Network resource pre-coverage method and computer program product

Abstract

The invention provides a network resource pre-coverage method and a computer program product. The method comprises the following steps: acquiring data of a first enterprise whose address is changed within a preset time; and generating a network resource coverage prediction model which can be used for predicting the port variable quantity by using the user data of the first enterprise, the network change index, the port demand quantity, the port margin and the port variable quantity. The network resource coverage prediction model can predict the port variable quantity of the target building when an enterprise migrates into the target building, so that a guidance basis is provided for network resource pre-coverage of the target building. Compared with artificial experience, the method can provide more accurate data support, ensures that back-end resources can meet front-end requirements, and achieves efficient utilization of the resources.

Description

Technical Field

[0001] This disclosure relates to the field of wireless communication technology, and in particular to a network resource pre-coverage method and computer program product. Background Technology

[0002] Against the backdrop of digital transformation, intelligent upgrading, and integrated innovation across all sectors of society, optical communication networks, with their numerous advantages such as high bandwidth, long distance, and low cost, are the preferred solution for high-speed data transmission in fixed locations.

[0003] Currently, when there is a new access demand in the communication network, the network personnel often rely on their experience to determine how to plan network resources. This process often lacks strong data support, has a large error relative to the actual network usage, and cannot accurately determine the pre-coverage needs of the building's network resources. Summary of the Invention

[0004] This disclosure provides a network resource pre-coverage method and a computer program product. It constructs a network resource coverage prediction model using historical data, thereby providing an accurate theoretical basis for the pre-coverage construction of target buildings.

[0005] In one aspect, this embodiment provides a network resource pre-coverage method, including: acquiring user data of a first enterprise, first network resource data of a first building, second network resource data of a second building, and third network resource data of the first enterprise, wherein the first building is the building where the first enterprise was located before changing its address, and the second building is the building where the first enterprise is located after changing its address; and determining a network resource coverage prediction model based on the user data, the first network resource data, the second network resource data, and the third network resource data, so as to use the network resource coverage prediction model to carry out network resource pre-coverage construction for the target building.

[0006] In the embodiments of this disclosure, the first enterprise is configured as an enterprise that has undergone an address change within a preset time period, and the user data includes at least one or more of the enterprise's industry, enterprise size, and broadband tariff.

[0007] In embodiments of this disclosure, the first network resource data includes at least the network access information of the first building.

[0008] In the embodiments of this disclosure, the second network resource data includes at least the network access information of the second building, the port availability of the second building when the first enterprise enters the second building but does not occupy any ports in the second building, and the port change of the second building after the first enterprise enters the second building.

[0009] In embodiments of this disclosure, the third network resource data includes at least the number of ports required by the first enterprise in the first building.

[0010] In the embodiments of this disclosure, obtaining user data of a first enterprise, first network resource data of a first building, second network resource data of a second building, and third network resource data of a first enterprise includes: identifying multiple first enterprises that have undergone address changes within a preset time period; identifying the first and second buildings corresponding to each first enterprise; and obtaining user data and third network resource data of each first enterprise, first network resource data of each first building, and second network resource data of each second building.

[0011] In embodiments of this disclosure, determining a network resource coverage prediction model based on user data, first network resource data, second network resource data, and third network resource data includes: determining a training set, the training set including at least input data and output data; constructing a prediction model; and training the prediction model based on the training set to determine the network resource coverage prediction model.

[0012] In embodiments of this disclosure, a training set is determined, which includes at least input data and output data, including: determining user data, the network change index of a first enterprise, port availability, and port demand as input variables; determining port change as an output variable; and determining the training set from multiple user data, multiple first network resource data, multiple second network resource data, and multiple third network resource data based on the input variables and output variables.

[0013] In embodiments of this disclosure, determining the network change index includes: determining the network access change amount of the first enterprise based on the network access information of the first building corresponding to the first enterprise and the network access information of the second building, wherein the network access information includes at least the total network access of the buildings; determining the network access change amounts of multiple first enterprises; and determining the network change index of the first enterprise based on the minimum value among the network access change amounts of multiple buildings, the maximum value among the network access change amounts of multiple buildings, and the network access change amount of the first enterprise.

[0014] In embodiments of this disclosure, constructing a prediction model includes: determining the structure of the prediction model, a training function, and the number of training iterations. The structure of the prediction model includes at least the number of input layer neurons, the number of output layer neurons, and the number of hidden layer neurons. The number of hidden layer neurons is determined at least based on the number of input layer neurons and the number of output layer neurons.

[0015] In embodiments of this disclosure, the method further includes: determining a test set from multiple user data, multiple first network resource data, multiple second network resource data, and multiple third network resource data based on input variables and output variables; and verifying the network resource coverage prediction model based on the test set.

[0016] In the embodiments of this disclosure, verifying the network resource coverage prediction model based on a test set includes: dividing the test set into multiple test groups based on the difference in the number of ports occupied by the first enterprise before and after the address change; training the network resource coverage prediction model using user data, network change index, port availability, and port demand in the test groups to obtain the predicted port change amount of the second building corresponding to the first enterprise; and verifying the network resource coverage prediction model based on the predicted port change amount and the port change amount of the second building in the test group.

[0017] In embodiments of this disclosure, the method further includes: identifying a second enterprise that has recently moved into the target building; determining whether the second enterprise is a newly registered enterprise based on its data information; in response to the second enterprise being a newly registered enterprise, identifying a third enterprise that is similar to the second enterprise; and training a network resource coverage prediction model based on the third enterprise's user data, network change index, number of port requirements, and port availability of the second building corresponding to the third enterprise, so as to carry out pre-coverage construction of network resources for the target building.

[0018] In embodiments of this disclosure, the method further includes: in response to the fact that the second enterprise is a registered enterprise, determining the user data, network change index, port demand quantity, and port availability of the second building corresponding to the second enterprise; and training a network resource coverage prediction model based on the user data, network change index, port availability, and port demand quantity of the second enterprise to carry out pre-coverage construction of network resources for the target building.

