Cell value prediction method and device, electronic equipment and storage medium

Abstract

The application provides a cell value prediction method and device, electronic equipment and storage medium. The method comprises: acquiring input feature indexes of a base station cell at any time period of the day, wherein the input feature indexes comprise communication state information, value labels and basic information of the base station cell, and time features of the time period; taking the input feature indexes of the base station cell as input data of a value prediction model, and obtaining a value prediction result of the base station cell output by the value prediction model; wherein the value prediction model is established based on historical feature indexes of each base station cell through machine learning. Through the above scheme, an accurate value prediction model can be established based on a large number of historical feature indexes through machine learning, so as to realize accurate prediction of cell value by using the value prediction model.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of communication networks, and particularly relates to a cell value prediction method and device, electronic equipment and storage medium. BACKGROUND

[0002] At present, mobile communication network technology is developing rapidly. With the advent of the 5G era, multi-standard networks coexist, and cloud-based network architecture is gradually developing. Various new services are emerging, and device connections are massive, which are constantly driving the explosive growth of mobile communication network data traffic.

[0003] However, the rapid development of mobile communication network technology is accompanied by rapid growth in the number of base stations and energy consumption. Traditional base station energy saving schemes are mainly based on artificial experience. By analyzing relevant historical data, such as uplink and downlink traffic, online user number, etc., relevant threshold values of base station cell energy saving are given to implement energy saving. In the face of the development trend of communication network complexity and business scene diversification, it is difficult to achieve precise energy saving while ensuring user business performance and not affecting user perception. SUMMARY

[0004] The present application provides a cell value prediction method and device, electronic equipment and storage medium to achieve precise prediction of cell value.

[0005] In a first aspect, the present application provides a cell value prediction method, comprising: obtaining input feature indicators of a base station cell in any time period of the day, the input feature indicators including communication state information, value labels and basic information of the base station cell, and time characteristics of the time period; taking the input feature indicators of the base station cell as input data of a value prediction model to obtain a value prediction result of the base station cell output by the value prediction model; wherein the value prediction model is established based on historical feature indicators of each base station cell through machine learning; the historical feature indicators include communication state, historical value labels and basic information of the base station cell in a historical first time period, and time characteristics of the predetermined time period.

[0006] In a possible implementation, the method further includes: obtaining historical feature indexes of the base station cell in a first time period of a historical date; selecting, from the historical feature indexes, historical feature indexes with a date before a preset first threshold as a training set, selecting a predetermined number of historical feature indexes from historical feature indexes other than the training set as a validation set, and selecting the remaining historical feature indexes as a test set; obtaining an initial value prediction model by using a machine learning algorithm based on the historical feature indexes in the training set; performing model adjustment on the initial value prediction model based on the historical feature indexes in the validation set; and testing the current value prediction model according to the historical feature indexes in the test set, and completing establishment of the value prediction model if the test is passed.

[0007] In a possible implementation, the method further includes: fitting each communication state information of the base station cell by using an exponential distribution, obtaining an exponential probability density function of the each communication state information, and performing integral processing on the exponential probability density function to obtain a cumulative probability distribution of the each communication state information; and calculating the value degree of the base station cell based on a value degree formula, where the value degree formula includes:

[0008]

[0009] where n is the number of the first communication state information other than the measurement report coverage rate, Wi is the weight of the i th first communication state information, F i (x) is the cumulative probability value of the i th first communication state information when the value of the i th first communication state information is x, W mr is the weight of the measurement report coverage rate, and M is the measurement report coverage rate. If the value degree of the base station cell is less than a preset second threshold, the base station cell is added with a low-value cell label; and if the value degree of the base station cell is not less than the second threshold, the base station cell is added with a high-value label.

[0010] In a possible implementation, the basic information includes the location, type, and identification of the base station cell, and the communication state information includes at least one of the following: uplink network resource utilization, downlink network resource utilization, uplink traffic, downlink traffic, user number, VIP user number, hourly tariff, and measurement report coverage rate.

[0011] In a possible implementation, after the input feature indicators of the base station cell in any time period of the day are acquired, the method further includes: if the base station cell belongs to a predetermined specific cell set, taking the input feature indicators of the base station cell as input data of a first value prediction model corresponding to the base station cell, and obtaining a value prediction result of the base station cell output by the first value prediction model; wherein the first value prediction model is obtained by machine learning based on historical feature indicators of the base station cell and corresponding known results; and wherein each base station cell in the specific cell set corresponds to a pre-established value prediction model.

[0012] In a possible implementation, the method further includes: if the value prediction result of the base station cell in any time period in the future is a low-value cell, executing a base station shutdown instruction on the base station cell in the predetermined time period thereafter; and if the value prediction result of the base station cell in any time period in the future is a high-value cell, not executing processing on the base station cell.

[0013] In a second aspect, the present application provides a cell value prediction device, including: an acquisition module, configured to acquire input feature indicators of a base station cell in any time period of the day, the input feature indicators including communication state information, value labels and basic information of the base station cell, and time features of the time period; and a processing module, configured to take the input feature indicators of the base station cell as input data of a value prediction model, and obtain a value prediction result of the base station cell output by the value prediction model; wherein the value prediction model is established by machine learning based on historical feature indicators of each base station cell; and the historical feature indicators include communication states, historical value labels and basic information of a base station cell in a historical first time period, and time features of the predetermined time period.

