CN112990520A - Mesh point connection quantity prediction method and device, computer equipment and storage medium - Google Patents
Mesh point connection quantity prediction method and device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN112990520A CN112990520A CN201911281445.0A CN201911281445A CN112990520A CN 112990520 A CN112990520 A CN 112990520A CN 201911281445 A CN201911281445 A CN 201911281445A CN 112990520 A CN112990520 A CN 112990520A
- Authority
- CN
- China
- Prior art keywords
- data
- sample
- configuration data
- service data
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 67
- 238000012549 training Methods 0.000 claims description 55
- 238000012360 testing method Methods 0.000 claims description 16
- 238000004590 computer program Methods 0.000 claims description 15
- 238000001914 filtration Methods 0.000 claims description 8
- 238000002790 cross-validation Methods 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000007637 random forest analysis Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 5
- 238000003066 decision tree Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- 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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/067—Enterprise or organisation modelling
-
- 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
Abstract
The application relates to a method and a device for predicting the connection quantity of a mesh point, computer equipment and a storage medium. The mesh point connection quantity prediction method comprises the following steps: acquiring service data and configuration data of a target network point in a preset time period; inputting the service data and the configuration data into a connection component prediction model to obtain a connection component prediction range corresponding to the target network point; the interface quantity prediction model comprises a plurality of cascaded predictor models. By adopting the method, the problem of poor data reliability of the predicted connection piece quantity caused by manually predicting the connection piece quantity of the network points according to experience in the traditional technology can be solved. By the adoption of the method, the data reliability of the predicted mesh point connection quantity can be improved, and the accuracy of the output result of the connection model is further improved.
Description
Technical Field
The invention relates to the technical field of logistics, in particular to a method and a device for predicting the quantity of network point connection parts, computer equipment and a storage medium.
Background
In recent years, with the rapid development of internet technology, the logistics industry also enters a rapid development period. The connection refers to the quantity of the parts conveyed between the connection points and the distribution points by the connection personnel, the connection can effectively relieve the receiving and dispatching pressure of the receiving and dispatching personnel, and the receiving and dispatching efficiency is improved.
The existing connection model, such as a connection line model and a connection shift model, needs to use the predicted connection quantity of the network points as input data, so that the connection model can conveniently arrange the connection line, the connection shift and the planning of connection personnel according to the connection quantity, and the like. At present, the amount of the connection piece input into the connection model is predicted by a human according to experience.
However, in the method for manually predicting the amount of the connection piece of the network point according to experience, the predicted data reliability of the amount of the connection piece is poor, so that the accuracy of an output result of the connection model is poor.
Disclosure of Invention
In view of the foregoing, it is desirable to provide a method, an apparatus, a computer device, and a storage medium for predicting a dot patch amount, which can improve data reliability of the predicted dot patch amount.
In a first aspect, an embodiment of the present application provides a mesh point connection quantity prediction method, where the mesh point connection quantity prediction method includes:
acquiring service data and configuration data of a target network point in a preset time period;
inputting the service data and the configuration data into a connection component prediction model to obtain a connection component prediction range corresponding to the target network point; the interface quantity prediction model comprises a plurality of cascaded predictor models.
In one embodiment, the interface quantity prediction model comprises a first predictor model and a second predictor model, the first predictor model comprises a plurality of base classifiers;
the inputting the service data and the configuration data into a connection component prediction model to obtain a connection component prediction range corresponding to the target network point includes:
respectively inputting the service data and the configuration data into each base classifier to obtain a predicted component range output by each base classifier;
and inputting the predicted quantity range output by each base classifier into the second prediction submodel to obtain the connection quantity predicted range corresponding to the target mesh point.
In one embodiment, the second predictor model includes a target classifier, and the step of inputting the predicted quantity range output by each base classifier into the second predictor model to obtain the connection quantity predicted range corresponding to the target mesh point includes:
and inputting the predicted component range output by each base classifier into the target classifier to obtain the connection component prediction range corresponding to the target mesh point.
In one embodiment, the training process of the base classifier includes:
acquiring sample service data and sample configuration data of a plurality of network points in a preset time period;
and training an initial network model according to the sample service data and the sample configuration data to obtain the base classifier.
In one embodiment, the method further comprises:
respectively preprocessing the sample service data and the sample configuration data according to a preset filtering rule to obtain preprocessed sample service data and preprocessed sample configuration data;
the training an initial network model according to the sample service data and the sample configuration data to obtain the base classifier comprises:
and training an initial network model according to the preprocessed sample service data and the preprocessed sample configuration data to obtain the base classifier.