[0019] In another aspect, embodiments of this disclosure provide a computer program product, including a computer program that, when executed by a processor, implements the aforementioned network resource pre-coverage method. Attached Figure Description

[0020] The above and other objects, features, and advantages of this disclosure will become more apparent from the more detailed description of the embodiments thereof in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this disclosure and form part of the specification. They are used together with the embodiments of this disclosure to explain the disclosure and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0021] Figure 1 The schematic diagram illustrates an environmental application according to an embodiment of the present disclosure.

[0022] Figure 2 A flowchart illustrating a network resource pre-coverage method according to an embodiment of the present disclosure is shown.

[0023] Figure 3 A flowchart illustrating another network resource pre-coverage method according to an embodiment of the present disclosure is shown.

[0024] Figure 4 A flowchart illustrating the determination of a network resource coverage prediction model according to an embodiment of the present disclosure is shown.

[0025] Figure 5 The flowchart illustrating the determination of the training set according to an embodiment of the present disclosure is shown schematically.

[0026] Figure 6 A flowchart illustrating the determination of a network change index according to an embodiment of the present disclosure is shown.

[0027] Figure 7 A flowchart illustrating a verification network resource coverage prediction model according to an embodiment of the present disclosure is shown.

[0028] Figure 8 An error curve diagram illustrating the training process of the prediction model according to an embodiment of the present disclosure is shown.

[0029] Figure 9 A flowchart illustrating another network resource pre-coverage method according to an embodiment of the present disclosure is shown.

[0030] Figure 10 A block diagram illustrating a network resource pre-overlay device according to an embodiment of the present disclosure is shown.

[0031] Figure 11 A block diagram illustrating a computer program product according to an embodiment of the present disclosure is shown schematically. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of this disclosure more apparent, exemplary embodiments according to this disclosure will now be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this disclosure, and not all embodiments of this disclosure. It should be understood that this disclosure is not limited to the exemplary embodiments described herein.

[0033] Figure 1 The schematic diagram illustrates an environmental application according to an embodiment of the present disclosure.

[0034] like Figure 1 As shown, it includes a target building 101 and a network resource pre-coverage device 102. The target building 101 may house multiple office companies, and the network resource pre-coverage device 102 can be used to plan network resource pre-coverage.

[0035] In some embodiments, when a new enterprise moves into target building 101, it is necessary to plan the network resource coverage for that enterprise, such as determining the number of ports allocated to the enterprise, including the ports of network equipment such as fiber optic distribution boxes and optical splitters required when the enterprise accesses broadband. In this case, the network resource pre-coverage device 102 can obtain network resource data of target building 101 from the database of target building 101 or other databases, such as port availability, network access status, and other information, as well as relevant information about the newly moved-in enterprise.

[0036] Furthermore, the network resource pre-coverage device 102 can predict the port changes in the target building 101 based on network resource data and relevant enterprise information. Furthermore, based on the port changes in the target building 101 and the enterprise's port requirements, it can determine the number of ports to allocate to the enterprise. Thus, the predictive model of the network resource pre-coverage device 102 can be used to plan the allocation of network resources for the enterprise and the target building 101, eliminating reliance on the experience of network personnel and providing predictive guidance for pre-coverage based on actual data.

[0037] The following is combined Figures 2 to 9 The process of the network resource pre-coverage method is described.

[0038] Figure 2 A flowchart illustrating a network resource pre-coverage method according to an embodiment of the present disclosure is shown.

[0039] like Figure 2 The method shown is applied to a network resource pre-overlay device, including S201 and S202: S201. Obtain the user data of the first enterprise, the first network resource data of the first building, the second network resource data of the second building, and the third network resource data of the first enterprise. The first building is the building where the first enterprise was located before the address was changed, and the second building is the building where the first enterprise is located after the address was changed.

[0040] In embodiments of this disclosure, the first enterprise is configured as an enterprise that has undergone an address change within a preset time period. The preset time period can be a pre-defined time frame, such as five years, six years, or three years. In some embodiments, the first enterprise can be an enterprise that has undergone a network address change within the past five years. For example, if the first enterprise relocates to a new building, the operator needs to redeploy access lines and access point locations (such as the installation location of the optical modem and the switch port) to ensure that the physical lines match the new access points.

[0041] In some embodiments, after the first enterprise changes its address, its previous address and its current address are commercial buildings in different locations. In some embodiments, the first enterprise cannot be an enterprise that owns a standalone office building.

[0042] In some embodiments, the network resource pre-overlay device can retrieve enterprise address data from the network resource database and filter out the first enterprise that has undergone address changes within a preset time period based on the enterprise address data.

[0043] In some embodiments, user data can be used to describe the broadband user profile of the first enterprise, and may include indicators used to characterize the broadband information of the first enterprise, such as bandwidth, etc. Third network resource data can be used to describe the network resource information used by the building where the first enterprise was located before the address change, such as accessed broadband information, occupied fiber distribution boxes, optical splitter information, etc.

[0044] In the embodiments of this disclosure, user data includes at least one or more of the following: enterprise industry, enterprise size, and broadband tariff. Enterprise industry may include: finance, real estate, scientific research and technical services, computer and software, education, trading companies, etc. Enterprise size may include: medium-sized enterprises, small enterprises, and micro-enterprises. Since large enterprises are generally less likely to relocate to commercial buildings, large enterprises are not included in the obtained enterprise size data. Broadband tariff refers to the fee that users need to pay for using broadband network services. Based on the obtained enterprise broadband tariff, enterprises can be classified as high-tariff enterprises, medium-tariff enterprises, and low-tariff enterprises.

[0045] In some embodiments, user data of the first enterprise can be obtained from a database. For example, the user's industry and size can be obtained from a business registration information system such as the National Enterprise Credit Information Publicity System or Tianyancha, and the enterprise's broadband tariffs can be obtained from the database.