[0014] In a possible implementation, the device further includes a modeling module; the acquisition module is further configured to acquire historical feature indicators of a base station cell in a first time period of a historical date; the acquisition module is further configured to select historical feature indicators with dates before a preset first threshold as a training set from the historical feature indicators, select a predetermined number of historical feature indicators other than those in the training set as a verification set, and select the remaining historical feature indicators as a test set; the modeling module is configured to obtain an initial value prediction model by machine learning based on the historical feature indicators in the training set, and perform model adjustment on the initial value prediction model based on the historical feature indicators in the verification set; and the modeling module is further configured to test the current value prediction model according to the historical feature indicators in the test set, and complete establishment of the value prediction model if the test is passed.

[0015] In a possible implementation, the apparatus further includes a calculation module configured to fit each communication state information of the base station cell with an exponential distribution, obtain an exponential probability density function of the each communication state information, and integrate the exponential probability density function to obtain a cumulative probability distribution of the each communication state information; and the calculation module is further configured to calculate the value degree of the base station cell based on a value degree formula, where the value degree formula includes:

[0016]

[0017] where n is a number of the first communication state information other than the measurement report coverage rate, Wi is a weight of the ith first communication state information, F i (x) is a cumulative probability value of the ith first communication state information when the value is x, W mr is a weight of the measurement report coverage rate, and M is the measurement report coverage rate; the calculation module is further configured to add a low-value cell label to the base station cell if the value degree of the base station cell is less than a preset second threshold, and add a high-value label to the base station cell if the value degree of the base station cell is not less than the second threshold.

[0018] In a possible implementation, the basic information includes a location, a type, and an identifier of the base station cell; and the communication state information includes at least one of an uplink network resource utilization rate, a downlink network resource utilization rate, uplink traffic, downlink traffic, a number of users, a number of VIP users, an hourly tariff, and a measurement report coverage rate.

[0019] In a possible implementation, the processing module is further configured to, if the base station cell belongs to a predetermined specific cell set, use an input feature index of the base station cell as input data of a first value prediction model corresponding to the base station cell to obtain a value prediction result of the base station cell output by the first value prediction model, where the first value prediction model is obtained through machine learning based on historical feature indexes of the base station cell and corresponding known results; and each base station cell in the specific cell set corresponds to a pre-established value prediction model.

[0020] In a possible implementation, the processing module is further configured to, if the value prediction result of the base station cell in any future time period is a low-value cell, execute a base station shutdown instruction on the base station cell in a subsequent predetermined time period; and the processing module is further configured to, if the value prediction result of the base station cell in any future time period is a high-value cell, not perform processing on the base station cell.

[0021] In a third aspect, the present application provides an electronic device, comprising: a processor, and a memory connected with the processor in communication; the memory stores computer execution instructions; the processor executes the computer execution instructions stored in the memory to implement the method in any one of the first aspect.

[0022] In a fourth aspect, the present application provides a computer readable storage medium, wherein the computer readable storage medium stores computer execution instructions, and the computer execution instructions are executed by a processor to implement the method in any one of the first aspect.

[0023] The cell value prediction method, device, electronic device and storage medium provided by the present application obtain input feature indexes of a base station cell at any time period of the day, wherein the input feature indexes include communication state information, value labels and basic information of the base station cell, and time features of the time period; the input feature indexes of the base station cell are taken as input data of a value prediction model, and a value prediction result of the base station cell output by the value prediction model is obtained; wherein the value prediction model is established based on historical feature indexes of each base station cell through machine learning; the historical feature indexes include communication states, historical value labels and basic information of the base station cell at a historical first time period, and time features of the predetermined time period. Through the above scheme, an accurate value prediction model can be established based on a large number of historical feature indexes through machine learning, so as to realize accurate prediction of cell value by using the value prediction model. BRIEF DESCRIPTION OF DRAWINGS

[0024] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments consistent with the present application and serve to explain the principles of the present application together with the specification.

[0025] Figure 1 An application scenario diagram of the cell value prediction method provided by the present application is shown in the following figure:

[0026] Figure 2 A flowchart of the cell value prediction method provided by the first embodiment of the present application is shown in the following figure:

[0027] Figure 3 A data set acquisition example provided by the present application is shown in the following table:

[0028] Figure 4 A flowchart of the cell value prediction method provided by the second embodiment of the present application is shown in the following figure:

[0029] Figure 5 A structure diagram of the cell value prediction device provided by the third embodiment of the present application is shown in the following figure:

[0030] Figure 6 A device block diagram of a cell value prediction device provided by the fifth embodiment of the present application is shown in the following figure:

[0031] Figure 7 A structural schematic diagram of an electronic device provided in Embodiment Six of the present application is shown in the following figure:

[0032] The specific embodiments of the present application have been shown and described in the above figures, and will be described in more detail hereinafter. These figures and the written description are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0033] The exemplary embodiments will be described in detail herein with reference to the attached drawings. When the description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The following exemplary embodiments described are not meant to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with aspects related to the present application.

[0034] First, the terms involved are explained:

[0035] CatBoost: a machine learning algorithm based on gradient boosting decision trees.