In one embodiment, the training an initial network model according to the preprocessed sample service data and the preprocessed sample configuration data to obtain the base classifier includes:
acquiring a training data set and a testing data set from the preprocessed sample service data and the preprocessed sample configuration data by adopting a cross validation method;
training an initial network model by using the training data set to obtain a trained network model;
and verifying the prediction accuracy of the trained network model by adopting the test data set to obtain the base classifier.
In one embodiment, the method further comprises:
acquiring the priority of each feature data in the sample service data and the sample configuration data according to the sample service data, the sample configuration data and the base classifier;
adjusting the proportion of each feature data in the sample service data and the sample configuration data according to the priority of each feature data;
and adjusting the parameters of the base classifier according to the adjusted sample service data and the adjusted sample configuration data.
In a second aspect, an embodiment of the present application provides a device for predicting a mesh point connection quantity, where the device includes:
the first acquisition module is used for acquiring service data and configuration data of a target network point within a preset time period;
the prediction module is used for inputting the service data and the configuration data into a connection component prediction model to obtain a connection component prediction range corresponding to the target network point; the interface quantity prediction model comprises a plurality of cascaded predictor models.
In a third aspect, an embodiment of the present application provides a computer device, including a memory and a processor, where the memory stores a computer program, and the processor implements the steps of the method according to the first aspect when executing the computer program.
In a fourth aspect, embodiments of the present application provide a computer-readable storage medium, on which a computer program is stored, which, when executed by a processor, implements the steps of the method according to the first aspect.
The beneficial effects brought by the technical scheme provided by the embodiment of the application at least comprise:
acquiring service data and configuration data of a target network point in a preset time period; inputting the service data and the configuration data into a connection component prediction model to obtain a connection component prediction range corresponding to the target network point; the connection quantity prediction model comprises a plurality of cascaded predictor models; therefore, the service data and the configuration data of the target network point are input into the connection quantity prediction model, and the connection quantity prediction range corresponding to the target network point can be obtained; the problem that in the traditional technology, the data reliability of the predicted connection piece quantity is poor due to the fact that the connection piece quantity of the network points is predicted manually according to experience is solved. According to the method and the device, the data reliability of the predicted mesh point connection quantity can be improved, and the accuracy of the output result of the connection model is improved.
Drawings
Fig. 1 is an application environment diagram of a mesh point connection quantity prediction method according to an embodiment;
fig. 2 is a schematic flow chart of a mesh point connection quantity prediction method according to an embodiment;
fig. 3 is a schematic flow chart of a mesh point connection quantity prediction method according to an embodiment;
FIG. 4 is a flowchart of a base classifier training process according to an embodiment;
FIG. 5 is a flowchart of a base classifier training process provided by one embodiment;
FIG. 6 is a flow diagram that illustrates a process for training a base classifier, according to an embodiment;
FIG. 7 is a flowchart of a base classifier training process provided by one embodiment;
fig. 8 is a block diagram of a device for predicting a mesh point connection quality according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
The method for predicting the amount of the network point connection component provided by the application can be applied to the computer device shown in fig. 1, the computer device can be a server, and the internal structure diagram can be shown in fig. 1. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer equipment is used for storing data of the network access piece quantity prediction method.
The mesh point connection quantity prediction method, the mesh point connection quantity prediction device, the computer equipment and the storage medium aim at solving the technical problem that data reliability of predicted mesh point connection quantity is poor due to the fact that connection quantity of mesh points is predicted manually according to experience in the traditional technology. The following describes in detail the technical solutions of the present application and how the technical solutions of the present application solve the above technical problems by embodiments and with reference to the drawings. The following specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments.
It should be noted that, in the method for predicting a mesh point connection component amount provided in the embodiment of the present application, an execution subject may be a mesh point connection component amount prediction device, and the mesh point connection component amount prediction device may be implemented as part or all of a computer device in a software, hardware, or a combination of software and hardware. In the following method embodiments, the execution subject is a computer device as an example.
Referring to fig. 2, which shows a schematic flow chart of a mesh point connection quality prediction method provided in the embodiment of the present application, as shown in fig. 2, the mesh point connection quality prediction method of the present embodiment may include the following steps:
step S100, acquiring service data and configuration data of a target network point in a preset time period.
Based on actual service requirements, if the range of the connection quantity of the target network point needs to be predicted, the computer equipment acquires service data and configuration data of the target network point within a preset time period.