[0046] As shown in Table 1 below, Table 1 contains the user data obtained from the first company: Table 1

[0047] In the embodiments of this disclosure, data on enterprise industry, enterprise size, and broadband tariffs from different sources can also be integrated; for example, data on "enterprise industry," "enterprise size," and corresponding "broadband tariffs" (such as broadband packages, bandwidth, usage time, payment records, etc.) scattered across different channels and formats can be integrated into a unified data system.

[0048] Next, data cleaning is performed to remove invalid data and fill in blanks. For example, a company's industry might be labeled as "education," but its broadband installation address is listed as "a factory area." Another example is duplicate records: the same company might be registered multiple times, resulting in inconsistent information. The merged raw data undergoes "quality filtering" to resolve issues such as data errors, missing data, and duplication, ensuring data accuracy and reliability.

[0049] Finally, data augmentation is performed to enhance the richness and diversity of the data. For example, based on the cleaned data, data dimensions can be increased through calculations, derivations, and extensions, making the data more comprehensive and more valuable for analysis. Thus, by fusing, cleaning, and augmenting the original data, it can be transformed into high-quality, high-value, and usable data.

[0050] In embodiments of this disclosure, the first network resource data includes at least the network access information of the first building.

[0051] After a company relocates to a new building, it needs to pre-cover port resources. For example, before the company connects to broadband, it needs to allocate and plan the ports of network resources such as fiber optic boxes and optical splitters that the company will need. In the embodiments of this disclosure, network resource data may include the number of ports of network devices such as fiber optic boxes and optical splitters used by the company when connecting to broadband, and may also include building network access information. It is understood that a port can refer to a physical port (such as a hardware interface) or a virtual port.

[0052] In embodiments of this disclosure, building network access information may include internet leased line information and business broadband information for the building. In some embodiments, the network access information of a first building may include the sum of the internet leased line and business broadband of the first building. For example, if a building has an internet leased line of 10M and a business broadband of 5000M, then the network access information of that building is internet leased line 10M + business broadband 5000M = 5010M.

[0053] In the embodiments of this disclosure, the second network resource data includes at least the network access information of the second building, the port availability of the second building when the first enterprise enters the second building but does not occupy any ports in the second building, and the port change of the second building after the first enterprise enters the second building.

[0054] In some embodiments, the network access information of the second building may include the sum of the second building's internet leased line and business broadband.

[0055] Port availability refers to the number or capacity of unused ports among the various types of ports (such as network, telephone, low-voltage, and power interfaces) deployed within a building. When a company enters a building but does not occupy any ports in that building, the port availability of that building can be obtained by retrieving the port availability of the building before the first company occupied a port from the database. For example, if the port availability of building B before company A entered was 'a', and company A occupied 'b' ports, leaving 'c' ports remaining, then 'a' represents the port availability of building B, and 'c' represents the port availability of companies entering building B after company A.

[0056] After a company moves into a new building, the new building will allocate ports to it based on the company's network needs. The port change refers to the number of ports that change in the building before and after the company moves in. The port change can be obtained by retrieving the number of newly deployed ports for the company in the second building after the first company moves in from the database.

[0057] In embodiments of this disclosure, the third network resource data includes at least the number of ports required by the first enterprise in the first building.

[0058] In some embodiments, the port requirement refers to the total number of physical and logical ports required by an enterprise in its daily operations to meet the needs of various device connections, data transmission, and service operation. The port requirement is the "total number of ports that must be configured or reserved" calculated by the enterprise based on its business scale, number of devices, network architecture, and other factors. It is the core basis for building port pre-coverage and network planning. The port requirement can be obtained by retrieving the number of ports occupied by the enterprise in its building before the address change from a database.

[0059] There is an inseparable relationship between port availability (denoted as A), port demand (denoted as B), and port change (denoted as C). The port demand (B) of an enterprise in its existing building typically serves as a crucial reference for port planning in the new building. For example, if the original demand is 100 ports, the new building might be planned for 100-120 ports. Therefore, the theoretical value of C is approximately equal to the "actual number of newly occupied ports," and is likely close to B. When an enterprise decides to use ports in the new building, the maximum value of C is limited by A. For example, if A ≥ the planned number of ports (refer to B), then C = the actual number used (e.g., 100 ports), and the port availability in the new building becomes AC. If A < the planned number of ports, then port resources need to be expanded, and C = A + the number of newly expanded ports. Therefore, these three factors are inextricably linked and jointly determine the planning, allocation, and adjustment strategies for port resources in the new building.

[0060] In embodiments of this disclosure, such as Figure 3 As shown, S201 includes S301, S302, and S303. Figure 3 A flowchart illustrating another network resource pre-coverage method according to an embodiment of the present disclosure is shown.

[0061] S301. Identify multiple first enterprises that have undergone address changes within a preset time period.

[0062] In the embodiments of this disclosure, some or all of the first enterprises that have undergone network address changes within a preset time period, such as the past five years, can be selected from the network resource database.

[0063] S302. Determine the first building and second building corresponding to each first enterprise.

[0064] In embodiments of this disclosure, the first building where each first enterprise was located before changing its address and the second building where it was located after changing its address are determined from the database.

[0065] S303, Obtain user data and third network resource data for each first enterprise, first network resource data for each first building, and second network resource data for each second building.

[0066] In the embodiments of this disclosure, the network access information of the building is obtained based on the building information, and the network access information of the building before and after the address change of the first enterprise is retrieved from the database, as shown in Table 2 below: Table 2

[0067] It is understandable that a first-level enterprise corresponds to a first-level building and a second-level building in a single address change. Therefore, a first-level enterprise corresponds to one user data and one third-level network resource data, one first-level building, and one second-level building. The buildings A1, A2, A3...Ai in Table 2 above can represent all the buildings involved in the address changes of all first-level enterprises.