[0036] Figure 1 An application scenario diagram of the cell value prediction method provided in the present application is shown in the following figure: Figure 1 As shown in the figure, the scenario includes a user terminal 1, a base station cell 2, and a cell value prediction device 3.

[0037] Taking the illustrated scenario as an example: the user terminal 1 sends a data request signal to the base station cell 2, the base station cell 2 sends a corresponding data signal according to the data request signal of the user terminal 1, and the user terminal 1 receives the data signal to realize mobile communication online. The cell value prediction device 3 obtains the communication state information, value label, and basic information of the base station cell 2, and accurately predicts whether the base station cell 2 is a low-value cell through the value prediction model established by machine learning.

[0038] In actual application, predicting the value of the base station cell can be used for base station energy saving. For example, if it is predicted that a certain base station cell is a low-value cell, it means that the actual usage of the user terminal to the base station cell is low, and the user terminal will not be affected even if it only receives the data signal of other base station cells. At this time, the low-value cell can be executed with a shutdown instruction to stop the operation of the low-value cell without affecting the user's use, thereby realizing accurate energy saving.

[0039] The scheme of the embodiments of the present application will be exemplarily introduced in combination with the following embodiments.

[0040] Embodiment One

[0041] Figure 2 A flowchart of a cell value prediction method provided for Embodiment One of the present application is shown in the figure. The method comprises the following steps:

[0042] S101, input feature indicators of a base station cell at any time period of the day are obtained, the input feature indicators comprising communication state information, value labels and basic information of the base station cell, and time features of the time period;

[0043] S102, the input feature indicators of the base station cell are taken as input data of a value prediction model, and a value prediction result of the base station cell output by the value prediction model is obtained; wherein the value prediction model is established based on historical feature indicators of each base station cell through machine learning; the historical feature indicators comprising communication states, historical value labels and basic information of the base station cell at a historical first time period, and time features of the predetermined time period.

[0044] In actual application, the execution subject of this embodiment can be a cell value prediction device, and the cell value prediction device can be implemented in multiple ways. For example, it can be a program software, or a medium storing a related computer program, such as a U disk, etc.; or the device can also be an entity device integrated or installed with a related computer program, which supports USB interface connection, such as a chip, a smart terminal, a computer, a server, etc.

[0045] In one example, S101 specifically comprises: obtaining historical feature indicators of a base station cell at a first time period of a historical date; selecting historical feature indicators with dates located before a preset first threshold value from the historical feature indicators as a training set, and selecting a predetermined number of historical feature indicators from historical feature indicators other than the training set as a validation set, and the rest of the historical feature indicators as a test set; obtaining an initial value prediction model based on the historical feature indicators in the training set by using a machine learning algorithm; adjusting the initial value prediction model based on the historical feature indicators in the validation set; testing the current value prediction model according to the historical feature indicators in the test set, and if the test is passed, the establishment of the value prediction model is completed.

[0046] Optionally, as shown in Figure 3 , the method further comprises the following steps: Figure 3For the data set, an example is obtained. The historical feature indicators of the base station cell for 30 days are obtained. The data from the 12th day to the 30th day in the historical feature indicators is selected as a training set. The data from the 6th day to the 12th day in the historical feature indicators is selected as a validation set. The remaining historical feature indicators are selected as a test set. An initial value prediction model is obtained based on the historical feature indicators in the training set by using a CatBoost machine learning algorithm. The historical feature indicators in the validation set are brought into the initial value prediction model for model adjustment. The performance of the initial value prediction model is adjusted and optimized in the training process. The historical feature indicators in the test set are brought into the current optimized value prediction model. Whether the value prediction model meets the target requirement is evaluated by testing. If the target requirement is met, the establishment of the value prediction model is completed.

[0047] Alternatively, the historical feature indicators of the base station cell for 1 month are obtained. 18 days of data are randomly selected from the historical feature indicators as a training set. 6 days of data are randomly selected from the remaining historical feature indicators as a validation set. The remaining historical feature indicators are selected as a test set.

[0048] Based on the above embodiments, a large number of historical feature indicators can be obtained, so that an accurate value prediction model can be established by machine learning.

[0049] In one example, the establishment of the value prediction model includes: fitting each communication state information of the base station cell with an exponential distribution, obtaining an exponential probability density function of the each communication state information, and performing integral processing on the exponential probability density function to obtain a cumulative probability distribution of the each communication state information; calculating a value degree of the base station cell based on a value degree formula; wherein the value degree formula includes:

[0050]

[0051] wherein n is the number of the first communication state information except the measurement report coverage rate, Wi is the weight of the i-th first communication state information, F i (x) is the cumulative probability value of the i-th first communication state information when the value is x, W mr is the weight of the measurement report coverage rate, and M is the measurement report coverage rate. If the value degree of the base station cell is less than a preset second threshold, a low-value cell label is added to the base station cell. If the value degree of the base station cell is not less than the second threshold, a high-value label is added to the base station cell.

[0052] In combination with a scenario example, 1-month historical communication state information is acquired, and statistical analysis shows that each historical communication state information is approximately exponentially distributed, so that exponential distribution fitting is performed on the historical communication state information to obtain an exponential probability density function of the historical communication state information, and integral processing is performed on the exponential probability density function to obtain a cumulative probability distribution function F(X) of each communication state information:

[0053] F(X) = P{X≤x}

[0054] where X is a random variable, x is an arbitrary real number, and the cumulative probability distribution function F(X) represents a probability that the random variable X takes a value less than or equal to x. The cumulative probability distribution functions F(X) of the communication state information are weighted and calculated according to preset corresponding weights to obtain a base station cell value degree.