In this embodiment, the preset time period is one day, and the computer device obtains the service data and the configuration data of the target network point in the day corresponding to the specified date. The specified date may be any day before, and in other embodiments, the preset time period may also be other time periods, which are not specifically limited herein.
In this embodiment, the service data includes a target mesh point receiving and dispatching quantity, a target type quantity and a historical connection quantity. The target network point receiving and dispatching quantity comprises total receiving and dispatching quantity, receiving quantity, dispatching quantity and average weight of each target network point in one day; the target type quantity of the target network point comprises a quantity aged to be the next day type and a quantity aged to be the document bill type; the historical connection quantity of the target mesh point comprises a historical average connection quantity, a historical maximum connection quantity and a historical minimum connection quantity, and in the embodiment, the historical connection quantity can be a daily average connection quantity, a daily maximum connection quantity and a daily minimum connection quantity in the previous week.
In this embodiment, the configuration data includes configuration information of a target node access person and configuration information of a target node unit area. The target website lightering personnel configuration information comprises the number of lightering personnel of the target website, the average age of the lightering personnel, the average working age of the lightering personnel and the number of full-time lightering personnel; the target mesh point unit area configuration information comprises the number of unit areas in the target mesh point, the number of C-type unit areas, the number of CBD-type unit areas and the average distance from the target mesh point to each unit area. The type C unit area refers to a residential area, a mixed business area, a hospital, and a campus-type unit area, and the type CBD unit area refers to an office-type unit area.
In other embodiments, the traffic data may also include one or more of a target mesh point pick-and-place quantity, a target type quantity, a historical interface quantity; the configuration data may also include one or more of target site dockee configuration information and target site unit area configuration information, which is not specifically limited herein.
Step S200, inputting the service data and the configuration data into a connection component quantity prediction model, and obtaining a connection component quantity prediction range corresponding to the target network point.
The connection component prediction model comprises a plurality of cascaded predictor models.
And the computer equipment inputs the service data and the configuration data into the connection quantity prediction model to obtain a connection quantity prediction range corresponding to the target network point.
In this embodiment, the connection workload prediction model includes a plurality of cascaded prediction submodels, the computer device inputs the service data and the configuration data into a first layer of prediction submodel, and the first layer of prediction submodel outputs a connection workload prediction range corresponding to a target node; and the computer equipment takes the output of the first-layer predictor model as the input of the second-layer predictor model, inputs the input into the second-layer predictor model, and repeats the process until the last-layer predictor model outputs the connection condition prediction range corresponding to the target network point. For example, the predicted range of the amount of the connection pieces corresponding to the output target mesh point is 0-2000 or 2000-4000 pieces, and so on. Therefore, the connection quantity prediction model is formed through the cascade of the multilayer prediction submodels, and the prediction accuracy of the prediction range of the connection quantity is improved.
In the embodiment, the service data and the configuration data of the target network point in a preset time period are acquired; inputting the service data and the configuration data into a connection condition prediction model formed by cascading a plurality of prediction sub-models, and acquiring a connection condition prediction range corresponding to a target network point; therefore, the service data and the configuration data of the target network point are input into the connection quantity prediction model, and the connection quantity prediction range corresponding to the target network point can be obtained; the problem of poor data reliability of the connection quantity caused by manually predicting the connection quantity of the network points according to experience in the traditional technology is solved. The data reliability of the predicted mesh point connection quantity can be improved, the accuracy of the output result of the connection model is further improved, the existing connection model is analyzed and improved, and the connection efficiency is improved.
Fig. 3 is a schematic flow chart of a mesh point connection quantity prediction method according to another embodiment. On the basis of the embodiment shown in fig. 2, in this embodiment, the connection component prediction model specifically includes a first predictor model and a second predictor model, and the first predictor model includes a plurality of basis classifiers. As shown in fig. 3, in the present embodiment, step S200 includes step S210 and step S220, specifically:
step S210, respectively inputting the service data and the configuration data into each base classifier to obtain the predicted component range output by each base classifier.
In this embodiment, as an implementation manner, the connection quantity prediction model specifically includes two cascaded submodels, that is, a first predictor model and a second predictor model. Wherein the first predictor model comprises a plurality of base classifiers.
And the computer equipment respectively inputs the service data and the configuration data into each base classifier to obtain the predicted component range output by each base classifier.