[0068] In the embodiments of this disclosure, the collected user data, first network resource data, second network resource data, and third network resource data can also undergo data preprocessing. For example, one-hot encoding can be performed on unordered variables such as enterprise industry and broadband tariffs, as shown in Table 3 below: Table 3

[0069] Ordinal encoding is performed on ordinal variables sorted by certain characteristics, such as enterprise size or broadband tariff, as shown in Table 4 below: Table 4

[0070] S202. Based on user data, first network resource data, second network resource data, and third network resource data, determine a network resource coverage prediction model, and use the network resource coverage prediction model to carry out network resource pre-coverage construction for the target building.

[0071] In embodiments of this disclosure, a network model can be trained based on acquired user data, first network resource data, second network resource data, and third network resource data to determine a network resource coverage prediction model. The network resource coverage prediction model can predict the port information of a target building, and its output prediction data can be used to guide the target building in pre-covering network resources.

[0072] According to embodiments of this disclosure, a network resource coverage prediction model is determined based on the user data of the first enterprise and the network resource data of the first and second buildings corresponding to the first enterprise before and after the address change. This model is then used to plan network resource pre-coverage for the target buildings. This allows for the use of historical data to support network resource pre-coverage, overcoming the limitation that pre-coverage planning relies solely on the experience of network personnel.

[0073] In embodiments of this disclosure, such as Figure 4 As shown, S202 includes S401, S402, and S403. Figure 4 A flowchart illustrating the determination of a network resource coverage prediction model according to an embodiment of the present disclosure is shown.

[0074] S401. Determine the training set, which includes at least input data and output data.

[0075] In the embodiments of this disclosure, a training set is determined by acquiring user data, first network resource data, second network resource data, and third network resource data. That is, input data and output data are determined from the acquired data information, and these input and output data can be input into the prediction model to optimize its parameters, thereby determining the nonlinear relationship between the input and output data.

[0076] In embodiments of this disclosure, such as Figure 5 As shown, S401 includes S501, S502, and S503. Figure 5 The flowchart illustrating the determination of the training set according to an embodiment of the present disclosure is shown schematically.

[0077] S501. Determine the user data, the network change index of the first enterprise, the port availability, and the number of ports required as input variables.

[0078] After a company relocates, some companies will not simply reuse the ports they used in the original building. This means that the network resources a company uses in the new building will differ from those in the original building. This difference is, to some extent, limited by the network access information in the new building. Therefore, embodiments of this disclosure utilize the parameter of network access change to reflect the difference in network access information between the two buildings. The number of ports a company needs and can use in the new building is, to some extent, limited by the network access change. Therefore, embodiments of this disclosure consider the impact of the network access change parameter on network resource pre-coverage.

[0079] In embodiments of this disclosure, the network change index of the first enterprise is configured to be determined based on the parameter of network access change. For example... Figure 6 As shown, network change indices are determined, including S601, S602, and S603. Figure 6 A flowchart illustrating the determination of a network change index according to an embodiment of the present disclosure is shown.

[0080] S601. Based on the network access information of the first building corresponding to the first enterprise and the network access information of the second building, determine the change in network access of the first enterprise. The network access information shall include at least the total network access of the buildings.

[0081] In embodiments of this disclosure, a first building where the first enterprise was located before its address change is determined, and the network access information of that first building is determined. In some embodiments, the network access information includes at least the total network access of the building, including leased internet lines and business broadband. That is, the network access information is the sum of leased internet lines and business broadband. Further, a second building where the first enterprise is located after its address change can also be determined, and the network access information of that second building can be determined.

[0082] In the embodiments of this disclosure, the change in network access for the first enterprise is determined by the difference between the network access information of the first building and the network access information of the second building. That is, the change in network access I = network access information of the second building - network access information of the first building.

[0083] For example, Company D relocates from Building 1 to Building 2. The network access in Building 1, A1, is 10M dedicated internet line + 5000M business broadband = 5010M. The network access in Building 2, A2, is 20M dedicated internet line + 8000M business broadband = 8020M. Therefore, the change in Company D's network access, I, is 8020M - 5010M = 3010M.

[0084] S602. Determine the changes in network access for multiple first enterprises.

[0085] In embodiments of this disclosure, the network access change amount I of multiple first enterprises can be determined using the method described above. In some embodiments, the network access change amount I of all first enterprises can be determined using the method described above.

[0086] S603. Based on the minimum value among the network access changes between multiple buildings, the maximum value among the network access changes between multiple buildings, and the network access change of the first enterprise, determine the network change index of the first enterprise.

[0087] In embodiments of this disclosure, the change in network access between buildings can be the difference in network access information between two buildings. Multiple buildings can be determined by a plurality of first enterprises. Based on these multiple buildings, the change in network access between each pair of buildings is determined, i.e., the change in network access between multiple buildings is determined. For example, buildings A1, A2, A3, ..., Ai can be determined by the plurality of first enterprises. Further, the change in network access between each pair of buildings is determined, i.e., E1-E2, E1-E3, ..., E1-E i E2-E3, ..., E2-E i ... E i-1 -E i E1, E2, E3, etc., represent the network access status of the building.

[0088] In embodiments of this disclosure, a minimum and a maximum value are determined from the determined network access changes among multiple buildings. Further, the network change index of the first enterprise is calculated by subtracting the minimum value among the network access changes among multiple buildings from the network access change of the first enterprise, and then dividing by the maximum value among the network access changes among multiple buildings minus the minimum value among the network access changes among multiple buildings.

[0089] For example, the network access change I of the first enterprise D is 3010M, and the minimum value among the network access changes between multiple buildings is MIN[|E1-E2|,…,|E1-E2|]. i |,…,|E i-1 -E i |]=155M, the maximum value is MAX[|E1-E2|,…,|E1-E i |,…,|E i-1 -E i |]=8645M. Then the network change index P of the first enterprise is P=(3010M-155M) / (8645M-3010M)=0.51.

[0090] In the embodiments of this disclosure, the network change index P of multiple first enterprises can be determined by the method described above.

[0091] S502. Determine the port change as the output variable.