[0055] Based on the above implementation, a variety of types of data can be comprehensively accumulated to calculate a cumulative probability distribution function, so that a base station cell value degree closer to a real situation is obtained.

[0056] In one example, the basic information includes a location, a type, and an identifier of the base station cell; the communication state information includes at least one of an uplink network resource utilization rate, a downlink network resource utilization rate, uplink traffic, downlink traffic, a number of users, a number of VIP users, an hourly tariff, and a measurement report coverage rate; the value label includes a low-value cell label and a high-value cell label; and the time characteristics of the time period include an hour of the time period, a day of the time period, a day of the week of the time period, and whether it is a holiday.

[0057] The application of various input feature indicators is described below in combination with multiple scenarios.

[0058] In combination with a scenario example, the basic information is an inherent feature of each base station, which is used to associate with a corresponding base station for classification and searching. For example, the location and type of the basic information can be used to add a filtering field, and an administrator can query a base station cell according to the field filtering.

[0059] In combination with a scenario example, the uplink network resource utilization rate, the downlink network resource utilization rate, the uplink traffic, the downlink traffic, the number of users, the number of VIP users, and the hourly tariff in the communication state information reflect the usage amount of users on the base station cell. For example, if the uplink network resource utilization rate, the downlink network resource utilization rate, the uplink traffic, and the downlink traffic are high, it indicates that the base station cell actually uses more resources and has less idle resources, which means that the users have high dependence on the base station cell and have a large usage amount, and the value degree of the base station cell is also higher. If the number of users is large, it indicates that more users have high dependence on the base station cell and have a large usage amount, and the value degree of the base station cell is also higher.

[0060] In combination with a scenario example, the measurement report coverage in the communication state information represents the resource amount provided by the base station cell. For example, if the measurement report coverage is high, it indicates that the signal strength of the base station cell is higher, and in the case where the user usage amount is unchanged, the higher the signal strength of the base station cell, the lower the resource utilization rate, the higher the energy consumption, and the lower the value of the base station cell. Therefore, the measurement report coverage is taken as the reciprocal in the base station cell value degree formula for calculation.

[0061] Optionally, the weights of the communication state information can be adjusted according to different business needs. For example, if more attention is paid to the download business of users or the demand of users in the coverage area of the base station cell for the download business is greater, the weights of the downlink network resource utilization or the downlink traffic in the communication state information can be increased.

[0062] In combination with a scenario example, the value label is used to provide information for machine learning to predict the cell value. For example, if a cell is evaluated as a low-value cell multiple times in history, the probability that the cell is predicted as a low-value cell in the future is higher.

[0063] In combination with a scenario example, the time feature of the period is used to provide time regularity information for machine learning to predict the cell value. For example, if a cell is evaluated as a low-value cell multiple times from 3 p.m. to 5 p.m. on Monday to Friday in history, the probability that the cell is predicted as a low-value cell from 3 p.m. to 5 p.m. on Monday to Friday in the future is higher.

[0064] Based on the above embodiments, multiple types of input feature indicators can be comprehensively analyzed instead of being independently analyzed, and in the case of a communication network with diversified business scenarios, more accurate prediction can be achieved.

[0065] In one example, S101 further includes: if the base station cell belongs to a predetermined specific cell set, taking the input feature indicators of the base station cell as input data of a first value prediction model corresponding to the base station cell, obtaining a value prediction result of the base station cell output by the first value prediction model; wherein the first value prediction model is obtained through machine learning based on historical feature indicators of the base station cell and corresponding known results; and wherein each base station cell in the specific cell set corresponds to a pre-established value prediction model.

[0066] In combination with a scenario example, for specific cells such as government office areas, hospitals, and the like, due to the particularity of their businesses, the base station cells are modeled separately, the input feature indicators of the base station cells are taken as input data of a first value prediction model corresponding to the base station cells, and a value prediction result of the base station cells output by the first value prediction model is obtained.

[0067] Based on the above embodiment, a specific cell is modeled separately, only the historical feature indicators of the specific cell are obtained to perform machine learning modeling to obtain a value prediction model, thereby excluding the interference of the historical feature indicators of other business different base station cells on the value prediction model, and realizing the accurate identification of low-value base station cells in different scenes.

[0068] In one example, S102 specifically includes: inputting the input feature indicators of the base station cell as input data of a value prediction model, the value prediction model outputting a predicted probability that the base station cell is a low-value cell in a future period, and if the predicted probability that the base station cell is a low-value cell in a future period is greater than a preset third threshold, it is determined that the base station cell is a low-value cell in a future period.

[0069] For example, if the predicted probability that a certain cell is a low-value cell at five o'clock tomorrow is 0.62, which is less than the third threshold 0.80, it is determined that the cell is a high-value cell at five o'clock tomorrow.

[0070] Based on the above embodiment, the result output by the value prediction model can be judged by the method of the preset third threshold, the third threshold can be adjusted according to the actual scene, thereby effectively improving the accuracy of value prediction.