And S220, inputting the predicted quantity ranges output by the base classifiers into a second prediction submodel to obtain the connection quantity predicted ranges corresponding to the target mesh points.
After the predicted quantity ranges output by the base classifiers are obtained, the computer equipment inputs the predicted quantity ranges output by the base classifiers into a second prediction submodel to obtain a connection quantity predicted range corresponding to a target mesh point output by the second prediction submodel.
In this embodiment, as an implementation manner, the second predictor model includes a target classifier, and the step S220 can be specifically implemented by inputting the predicted quantity range output by each base classifier into the target classifier to obtain the connection quantity predicted range corresponding to the target website.
In other embodiments, as an implementation, the base classifier and the target classifier may each be any one of extreme gradient boost Xgboost, light gradient boost frame LightGBM, or Random Forest. In this embodiment, specifically, the first predictor model includes three base classifiers, which are Xgboost, LightGBM, and Random Forest, respectively; the target classifier is Random Forest.
In the embodiment, the service data and the configuration data of the target network point in a preset time period are acquired; respectively inputting the service data and the configuration data into each base classifier to obtain a predicted component range output by each base classifier; inputting the predicted quantity range output by each base classifier into a second prediction submodel to obtain a connection quantity predicted range corresponding to the target mesh point; therefore, the prediction accuracy of the connection quantity prediction range corresponding to the target mesh point is improved through the cascaded first prediction submodel and the second prediction submodel.
Based on the embodiment shown in fig. 3, referring to fig. 4, fig. 4 is a flowchart illustrating a training process of the base classifier. As shown in fig. 4, the training process of the base classifier of the present embodiment includes steps S310 and S320, specifically:
step S310, sample service data and sample configuration data of a plurality of network points in a preset time period are obtained.
The computer equipment acquires sample service data and sample configuration data of a plurality of network points in a preset time period from the big data platform. The preset time period may be one day, and the computer device obtains the sample service data and the sample configuration data of the plurality of network points in any day before the preset time period to obtain the sample service data and the sample configuration data of each network point.
In this embodiment, the sample traffic data of the mesh point includes the receiving and dispatching quantity of the mesh point, the target type quantity of the mesh point, and the historical connecting quantity of the mesh point. The receiving and dispatching amount of the network point comprises the total receiving and dispatching amount, the receiving amount, the dispatching amount and the average weight of each network point in one day; the target type quantity of the network point comprises the quantity of the network point aged to the next day type and the quantity of the document bill type; the historical connection quantity of the mesh point comprises a historical average connection quantity, a historical maximum connection quantity and a historical minimum connection quantity of the mesh point; in this embodiment, the historical amount of connections may be the average amount of connections per day, the maximum amount of connections per day, and the minimum amount of connections per day for the mesh point during the week prior to the historical amount of connections.
In this embodiment, the sample configuration data of the website includes configuration information of a dockee of the website and configuration information of a unit area of the website. The configuration information of the connecting personnel of the network comprises the number of the connecting personnel, the average age of the connecting personnel, the average working age of the connecting personnel and the number of the full-time connecting personnel of the network; the unit area configuration information of the mesh point includes the number of unit areas in the mesh point, the number of C-type unit areas, the number of CBD-type unit areas, and the average distance from the mesh point to each unit area.
And step S320, training an initial network model according to the sample service data and the sample configuration data to obtain a base classifier.
The computer equipment takes the sample service data and the sample configuration data of each network point as a piece of data, and each piece of data comprises the data characteristics corresponding to the sample service data and the sample configuration data of the network point; and training an initial network model by the computer equipment according to a plurality of pieces of data corresponding to a plurality of network points to obtain a base classifier. In this embodiment, the initial network model may be Xgboost, LightGBM, or Random Forest initialized by parameters.
In this embodiment, if the initial network model is an Xgboost or LightGBM initialized by parameters, the computer device needs to obtain the following input parameters when training the initial network model: the number of decision trees, the regularization coefficient of L1, the regularization coefficient of L2, the learning rate, the depth of the decision trees, the proportion of data features used in training each tree to all data features, and the proportion of the number of data used in training each tree to multiple pieces of data. If the initial network model is a Random Forest initialized by parameters, the depth of the decision tree is required to be obtained as an input parameter when the computer equipment trains the initial network model. In this embodiment, the input parameters acquired by the computer device may be manually input to the computer device according to the service requirement.