[0092] In the embodiments of this disclosure, after the first enterprise changes its address, the number of ports that need to be added or allocated for the target building needs to be determined based on various factors. Therefore, the port change amount of the second building in the acquired data can provide guidance for pre-coverage of the target building to a certain extent. Therefore, when determining the prediction model, the port change amount can be used as an input variable to guide the target building in network resource pre-coverage.

[0093] In some embodiments, grey relational analysis can be used to screen out the main influencing factors affecting port changes. In some embodiments, four main influencing factors are selected as input variables for the prediction model: user data (user industry, size, broadband tariff), network change index P, port availability, and port demand; the target value is the pre-coverage network demand of the target building, i.e., the port change of the target building, which is used as the output of the prediction model.

[0094] S503. Based on input and output variables, determine the training set from multiple user data, multiple first network resource data, multiple second network resource data, and multiple third network resource data.

[0095] In embodiments of this disclosure, input and output data are determined from the acquired data using defined input and output variables to form a training set. For example, the training set is shown in Table 5 below: Table 5

[0096] According to embodiments of this disclosure, by setting network change index, port availability, and port demand as input variables, the changes in port availability and port quantity in buildings due to enterprise relocation are fully considered. Simultaneously, for the enterprise side, the port demand of the enterprise and the network access situation of the building before and after its relocation are fully considered. Therefore, when training the prediction model, multiple aspects such as the port demand and changes of buildings and enterprises can be fully considered, thereby obtaining a more accurate network resource coverage prediction model.

[0097] return Figure 4 S402, Construct a prediction model.

[0098] In embodiments of this disclosure, a neural network prediction model can be constructed, such as an ELman neural network prediction model. The ELman neural network prediction model is then trained using a training set to generate a network resource coverage prediction model.

[0099] In embodiments of this disclosure, S402 includes: determining the structure of the prediction model, wherein the structure of the prediction model includes at least the number of input layer neurons, the number of output layer neurons, and the number of hidden layer neurons; the number of hidden layer neurons is determined at least based on the number of input layer neurons and the number of output layer neurons.

[0100] In the embodiments of this disclosure, based on the input variables, the input layer neurons of the ELman neuron network model are determined to be 4, the output layer neurons are 1, and the ELman neuron network model is set to adopt a single hidden layer structure.

[0101] In some embodiments, the possible values ​​of the number of hidden layer neurons can be determined according to formula (1). For example, the range of the number of hidden layer neurons determined based on formula (1) is 3 to 7. Further, an Elman neural network model is constructed for each number of hidden layer neurons. Then, each Elman neural network model is trained using the same input and output data, enabling the model to learn the nonlinear relationship between the input and output data, where the input and output data are data from the training samples in the training set. After training, the mean squared error of the training results is used as the evaluation metric, and the number of hidden layer neurons corresponding to the model with the smallest mean squared error is selected as the final configuration.

[0102] (1) in, R 1 This represents the number of neurons in the input layer. S 2 This represents the number of neurons in the output layer. S 1 This represents the number of neurons in the hidden layer. m It is an adjustment coefficient, with a value ranging from 1 to 10.

[0103] In embodiments of this disclosure, S402 further includes: determining a training function and the number of training iterations.

[0104] In some embodiments, the transfer functions of the output layer and hidden layers can be set to the logarithmic sigmoid function. Simultaneously, the `traindgx` function is used to train the Elman neural network model to adjust its weights and bias parameters. A preset threshold is also set, meaning that training terminates after n training iterations when the mean squared error of the training set reaches the preset threshold.

[0105] S403. Train a prediction model based on the training set to determine the network resource coverage prediction model.

[0106] In the embodiments of this disclosure, since excessively large or small data are prone to falling into the saturation region of the Elman neuron network model's sigmoid activation function, the training speed of the neural network slows down. Therefore, the continuous variable data of the input and output data can be normalized. Among the aforementioned input and output variables, the continuous variable data includes port capacity, port demand quantity, and port change.

[0107] Furthermore, based on the determined structure, training function, and number of training iterations of the Elman neural network model, the normalized data and the data encoded as described above (one-hot encoding, ordinal encoding) are input into the Elman neural network model for training to initialize the parameters of the Elman neural network model. This generates a network resource coverage prediction model.

[0108] In the embodiments of this disclosure, the neural network is optimized by continuously learning and adjusting the weights of each layer, minimizing the sum of squared errors between the actual output and the expected output, thereby determining the nonlinear relationship between the input and output data. In some embodiments, the training set is divided into training samples and test samples, and operations and error indices are adjusted for different samples to achieve error testing of the model.

[0109] In embodiments of this disclosure, after determining the network resource coverage prediction model, the method may further include: determining a test set from multiple user data, multiple first network resource data, multiple second network resource data, and multiple third network resource data based on input variables and output variables; and verifying the network resource coverage prediction model based on the test set.

[0110] In the embodiments of this disclosure, the acquired data can be divided into a training set and a test set. The prediction model is trained using the training set to generate a network resource coverage prediction model. Furthermore, the trained network resource coverage prediction model is validated using the test set to verify its effectiveness.

[0111] According to embodiments of this disclosure, by dividing historical data into training and testing sets, and performing operations and error metric adjustments on different datasets, error verification of the prediction model can be achieved, thereby ensuring the accuracy of the prediction model.

[0112] Figure 7 A flowchart illustrating a verification network resource coverage prediction model according to an embodiment of the present disclosure is shown.

[0113] In embodiments of this disclosure, such as Figure 7 As shown, the network resource coverage prediction model is validated based on the test set, including S701, S702, and S703: S701. Based on the difference in the number of ports occupied by the first enterprise in the building before and after the address change, the test set is divided into multiple test groups.

[0114] In the embodiments of this disclosure, the number of ports B1 occupied by the first enterprise in the first building and the number of ports B2 occupied by the first enterprise in the second building can be obtained. The difference in the number of ports occupied by the first enterprise before and after the address change, |B2-B1|, can be determined. The difference in the number of ports can, to a certain extent, have a similar trend to the change in the number of ports in the target building: a large difference in the number of ports indicates a large change in the number of ports, and a small difference in the number of ports indicates a small change in the number of ports.