[0071] In the cell value prediction method provided in the embodiment, the input feature indicators of the base station cell in any period of the day are obtained, the input feature indicators include the communication state information, value label and basic information of the base station cell, and the time characteristics of the period. The input feature indicators of the base station cell are input as input data of a value prediction model, and the value prediction result of the base station cell output by the value prediction model is obtained. The value prediction model is established based on the historical feature indicators of each base station cell through machine learning; the historical feature indicators include the communication state, historical value label and basic information of the base station cell in a historical first period, and the time characteristics of the predetermined period. Through the above scheme, an accurate value prediction model can be established based on a large number of historical feature indicators through machine learning, so as to realize accurate prediction of cell value by using the value prediction model.

[0072] Embodiment Two

[0073] Figure 4 A flowchart of a cell value prediction method provided in Embodiment Two of the present application is shown in the figure, and based on Embodiment One, the processing flow of the base station cell after obtaining the value prediction result is exemplarily illustrated, as shown in Figure 4 Based on Embodiment One, the method further includes:

[0074] S103, if the base station cell in any future period of the value prediction result is low value cell, then in the future the predetermined period of the base station cell is executed base station closing instruction;

[0075] S104, if the base station cell in any future period of the value prediction result is high value cell, then the base station cell is not executed processing.

[0076] In one example, S103 specifically includes: automatically sending closing instruction to the base station cell in the future period, the base station cell receiving the closing instruction and closing power supply, when the future period ends, the base station cell automatically starts power supply and restores to normal operation state.

[0077] In one example, S103 further includes: real-time monitoring the service load of the base station cell adjacent to the base station cell closing power supply, if it is detected that the service load of the adjacent base station cell exceeds the preset fourth threshold, then automatically sending opening instruction to the base station cell closing power supply to make it restore to normal operation state.

[0078] In combination with the scene example, if the load of the adjacent base station cell exceeds the fourth threshold, it means that the adjacent base station cell shares the service of the base station cell executing the closing instruction and exceeds the load upper limit of the adjacent base station cell itself, at this time, it may cause network fluctuation of users. Therefore, the opening instruction is automatically sent to the base station cell closing power supply to make it restore to normal operation state, so as to ensure normal use of users.

[0079] Based on the above embodiment, the service load of the base station cell adjacent to the base station cell closing power supply is monitored in real time, so as to realize that the user perception is not affected.

[0080] In the cell value prediction method provided by the embodiment, if the base station cell in any future period of the value prediction result is low value cell, then in the future the predetermined period of the base station cell is executed base station closing instruction. If the base station cell in any future period of the value prediction result is high value cell, then the base station cell is not executed processing. Through the above scheme, the corresponding processing mode can be executed according to different prediction results, so as to realize precise energy saving under the premise of guaranteeing user service performance.

[0081] Embodiment three

[0082] Figure 5 The structure schematic diagram of the cell value prediction device provided by the embodiment three of the application is shown as Figure 5 The device includes:

[0083] The acquisition module 61 is configured to acquire input feature indexes of a base station cell in any time period of a day, wherein the input feature indexes comprise communication state information, a value label and basic information of the base station cell, and time features of the time period.

[0084] The processing module 62 is configured to take the input feature indexes of the base station cell as input data of a value prediction model, and obtain a value prediction result of the base station cell output by the value prediction model, wherein the value prediction model is established based on historical feature indexes of each base station cell through machine learning, and the historical feature indexes comprise communication states, historical value labels and basic information of the base station cell in a historical first time period, and time features of the predetermined time period.

[0085] In actual application, the cell value prediction device can be implemented in various ways. For example, the device can be a program software, a medium storing a related computer program, such as a U disk, or an entity device integrating or installing a related computer program, which supports USB interface connection, such as a chip, a smart terminal, a computer or a server.

[0086] In one example, the device further comprises a modeling module 63, and the acquisition module 61 is further configured to acquire historical feature indexes of a base station cell in a first time period of a historical date; the acquisition module 61 is further configured to select historical feature indexes with dates before a preset first threshold as a training set from the historical feature indexes, select a predetermined number of historical feature indexes as a validation set from historical feature indexes other than the training set, and select the remaining historical feature indexes as a test set; the modeling module 63 is further configured to obtain an initial value prediction model by using a machine learning algorithm based on the historical feature indexes in the training set, adjust the initial value prediction model based on the historical feature indexes in the validation set, and test the current value prediction model according to the historical feature indexes in the test set, and if the test is passed, the establishment of the value prediction model is completed.

[0087] In one example, the device further comprises a calculation module 64 configured to fit each communication state information of the base station cell by using an exponential distribution, obtain an exponential probability density function of the each communication state information, and perform integral processing on the exponential probability density function to obtain a cumulative probability distribution of the each communication state information; and the calculation module 64 is further configured to calculate a value degree of the base station cell based on a value degree formula, wherein the value degree formula comprises:

[0088]

[0089] wherein n is the number of the first communication state information other than the measurement report coverage, Wi is the weight of the i-th first communication state information, F i (x) is the cumulative probability value of the i-th first communication state information taking value x, W mr is the weight of the measurement report coverage, and M is the measurement report coverage; the calculation module 64 is further configured to add a low-value cell label to the base station cell if the value degree of the base station cell is less than a preset second threshold, and add a high-value label to the base station cell if the value degree of the base station cell is not less than the second threshold.