Referring to fig. 5, fig. 5 is a schematic flowchart of a training process of a base classifier according to another embodiment. On the basis of the embodiment shown in fig. 4, as shown in fig. 5, the training process of the base classifier of this embodiment further includes step S400:
step S400, respectively preprocessing the sample service data and the sample configuration data according to a preset filtering rule to obtain preprocessed sample service data and preprocessed sample configuration data.
In this embodiment, the filtering rule includes: deleting sample business data and sample configuration data corresponding to a network node with zero number of the connetors, deleting sample business data and sample configuration data corresponding to a network node with missing target data, deleting sample business data and sample configuration data corresponding to a network node with a receiving and dispatching quantity smaller than a first threshold, and deleting sample business data and sample configuration data corresponding to a network node with an average connecting quantity smaller than a second threshold.
And the computer equipment preprocesses the sample service data and the sample configuration data corresponding to each network point according to the filtering rule to obtain preprocessed sample service data and preprocessed sample configuration data.
In other embodiments, the filtering rule may also be to delete one or more of the sample service data and the sample configuration data corresponding to a website with zero number of dockees, delete the sample service data and the sample configuration data corresponding to a website with missing target data, delete the sample service data and the sample configuration data corresponding to a website with a smaller receiving and dispatching quantity than a first threshold, and delete the sample service data and the sample configuration data corresponding to a website with an average dockee quantity smaller than a second threshold, which is not limited herein.
Referring to fig. 5, in the present embodiment, the step S320 includes a step S321:
step S321, training an initial network model according to the preprocessed sample service data and the preprocessed sample configuration data to obtain a base classifier.
Invalid data are removed from the preprocessed sample business data and the preprocessed sample configuration data, and the computer device trains an initial network model according to the preprocessed sample business data and the preprocessed sample configuration data, so that the prediction accuracy of the trained base classifier is improved.
Fig. 6 is a flowchart illustrating a training process of a base classifier according to another embodiment. On the basis of the above-described embodiment shown in fig. 5, the step S321 of the present embodiment includes the step S321a, the step S321b and the step S321c, specifically:
step S321a, a cross validation method is used to obtain a training data set and a testing data set from the preprocessed sample service data and the preprocessed sample configuration data.
And the computer equipment acquires a training data set and a testing data set from the preprocessed sample service data and the preprocessed sample configuration data of each network point.
In this embodiment, the computer device adopts a five-fold cross validation method, divides the preprocessed sample service data and the preprocessed sample configuration data corresponding to the multiple mesh points into five parts, extracts four different parts as a training data set during each iteration, and uses the remaining part as a test data set.
Step S321b, training the initial network model by using the training data set to obtain a trained network model.
The computer device inputs the extracted training data set into the initial network model, and trains the initial network model. In the training process, the computer equipment searches GridSearchCV by adopting grids to find out more optimal model parameters so as to obtain a trained network model.
And step S321c, verifying the prediction accuracy of the trained network model by using the test data set to obtain a base classifier.
After the trained network model is obtained, the computer equipment carries out prediction accuracy verification on the trained network model. Specifically, the computer equipment inputs the test data set into the trained network model to obtain the connection quantity prediction range of each network point in the test data set, compares the connection quantity prediction range of each network point in the test data set with the actual connection quantity range of each network point in the test data set, and takes the trained network model with the prediction accuracy higher than the threshold value as the trained base classifier.
As an implementation mode, the computer device trains to obtain three base classifiers, namely Xgboost, LightGBM and Random Forest, and the three base classifiers are used as a first prediction submodel of the connection quantity prediction model; further, one of the three base classifiers is arbitrarily selected, for example, Random Forest is selected as a target classifier included in the second predictor model, thereby obtaining the connection quality prediction model.
The embodiment acquires sample service data and sample configuration data of a plurality of network points in a preset time period; respectively preprocessing the sample service data and the sample configuration data according to a preset filtering rule to obtain preprocessed sample service data and preprocessed sample configuration data; acquiring a training data set and a testing data set from the preprocessed sample service data and the preprocessed sample configuration data by adopting a cross validation method; training the initial network model by adopting a training data set to obtain a trained network model; verifying the prediction accuracy of the trained network model by adopting a test data set to obtain a base classifier; therefore, the base classifier is obtained by adopting cross validation method training, the model stability and the generalization capability of the base classifier are improved, and the model stability and the generalization capability of the connection component prediction model are further improved.