[0115] In the embodiments of this disclosure, multiple test groups are divided based on the port number difference. For example, the relevant data of the first enterprise corresponding to the port number difference greater than or equal to M is divided into test group 1, and the relevant data of the first enterprise corresponding to the port number difference less than or equal to N is divided into test group 2. Where M is greater than N. It can be understood that the test set can also be divided into multiple test groups based on port number difference X, port number difference Y, port number difference Z, etc., so that the effectiveness of the network resource coverage prediction model can be verified using multiple test groups, thereby improving the accuracy of the network resource coverage prediction model verification.

[0116] For example, as shown in Tables 6 and 7 below, Tables 6 and 7 schematically illustrate the division of test groups based on the difference in the number of ports. Table 6 shows the relevant data for test group 1, i.e., the difference in the number of ports is greater than or equal to M, where M can be 5. Table 7 shows the relevant data for test group 2, i.e., the difference in the number of ports is less than or equal to N, where N can be 2. Therefore, it can be seen that the test set can be divided into multiple test groups based on the difference in the number of ports.

[0117] Table 6

[0118] Table 7

[0119] As shown in Table 6 above, Company 1 has less port availability than Company 2, but the port change for Company 1 is greater than that for Company 2. This is to ensure that Company 1 has sufficient port availability. By determining that the difference between the number of ports occupied by the company before and after the address change and the difference between the port change are within the prediction deviation range, the effectiveness of the port demand prediction model can be verified.

[0120] S702. Using user data, network change index, port availability, and port demand in the test group, train a network resource coverage prediction model to obtain the predicted port change amount for the second building corresponding to the first enterprise.

[0121] In the embodiments of this disclosure, multiple test groups can be input into the network resource coverage prediction model to obtain the predicted port change amount of the second building. For example, for the first enterprise with enterprise code 1 in Table 6, the enterprise's user data, network change index, port demand quantity, and corresponding port availability of the second building can be input into the network resource coverage prediction model, and the network resource coverage prediction model outputs the port change amount.

[0122] S703. Based on the predicted port change and the port change in the second building in the test group, verify the network resource coverage prediction model.

[0123] In the embodiments of this disclosure, a network resource coverage prediction model outputs predicted port changes for multiple test groups. Further, the predicted port changes determine the number of ports occupied by the first enterprise in the second building, and the port number difference between the number of ports occupied by the first enterprise in the second building and the number of ports occupied in the first building. Then, based on the relationship between the determined port number difference and the port number difference used when dividing the test groups, the effectiveness of the test groups is verified.

[0124] For example, data from test group 1 is input into the network resource coverage prediction model to predict the port changes in the second building. When the number of predicted port changes greater than or equal to M in test group 1 reaches a preset proportion, the validity of test group 1 is verified. As another example, data from test group 2 is input into the network resource coverage prediction model to predict the port changes in the second building. When the number of predicted port changes less than or equal to N in test group 2 reaches a preset proportion, the validity of test group 2 is verified. It is understood that the preset proportion can be set based on different conditions and is not limited here.

[0125] It should be noted that the change in port count shows a similar trend to the difference in port count to some extent. Therefore, the predicted change in port count can be compared with the difference in port count to determine whether the test group is effective.

[0126] Furthermore, in the embodiments of this disclosure, if the test group is valid, the predicted port change of the second building and the actual port change of the second building are compared to verify whether the network resource coverage prediction model is effective.

[0127] For example, determine the difference between the predicted port change in test group 1 and the actual port change in the building after the first company moved in. When the difference is within the prediction deviation range, verify the effectiveness of the network resource coverage prediction model.

[0128] It is understandable that since Test Group 1 includes data from multiple companies, the difference between the predicted port change and the actual predicted port change for each company in Test Group 1 can be determined. Then, it is determined whether each difference is within the prediction deviation range. When the number of differences within the prediction deviation range accounts for a preset proportion of all differences, the network resource coverage prediction model can be considered effective. This preset proportion can be pre-set, for example, 95%.

[0129] In the embodiments of this disclosure, the applicability of the network resource coverage prediction model can also be determined based on test groups. For example, in the above verification process, if test group 1 verifies the network resource coverage prediction model as effective, and test group 2 verifies it as ineffective, then the network resource coverage prediction model is applicable to data with a port number difference greater than or equal to M, but not applicable to data with a port number difference less than or equal to N. As another example, if both test group 1 and test group 2 verify the network resource coverage prediction model as effective, then the model is applicable regardless of whether the port number difference changes significantly (e.g., a port number difference greater than or equal to M) or slightly (e.g., a port number difference less than or equal to N).

[0130] According to embodiments of this disclosure, by dividing the test set into different test groups, the network resource coverage prediction model can be validated under different changing conditions. Specifically, it can test whether the network resource coverage prediction model can cover situations where the number of ports occupied by the enterprise changes significantly or only slightly before and after the address change. By using more granular test groups to validate the network resource coverage prediction model, its applicability can be ensured through more detailed verification.

[0131] Figure 8 An error curve diagram illustrating the training process of the prediction model according to an embodiment of the present disclosure is shown.

[0132] like Figure 8 As shown, the mean squared error (MSE) of the training set error bar, validation set error bar, and test set error bar is displayed on a logarithmic scale as a function of training epochs. The horizontal axis (Epoch) represents the number of training epochs, which is the number of times the model has fully learned the training data (a total of 10 epochs). The vertical axis (Mean Squared Error, MSE) measures the deviation between the predicted value and the true value (the smaller the value, the more accurate the model prediction).