[0090] In one example, the basic information includes the location, type and identification of the base station cell; the communication state information includes at least one of the following: uplink network resource utilization, downlink network resource utilization, uplink traffic, downlink traffic, user number, VIP user number, hourly tariff and measurement report coverage; the value label includes a low-value cell label and a high-value cell label; and the time characteristics of the time period include the hour, day, day of the week of the time period and whether it is a holiday.

[0091] In one example, the acquisition module 61 is further configured to, if the base station cell belongs to a predetermined specific cell set, take the input feature indicators of the base station cell as input data of a first value prediction model corresponding to the base station cell, and obtain a value prediction result of the base station cell output by the first value prediction model; wherein the first value prediction model is obtained through machine learning based on historical feature indicators of the base station cell and corresponding known results thereof; and wherein each base station cell in the specific cell set corresponds to a pre-established value prediction model.

[0092] In one example, the processing module 62 is specifically configured to take the input feature indicators of the base station cell as input data of a value prediction model, and the value prediction model outputs a prediction probability that the base station cell is a low-value cell in a future time period; and if the prediction probability that the base station cell is a low-value cell in the future time period is greater than a preset third threshold, it is determined that the base station cell is a low-value cell in the future time period.

[0093] The cell value prediction device provided in the embodiment comprises an acquisition module, configured to acquire input feature indexes of a base station cell in any time period of a day, wherein the input feature indexes comprise communication state information, value labels and basic information of the base station cell, and time features of the time period; and a processing module, configured to take the input feature indexes of the base station cell as input data of a value prediction model, and obtain a value prediction result of the base station cell output by the value prediction model, wherein the value prediction model is established based on historical feature indexes of each base station cell through machine learning, and the historical feature indexes comprise communication states, historical value labels and basic information of the base station cell in a historical first time period, and time features of the predetermined time period. Through the above scheme, an accurate value prediction model can be established based on a large number of historical feature indexes through machine learning, so that the value prediction model is used to accurately predict the value of a cell.

[0094] Embodiment Four

[0095] The embodiment four of the application provides a multi-cluster container management device, which further comprises the following on the basis of the embodiment three:

[0096] The processing module is further configured to execute a base station shutdown instruction on the base station cell in a subsequent predetermined time period if the value prediction result of the base station cell in any future time period is a low-value cell.

[0097] The processing module is further configured to not execute processing on the base station cell if the value prediction result of the base station cell in any future time period is a high-value cell.

[0098] In one example, the processing module is specifically configured to automatically send a shutdown instruction to the base station cell in the future time period, and the base station cell shuts down power supply after receiving the shutdown instruction, and automatically starts power supply and returns to a normal operation state when the future time period ends.

[0099] In one example, the processing module is further configured to monitor a service load of a base station cell adjacent to the base station cell with shut down power supply in real time, and automatically send a start instruction to the base station cell with shut down power supply to make it return to a normal operation state if it is detected that the service load of the adjacent base station cell exceeds a preset fourth threshold.

[0100] The cell value prediction device provided in the embodiment, the processing module is further used for executing a base station closing instruction to the base station cell in the future any time period. The processing module is further used for not executing processing to the base station cell if the value prediction result of the base station cell in the future any time period is a high-value cell. Through the above scheme, corresponding processing modes can be executed according to different prediction results, so as to realize precise energy saving under the premise of guaranteeing user service performance.

[0101] Embodiment five

[0102] Figure 6 Fig. 1 is a device block diagram of a cell value prediction device according to an exemplary embodiment. The device can be a mobile phone, a computer, a digital broadcast terminal, a messaging equipment, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, etc.

[0103] The device 800 can include one or more of the following components: a processing component 802, a memory 804, a power supply component 806, a multimedia component 808, an audio component 810, an input / output (I / O) interface 812, a sensor component 814, and a communication component 816.

[0104] The processing component 802 usually controls overall operations of the device 800, such as operations associated with display, phone call, data communication, camera operation and recording operation. The processing component 802 can include one or more processors 820 to execute instructions to complete all or part of steps of the methods described above. Further, the processing component 802 can include one or more modules to facilitate interaction between the processing component 802 and other components. For example, the processing component 802 can include a multimedia module to facilitate the interaction between the multimedia component 808 and the processing component 802.

[0105] The memory 804 is configured to store various types of data to support operations of the device 800. Examples of these data include instructions for any application or method operating on the device 800, contact data, phonebook data, messages, pictures, videos, etc. The memory 804 can be implemented by any type of volatile or non-volatile storage devices or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0106] Power component 806 provides power to the various components of device 800. Power component 806 can include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for device 800.

[0107] Multimedia component 808 includes a screen providing an output interface between device 800 and a user. In some embodiments, the screen can include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen can be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touch, swiping, and gestures on the touch panel. The touch sensors can not only sense a boundary of a touching or swiping action, but also detect duration and pressure related to the touching or swiping action. In some embodiments, multimedia component 808 includes a front camera and / or a rear camera. When device 800 is in an operation mode, such as a shooting mode or a video mode, the front camera and / or the rear camera can receive external multimedia data. Each of the front and rear cameras can be a fixed optical lens system or have a focal length and optical zoom capability.