Fig. 7 is a flowchart illustrating a training process of a base classifier according to another embodiment. On the basis of the embodiment shown in fig. 4, in this embodiment, the training process of the base classifier further includes step S510, step S520, and step S530, specifically:
step S510, according to the sample service data, the sample configuration data, and the base classifier, obtaining priorities of each feature data in the sample service data and the sample configuration data.
And the computer equipment predicts the connection quantity range of each mesh point according to the base classifier to obtain the connection quantity prediction range of each mesh point. The method comprises the steps of inputting a predicted range of a connected part quantity of a network point, sample business data and sample configuration data of the network point into an existing feature importance analysis tool by adopting the existing feature importance analysis tool, obtaining the influence degree of each feature data in the sample business data and the sample configuration data of the network point on the predicted range of the connected part quantity, and setting the priority of the feature data with the larger influence degree to be the highest.
Step S520, adjusting the ratio of each feature data in the sample service data and the sample configuration data according to the priority of each feature data.
And the computer equipment correspondingly reduces the data proportion of the characteristic data with lower priority in the sample service data and the sample configuration data of each network point, namely the characteristic data with smaller influence degree on the predicted range of the receiving quantity.
For example, the adjusted sample business data and the adjusted sample configuration data are obtained by reducing the proportion of the feature data "CBD type unit area number" in the sample configuration data of the mesh point, reducing the proportion of the feature data "full-time number of dockees" in the sample business data of the mesh point, and the like.
Step S530, adjusting parameters of the base classifier according to the adjusted sample service data and the adjusted sample configuration data.
And the computer equipment inputs the adjusted sample service data and the adjusted sample configuration data into the base classifier for training again so as to improve the prediction accuracy of the finally obtained base classifier and optimize the prediction effect of the connection component prediction model.
It should be understood that although the various steps in the flow charts of fig. 2-7 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 2-7 may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performance of the sub-steps or stages is not necessarily sequential, but may be performed in turn or alternating with other steps or at least some of the sub-steps or stages of other steps.
In one embodiment, as shown in fig. 8, there is provided a mesh point connection quantity prediction apparatus, including:
the first obtaining module 10 is configured to obtain service data and configuration data of a target node within a preset time period;
the prediction module 20 is configured to input the service data and the configuration data into a connection quantity prediction model, and obtain a connection quantity prediction range corresponding to the target mesh point; the interface quantity prediction model comprises a plurality of cascaded predictor models.
Optionally, the interface quantity prediction model comprises a first predictor model and a second predictor model, the first predictor model comprises a plurality of base classifiers; the prediction module 20 includes:
the first prediction unit is used for respectively inputting the service data and the configuration data into each base classifier to obtain a prediction component range output by each base classifier;
and the second prediction unit is used for inputting the predicted quantity range output by each base classifier into the second prediction submodel to obtain the connection quantity predicted range corresponding to the target mesh point.
Optionally, the second predictor model includes an object classifier, and the second prediction unit includes:
and the predicting subunit is used for inputting the predicted quantity range output by each base classifier into the target classifier to obtain the connection quantity predicted range corresponding to the target mesh point.
Optionally, the apparatus further comprises:
the second acquisition module is used for acquiring sample service data and sample configuration data of a plurality of network points in a preset time period;
and the training module is used for training an initial network model according to the sample service data and the sample configuration data to obtain the base classifier.
Optionally, the apparatus further comprises:
the preprocessing module is used for respectively preprocessing the sample service data and the sample configuration data according to a preset filtering rule to obtain preprocessed sample service data and preprocessed sample configuration data;
the training module comprises:
and the training unit is used for training an initial network model according to the preprocessed sample service data and the preprocessed sample configuration data to obtain the base classifier.
Optionally, the training unit comprises:
an obtaining subunit, configured to obtain a training data set and a testing data set from the preprocessed sample service data and the preprocessed sample configuration data by using a cross validation method;
the training subunit is used for training an initial network model by adopting the training data set to obtain a trained network model;
and the verification subunit is used for verifying the prediction accuracy of the trained network model by adopting the test data set to obtain the base classifier.
Optionally, the apparatus further comprises:
a third obtaining module, configured to obtain, according to the sample service data, the sample configuration data, and the base classifier, a priority of each feature data in the sample service data and the sample configuration data;
the first adjusting module is used for adjusting the proportion of each feature data in the sample service data and the sample configuration data according to the priority of each feature data;
and the second adjusting module is used for adjusting the parameters of the base classifier according to the adjusted sample service data and the adjusted sample configuration data.