[0133] exist Figure 8 In this context, the training set error bar represents the model's error on the training data. From... Figure 8As can be seen, the error decreases rapidly as training progresses (especially in later stages), indicating that the model gradually "memorizes" the patterns in the training data. The validation set error bar evaluates the model using validation data that was not used in training. It fluctuates in the early stages but decreases rapidly after the 9th round, reflecting the change in the model's generalization ability. The test set error bar represents the error of the model on independent test data in the final evaluation, and it remains relatively stable throughout, reflecting the model's actual predictive performance. The optimal error bar marks the point where the validation set error is minimized (9th round, MSE=0.20824), representing the moment when the model has the "best generalization ability".

[0134] from Figure 8 As can be seen from the data, the optimal training round is the 9th round, when the validation set error (0.20824) reaches its minimum, and the model's generalization ability is optimal at this time.

[0135] In embodiments of this disclosure, Figure 9 A flowchart illustrating another network resource pre-coverage method according to an embodiment of the present disclosure is shown.

[0136] like Figure 9 As shown, the network resource pre-coverage method provided in this disclosure includes: S901. Obtain the user data of the first enterprise, the first network resource data of the first building, the second network resource data of the second building, and the third network resource data of the first enterprise. The first building is the building where the first enterprise was located before the address was changed, and the second building is the building where the first enterprise is located after the address was changed.

[0137] S902. Based on user data, first network resource data, second network resource data, and third network resource data, determine a network resource coverage prediction model, and use the network resource coverage prediction model to carry out network resource pre-coverage construction for the target building.

[0138] S903, Identify the second company that will relocate to the target building.

[0139] In the embodiments disclosed herein, for a newly registered second enterprise that moves into the target building, data information of the newly registered second enterprise can be obtained from the National Market Entity Registration Network or a third-party enterprise query tool. The data information includes enterprise industry, enterprise size, and other enterprise information.

[0140] S904. Based on the data information of the second enterprise, determine whether the second enterprise is a newly registered enterprise.

[0141] S905a. In response to the fact that the second company is a newly registered company, a third company that is similar to the second company is identified.

[0142] In the embodiments of this disclosure, if the second enterprise is determined to be a newly registered enterprise, at least one enterprise similar to the second enterprise can be matched from an existing broadband user profiling system. Similarity to the second enterprise can be in terms of industry, size, etc. Further, among the matched enterprises, the enterprise with the largest port usage is selected as the third enterprise. By matching the second enterprise with a third enterprise whose basic information is similar, the port size and broadband information of the third enterprise can be used to characterize the port size requirements of the second enterprise.

[0143] S906a: Based on user data, network change index, port demand, and port availability from third-party enterprises, a network resource coverage prediction model is trained to pre-construct network resource coverage for target buildings.

[0144] In the embodiments of this disclosure, user data and network change index of a third enterprise, as well as the port demand of the first building corresponding to the third enterprise and the port availability of the second building corresponding to the third enterprise, are obtained from the database. The user data, network change index, port demand, and port availability are input into the network resource coverage prediction model to obtain the port change of the target building, and port pre-coverage is performed on the target building based on the predicted port change.

[0145] S905b, in response to the second enterprise being a registered enterprise, determine the second enterprise's user data, network change index, port demand quantity, and port availability of the second building corresponding to the second enterprise.

[0146] In the embodiments of this disclosure, when the second enterprise is a registered enterprise, user data and network change index of the second enterprise, as well as the port demand of the first building corresponding to the second enterprise and the port availability of the second building corresponding to the third enterprise, can be obtained from the database. The user data, network change index, port demand, and port availability are input into the network resource coverage prediction model to obtain the port change of the target building, and port pre-coverage is performed on the target building based on the predicted port change.

[0147] S906b: Based on the user data, network change index, port availability, and port demand of the second enterprise, a network resource coverage prediction model is trained to carry out pre-coverage construction of network resources for the target building.

[0148] In the embodiments of this disclosure, user data, network change index, port demand quantity, and port availability are input into the network resource coverage prediction model to obtain the port change amount of the target building, and port pre-coverage is performed on the target building based on the predicted port change amount.

[0149] Figure 10 A block diagram illustrating a network resource pre-overlay device according to an embodiment of the present disclosure is shown.

[0150] like Figure 10 As shown, the network resource pre-coverage device 1000 of this embodiment includes a memory 1001 and a processor 1002.

[0151] The memory 1001 is used to store computer-readable instructions. The processor 1002 is used to execute the computer-readable instructions, causing the network resource pre-overlay device 1000 to perform the network resource pre-overlay method.

[0152] Figure 11 A block diagram illustrating a computer program product according to an embodiment of the present disclosure is shown schematically.

[0153] like Figure 11 As shown, a computer program product 1100 according to an embodiment of this disclosure includes a computer program 1101, which, when executed by a processor, implements the network resource pre-coverage method as described above.

[0154] The above description, with reference to the accompanying drawings, illustrates a network resource pre-coverage method and computer program product according to embodiments of the present disclosure. The method involves acquiring data on a first enterprise that has undergone address changes within a preset time period. Utilizing the first enterprise's user data, network change index, port demand, port availability, and port change amount, a network resource coverage prediction model is generated to predict port change amounts. This network resource coverage prediction model can predict port change amounts in a target building when an enterprise relocates there, thus providing guidance for network resource pre-coverage of the target building. Compared to manual experience, it provides more accurate data support, ensuring that backend resources meet frontend needs and achieving efficient resource utilization.

[0155] The basic principles of this disclosure have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this disclosure are merely examples and not limitations, and should not be considered as essential features of each embodiment of this disclosure. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the scope of this disclosure to the necessity of employing the aforementioned specific details for implementation.

[0156] The block diagrams of devices, apparatuses, devices, and systems disclosed herein are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.

[0157] Additionally, as used herein, the "or" used in a list of items beginning with "at least one" indicates a separate list, such that a list of, for example, "at least one of A, B, or C" means A or B or C, or AB or AC or BC, or ABC (i.e., A and B and C). Furthermore, the word "exemplary" does not imply that the described example is preferred or better than other examples.