[0108] Audio component 810 is configured to output and / or input audio signals. For example, audio component 810 includes a microphone (MIC) that is configured to receive external audio signals when device 800 is in an operation mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signals can be further stored in memory 804 or transmitted via communication component 816. In some embodiments, audio component 810 also includes a speaker for outputting audio signals.

[0109] I / O interface 812 provides an interface between processing component 802 and peripheral interface modules, which can be a keyboard, a click wheel, buttons, and the like. The buttons can include, but are not limited to, a home button, a volume button, a start button, and a lock button.

[0110] The sensor component 814 includes one or more sensors for providing status assessments for various aspects of the device 800. For example, the sensor component 814 can detect an open / closed position of the device 800, relative positioning of components, such as a display and keypad of the device 800, changes in position of the device 800 or a component of the device 800, presence or absence of user contact with the device 800, changes in orientation of the device 800 or acceleration / deceleration, and temperature changes of the device 800. The sensor component 814 can include proximity sensor(s) configured to detect presence of nearby objects without any physical contact. The sensor component 814 can also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor component 814 can also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0111] The communication component 816 is configured to facilitate wired or wireless communication between the device 800 and another device. The device 800 can access a wireless network based on a corresponding communication standard, such as WiFi, 2G, or 3G, or a combination thereof. In an example embodiment, the communication component 816 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an example embodiment, the communication component 816 also includes a Near Field Communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on Radio Frequency Identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0112] In an example embodiment, the device 800 can be implemented using one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), controllers, micro-controllers, microprocessors, or other electronic components, for performing the above-described methods.

[0113] In an example embodiment, a non-transitory computer-readable storage medium, such as the memory 804 including instructions, is also provided, which can be executed by the processor 820 of the device 800 to complete the above-described methods. For example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disc, and an optical data storage device, etc.

[0114] Embodiment Six

[0115] Figure 7 A structural schematic diagram of an electronic device provided in an embodiment of the present application is shown in FIG. 1. As shown in FIG. 1, the electronic device includes a processor 820, a memory 804, a display 802, a keypad 806, a communication component 816, a sensor component 814, and a power supply 812. Figure 7As shown, the electronic device includes:

[0116] The processor 291, the electronic device further includes a memory 292; can also include a communication interface 293 and a bus 294. Among them, the processor 291, the memory 292, the communication interface 293, can complete the mutual communication through the bus 294. The communication interface 293 can be used for information transmission. The processor 291 can call the logic instruction in the memory 294 to execute the method of the above-mentioned embodiment.

[0117] In addition, the logic instruction in the memory 292 described above can be realized in the form of a software function unit and sold or used as an independent product, which can be stored in a computer readable storage medium.

[0118] The memory 292 as a kind of computer readable storage medium can be used to store software programs, computer executable programs, such as the program instruction / module corresponding to the method in the embodiment of the application. The processor 291 executes the function application and data processing by running the software program, instruction and module stored in the memory 292, that is, realizes the method in the above-mentioned method embodiment.

[0119] The memory 292 can include a storage program area and a storage data area, wherein the storage program area can store an operating system, at least one application required by a function; the storage data area can store data created according to the use of the terminal device and the like. In addition, the memory 292 can include a high-speed random access memory, and can also include a nonvolatile memory.

[0120] The embodiment of the application provides a kind of non-transitory computer readable storage medium, the computer readable storage medium in it stores computer execution instruction, the computer execution instruction is executed by processor when for realizing the method as described in the preceding embodiment.

[0121] The embodiment of the application provides a kind of computer program product, including computer program, the computer program is executed by processor when realizing the method as described in the preceding embodiment.

[0122] Other embodiments of the application will be apparent to those skilled in the art from consideration of the specification and practice of the application disclosed herein. The application is intended to cover any variations, uses or adaptations of the application following, in general, the principles of the application and including such departures from the present disclosure as come within known or customary practice in the art to which the application pertains. The specification and examples are to be regarded as exemplary only, and the true scope and spirit of the application are indicated by the following claims.

[0123] It is to be understood that the application is not limited to the precise construction already described above and shown in the drawings, and that various modifications and changes can be made by those skilled in the art without departing from the scope of the application. The scope of the application should only be limited by the claims appended hereto.

Claims

1. A cell value prediction method characterized by, The method comprises: obtaining input feature indicators of a base station cell in any time period of the day, the input feature indicators comprising communication state information, value labels and basic information of the base station cell, and time features of the time period; using the input feature indicators of the base station cell as input data of a value prediction model to obtain a value prediction result of the base station cell output by the value prediction model; wherein the value prediction model is established based on historical feature indicators of each base station cell through machine learning; the historical feature indicators comprise communication states, historical value labels and basic information of a base station cell in a historical first time period, and time features of a predetermined time period; The method further comprises: fitting each communication state information of the base station cell with an exponential distribution to obtain an exponential probability density function of the communication state information, and integrating the exponential probability density function to obtain a cumulative probability distribution of the communication state information; the communication state information comprises at least one of uplink network resource utilization, downlink network resource utilization, uplink traffic, downlink traffic, user number, VIP user number, hourly tariff and measurement report coverage rate, and the measurement report coverage rate is used to indicate the resource amount provided by the base station cell; calculating the value degree of the base station cell based on a value degree formula; wherein the value degree formula comprises: wherein n is the number of first communication status information other than the measurement report coverage, Wi is the weight of the i-th first communication status information, F i (x) is the cumulative probability value when the i-th first communication status information takes the value x, W mr is the weight of the measurement report coverage, and M is the measurement report coverage. if the value degree of the base station cell is less than a preset second threshold, adding a low-value cell label to the base station cell; if the value degree of the base station cell is not less than the second threshold, adding a high-value label to the base station cell; The method further comprises: if the value prediction result of the base station cell in any future time period is a low-value cell, executing a base station shutdown instruction on the base station cell in the subsequent predetermined time period; if the value prediction result of the base station cell in any future time period is a high-value cell, not performing processing on the base station cell.

