graph
CN109977131A (en) A kind of house type matching system
CN113435900A (en) Transaction risk determination method and device and server
CN116703682B (en) Government affair data platform based on deep learning
CN116136989B (en) Port cargo flow direction statistical method and system based on BP neural network algorithm
CN110807159B (en) Data marking method and device, storage medium and electronic equipment
CN111860655B (en) User processing method, device and equipment
CN117216736A (en) Abnormal account identification method, data scheduling platform and graph computing platform
CN111984842B (en) Bank customer data processing method and device
US11042911B1 (en) Creation of high value data assets from undervalued data
CN111611981A (en) Information identification method and device and information identification neural network training method and device
CN111027296A (en) Report generation method and system based on knowledge base
CN116050508B (en) Neural network training method and device

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