The device for predicting the amount of a mesh point connection component provided in this embodiment may implement the above method for predicting the amount of a mesh point connection component, and the implementation principle and the technical effect are similar, which are not described herein again.
For specific limitations of the mesh point connection quantity prediction device, reference may be made to the above limitations on the mesh point connection quantity prediction method, which is not described herein again. All or part of the modules in the network point connection quantity prediction device can be realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, there is also provided a computer device as shown in fig. 1, which may be a server. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer equipment is used for storing the network point connection piece quantity prediction data. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to realize a network point connection quality prediction method.
Those skilled in the art will appreciate that the architecture shown in fig. 1 is a block diagram of only a portion of the architecture associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, and that a computing device may in particular include more or less components than those shown, or combine certain components, or have a different arrangement of components.
In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having a computer program stored therein, the processor implementing the following steps when executing the computer program:
acquiring service data and configuration data of a target network point in a preset time period; inputting the service data and the configuration data into a connection component prediction model to obtain a connection component prediction range corresponding to the target network point; the interface quantity prediction model comprises a plurality of cascaded predictor models.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware related to instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Ramb microsecond direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and bus dynamic RAM (RDRAM).
In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring service data and configuration data of a target network point in a preset time period; inputting the service data and the configuration data into a connection component prediction model to obtain a connection component prediction range corresponding to the target network point; the interface quantity prediction model comprises a plurality of cascaded predictor models.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above examples only show some embodiments of the present invention, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A method for predicting the amount of a mesh point connection component is characterized by comprising the following steps:
acquiring service data and configuration data of a target network point in a preset time period;
inputting the service data and the configuration data into a connection component prediction model to obtain a connection component prediction range corresponding to the target network point; the interface quantity prediction model comprises a plurality of cascaded predictor models.
2. The method of claim 1, wherein the interface quantity prediction model comprises a first predictor model and a second predictor model, the first predictor model comprising a plurality of basis classifiers;
the inputting the service data and the configuration data into a connection component prediction model to obtain a connection component prediction range corresponding to the target network point includes:
respectively inputting the service data and the configuration data into each base classifier to obtain a predicted component range output by each base classifier;
and inputting the predicted quantity range output by each base classifier into the second prediction submodel to obtain the connection quantity predicted range corresponding to the target mesh point.
3. The method of claim 2, wherein the second predictor model includes a target classifier, and the inputting the predicted quantity range output by each base classifier into the second predictor model to obtain the predicted range of the connected quantity corresponding to the target mesh point includes:
and inputting the predicted component range output by each base classifier into the target classifier to obtain the connection component prediction range corresponding to the target mesh point.
4. The method according to claim 2 or 3, wherein the training process of the base classifier comprises:
acquiring sample service data and sample configuration data of a plurality of network points in a preset time period;
and training an initial network model according to the sample service data and the sample configuration data to obtain the base classifier.
5. The method of claim 4, further comprising:
respectively preprocessing the sample service data and the sample configuration data according to a preset filtering rule to obtain preprocessed sample service data and preprocessed sample configuration data;
the training an initial network model according to the sample service data and the sample configuration data to obtain the base classifier comprises:
and training an initial network model according to the preprocessed sample service data and the preprocessed sample configuration data to obtain the base classifier.
6. The method of claim 5, wherein training an initial network model to obtain the base classifier according to the preprocessed sample traffic data and the preprocessed sample configuration data comprises:
acquiring a training data set and a testing data set from the preprocessed sample service data and the preprocessed sample configuration data by adopting a cross validation method;
training an initial network model by using the training data set to obtain a trained network model;
and verifying the prediction accuracy of the trained network model by adopting the test data set to obtain the base classifier.
7. The method of claim 4, further comprising:
acquiring the priority of each feature data in the sample service data and the sample configuration data according to the sample service data, the sample configuration data and the base classifier;
adjusting the proportion of each feature data in the sample service data and the sample configuration data according to the priority of each feature data;
and adjusting the parameters of the base classifier according to the adjusted sample service data and the adjusted sample configuration data.
8. A device for predicting the amount of a network node connection component, the device comprising:
the first acquisition module is used for acquiring service data and configuration data of a target network point within a preset time period;
the prediction module is used for inputting the service data and the configuration data into a connection component prediction model to obtain a connection component prediction range corresponding to the target network point; the interface quantity prediction model comprises a plurality of cascaded predictor models.
9. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor implements the steps of the method of any one of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911281445.0A CN112990520A (en) | 2019-12-13 | 2019-12-13 | Mesh point connection quantity prediction method and device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911281445.0A CN112990520A (en) | 2019-12-13 | 2019-12-13 | Mesh point connection quantity prediction method and device, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112990520A true CN112990520A (en) | 2021-06-18 |
Family
ID=76332278
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911281445.0A Pending CN112990520A (en) | 2019-12-13 | 2019-12-13 | Mesh point connection quantity prediction method and device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112990520A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107403198A (en) * | 2017-07-31 | 2017-11-28 | 广州探迹科技有限公司 | A kind of official website recognition methods based on cascade classifier |
| WO2019056499A1 (en) * | 2017-09-20 | 2019-03-28 | 平安科技（深圳）有限公司 | Prediction model training method, data monitoring method, apparatuses, device and medium |
| CN109961165A (en) * | 2017-12-25 | 2019-07-02 | 顺丰科技有限公司 | Part amount prediction technique, device, equipment and its storage medium |
| CN110019401A (en) * | 2017-12-25 | 2019-07-16 | 顺丰科技有限公司 | Part amount prediction technique, device, equipment and its storage medium |
| CN110046636A (en) * | 2018-12-11 | 2019-07-23 | 阿里巴巴集团控股有限公司 | Prediction technique of classifying and device, prediction model training method and device |
-
2019
- 2019-12-13 CN CN201911281445.0A patent/CN112990520A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107403198A (en) * | 2017-07-31 | 2017-11-28 | 广州探迹科技有限公司 | A kind of official website recognition methods based on cascade classifier |
| WO2019056499A1 (en) * | 2017-09-20 | 2019-03-28 | 平安科技（深圳）有限公司 | Prediction model training method, data monitoring method, apparatuses, device and medium |
| CN109961165A (en) * | 2017-12-25 | 2019-07-02 | 顺丰科技有限公司 | Part amount prediction technique, device, equipment and its storage medium |
| CN110019401A (en) * | 2017-12-25 | 2019-07-16 | 顺丰科技有限公司 | Part amount prediction technique, device, equipment and its storage medium |
| CN110046636A (en) * | 2018-12-11 | 2019-07-23 | 阿里巴巴集团控股有限公司 | Prediction technique of classifying and device, prediction model training method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109858737B (en) | Grading model adjustment method and device based on model deployment and computer equipment | |
| CN111126668B (en) | Spark operation time prediction method and device based on graph convolution network | |
| CN108647914B (en) | Production scheduling method and device, computer equipment and storage medium | |
| CN111191871A (en) | Project baseline data generation method and device, computer equipment and storage medium | |
| CN112465011A (en) | Project risk prediction method and system based on project research and development process | |
| CN111198907A (en) | Method and device for identifying potential defaulting user, computer equipment and storage medium | |
| CN112241951A (en) | Agricultural monitoring method, system and computer equipment based on raspberry pi and LORA | |
| CN112232594A (en) | Multi-satellite emergency observation task planning method and device and computer equipment | |
| CN115936895A (en) | Risk assessment method, device and equipment based on artificial intelligence and storage medium | |
| CN111242167A (en) | Distributed image annotation method and device, computer equipment and storage medium | |
| CN110765351A (en) | Target user identification method and device, computer equipment and storage medium | |
| CN109800975B (en) | Resource evaluation method and device, computer equipment and storage medium | |
| CN113537591A (en) | Long-term weather prediction method and device, computer equipment and storage medium | |
| CN113240255A (en) | Intelligent building full life cycle management method and system | |
| CN112990520A (en) | Mesh point connection quantity prediction method and device, computer equipment and storage medium | |
| CN115529315B (en) | Cloud edge cooperative system | |
| US20060136328A1 (en) | Method and system for analyzing the risk of a project | |
| CN112465646B (en) | Security data monitoring method and device, computer equipment and storage medium | |
| CN115049121A (en) | Bank customer loss risk prediction model generation method, device, equipment and medium | |
| CN113988804A (en) | Digital management method, device, equipment and storage medium for whole engineering construction process | |
| CN112435021A (en) | Transaction request processing method, device, equipment and medium based on artificial intelligence | |
| CN113191743A (en) | Modular enterprise management method and system based on block chain | |
| CN113051445A (en) | Industrial production data processing method and device, computer equipment and storage medium | |
| CN112488224A (en) | Model training method and device, computer equipment and storage medium | |
| CN112581295A (en) | Product data processing method, device, equipment and medium based on field splitting |
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 |