2. The method of claim 1, wherein, The method further comprises: obtaining historical feature indicators of a base station cell in a first time period of a historical date; selecting historical feature indicators with dates before a preset first threshold from the historical feature indicators as a training set, and selecting a predetermined number of historical feature indicators from historical feature indicators other than the training set as a validation set, and the remaining historical feature indicators as a test set; obtaining an initial value prediction model using a machine learning algorithm based on the historical feature indicators in the training set; adjusting the initial value prediction model based on the historical feature indicators in the validation set; testing the current value prediction model according to the historical feature indicators in the test set, and if the test is passed, completing the establishment of the value prediction model.

3. The method of claim 1, wherein, The basic information comprises the location, type and identification of the base station cell.

4. The method of claim 1, wherein, After obtaining the input feature indicators of the base station cell in any time period of the day, the method further comprises: If the base station cell belongs to a predetermined specific cell set, input feature indicators of the base station cell are taken as input data of a first value prediction model corresponding to the base station cell, and a value prediction result of the base station cell output by the first value prediction model is obtained; wherein the first value prediction model is obtained through machine learning based on historical feature indicators of the base station cell and corresponding known results thereof. Each base station cell in the specific cell set corresponds to a pre-established value prediction model.

5. A cell value prediction apparatus characterized by comprising: Comprise: The acquisition module is configured to acquire input feature indicators of a base station cell in any time period of the day, wherein the input feature indicators comprise communication state information, value labels and basic information of the base station cell, and time features of the time period. The processing module is configured to take the input feature indicators of the base station cell as input data of a value prediction model, and obtain a value prediction result of the base station cell output by the value prediction model; wherein the value prediction model is established through machine learning based on historical feature indicators of each base station cell; the historical feature indicators comprise communication states, historical value labels and basic information of the base station cell in a historical first time period, and time features of a predetermined time period. The device further comprises a calculation module configured to fit each communication state information of the base station cell with an exponential distribution, obtain an exponential probability density function of the each communication state information, and perform integral processing on the exponential probability density function to obtain a cumulative probability distribution of the each communication state information; the communication state information comprises at least one of uplink network resource utilization, downlink network resource utilization, uplink traffic, downlink traffic, user number, VIP user number, hourly tariff and measurement report coverage rate, and the measurement report coverage rate is used to indicate a resource amount provided by the base station cell. The calculation module is further configured to calculate a value degree of the base station cell based on a value degree formula; wherein the value degree formula comprises: wherein n is the number of first communication status information other than the measurement report coverage, Wi is the weight of the i-th first communication status information, F i (x) is the cumulative probability value when the i-th first communication status information takes the value x, W mr is the weight of the measurement report coverage, and M is the measurement report coverage. The calculation module is further configured to add a low-value cell label to the base station cell if the value degree of the base station cell is less than a preset second threshold, and add a high-value label to the base station cell if the value degree of the base station cell is not less than the second threshold. The processing module is further configured to execute a base station shutdown instruction on the base station cell in a predetermined time period after the base station cell if the value prediction result of the base station cell in any future time period is a low-value cell. The processing module is further configured to not perform processing on the base station cell if the value prediction result of the base station cell in any future time period is a high-value cell.

6. The apparatus of claim 5, wherein, The device further comprises a modeling module: The acquisition module is further configured to acquire historical feature indicators of a base station cell in a first time period of a historical date; The acquisition module is further configured to select historical feature indicators with dates before a preset first threshold as a training set from the historical feature indicators, select a predetermined number of historical feature indicators as a validation set from historical feature indicators other than the training set, and select the remaining historical feature indicators as a test set. The modeling module is configured to obtain an initial value prediction model based on historical feature indicators in the training set by using a machine learning algorithm, and to adjust the initial value prediction model based on historical feature indicators in the verification set. The modeling module is further configured to test the current value prediction model according to historical feature indicators in the test set, and to complete the establishment of the value prediction model if the test is passed.

7. The apparatus of claim 5, wherein, The basic information includes a location, a type, and an identifier of the base station cell.

8. The apparatus of claim 5, wherein, The processing module is further configured to, if the base station cell belongs to a predetermined specific cell set, take an input feature indicator of the base station cell as input data of a first value prediction model corresponding to the base station cell, and obtain a value prediction result of the base station cell output by the first value prediction model, wherein the first value prediction model is obtained based on historical feature indicators of the base station cell and corresponding known results by machine learning. Each base station cell in the specific cell set corresponds to a pre-established value prediction model.

9. An electronic device comprising: A processor, and a memory connected to the processor in communication; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method of any one of claims 1-4.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are executed by the processor to implement the method of any one of claims 1-4.

Images