[0158] It should also be noted that in the systems and methods of this disclosure, the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered as equivalent solutions to this disclosure.

[0159] Various changes, substitutions, and modifications can be made to the technology described herein without departing from the teachings defined by the appended claims. Furthermore, the scope of the claims of this disclosure is not limited to the specific aspects of the processes, machines, manufactures, events, means, methods, and actions described above. Currently existing or later-developed processes, machines, manufactures, events, means, methods, or actions that perform substantially the same function or achieve substantially the same result as the corresponding aspects described above can be utilized. Therefore, the appended claims include such processes, machines, manufactures, events, means, methods, or actions within their scope.

[0160] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of this disclosure. Therefore, this disclosure is not intended to be limited to the aspects shown herein, but rather to be carried out within the widest scope consistent with the principles and novel features disclosed herein.

[0161] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this disclosure to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations therein.

Claims

1. A method for pre-covering network resources, characterized in that, include: The system acquires user data of the first enterprise, first network resource data of the first building, second network resource data of the second building, and third network resource data of the first enterprise. The first building is the building where the first enterprise was located before changing its address, and the second building is the building where the first enterprise is located after changing its address. Based on the user data, the first network resource data, the second network resource data, and the third network resource data, a network resource coverage prediction model is determined, and the network resource coverage prediction model is used to carry out pre-coverage construction of network resources for the target building.

2. The network resource pre-coverage method according to claim 1, characterized in that, The first enterprise is configured as an enterprise that has undergone an address change within a preset time period, and the user data includes at least one or more of the following: enterprise industry, enterprise size, and broadband tariff.

3. The network resource pre-coverage method according to claim 2, characterized in that, The first network resource data includes at least the network access information of the first building.

4. The network resource pre-coverage method according to claim 3, characterized in that, The second network resource data includes at least the network access information of the second building, the port availability of the second building when the first enterprise enters the second building but does not occupy a port in the second building, and the port change of the second building after the first enterprise enters the second building.

5. The network resource pre-coverage method according to claim 4, characterized in that, The third network resource data includes at least the number of ports required by the first enterprise in the first building.

6. The network resource pre-coverage method according to claim 5, characterized in that, The acquisition of user data of the first enterprise, first network resource data of the first building, second network resource data of the second building, and third network resource data of the first enterprise includes: Identify multiple first enterprises that have undergone address changes within the preset time period; Determine the first building and the second building corresponding to each of the first enterprises; and Acquire the user data and third network resource data of each of the first enterprises, the first network resource data of each of the first buildings, and the second network resource data of each of the second buildings.

7. The network resource pre-coverage method according to claim 6, characterized in that, The step of determining a network resource coverage prediction model based on the user data, the first network resource data, the second network resource data, and the third network resource data includes: Determine a training set, which includes at least input data and output data; Building predictive models; and The prediction model is trained based on the training set to determine the network resource coverage prediction model.

8. The network resource pre-coverage method according to claim 7, characterized in that, The determined training set, which includes at least input data and output data, includes: The user data, the network change index of the first enterprise, the port availability, and the number of ports required are determined as input variables; The port change is determined as an output variable; and The training set is determined from multiple user data sets, multiple first network resource data sets, multiple second network resource data sets, and multiple third network resource data sets based on input and output variables.

9. The network resource pre-coverage method according to claim 8, characterized in that, Determining the network change index includes: Based on the network access information of the first building corresponding to the first enterprise and the network access information of the second building, the change in network access of the first enterprise is determined, wherein the network access information includes at least the total network access of the buildings; Determine the network access changes of multiple first enterprises; and The network change index of the first enterprise is determined based on the minimum value of the network access change among multiple buildings, the maximum value of the network access change among multiple buildings, and the network access change of the first enterprise.

10. The network resource pre-coverage method according to claim 7, characterized in that, The construction of the prediction model includes: determining the structure, training function, and number of training iterations of the prediction model, wherein the structure of the prediction model includes at least the number of input layer neurons, the number of output layer neurons, and the number of hidden layer neurons; the number of hidden layer neurons is determined at least based on the number of input layer neurons and the number of output layer neurons.

11. The network resource pre-coverage method according to claim 8, characterized in that, Also includes: Based on the input variables and the output variables, a test set is determined from multiple user data, multiple first network resource data, multiple second network resource data, and multiple third network resource data; as well as The network resource coverage prediction model is validated based on the test set.

12. The network resource pre-coverage method according to claim 11, characterized in that, The verification of the network resource coverage prediction model based on the test set includes: Based on the difference in the number of ports in the building occupied by the first enterprise before and after the address change, the test set is divided into multiple test groups; Using the user data, network change index, port availability, and port demand from the test group, the network resource coverage prediction model is trained to obtain the predicted port change amount for the second building corresponding to the first enterprise; and The network resource coverage prediction model is verified based on the predicted port change and the port change of the second building in the test group.

13. The network resource pre-coverage method according to claim 9, characterized in that, Also includes: Identify the second company that has recently moved into the target building; Based on the data information of the second company, determine whether the second company is a newly registered company; In response to the fact that the second enterprise is the newly registered enterprise, a third enterprise that is similar to the second enterprise is identified; as well as Based on the user data of the third enterprise, the network change index, the number of ports required, and the port availability of the second building corresponding to the third enterprise, the network resource coverage prediction model is trained to carry out pre-coverage construction of network resources for the target building.

14. The network resource pre-coverage method according to claim 13, characterized in that, Also includes: In response to the fact that the second enterprise is a registered enterprise, the user data of the second enterprise, the network change index, the number of ports required, and the port availability of the second building corresponding to the second enterprise are determined; as well as Based on the user data of the second enterprise, the network change index, the port availability, and the number of ports required, the network resource coverage prediction model is trained to carry out pre-coverage construction of network resources for the target building.

15. A computer program product, comprising a computer program, characterized in that, When a computer program is executed by a processor, it implements the network resource pre-coverage method as described in any one of claims 1 to 14.

Images