A method and device for generating bandwidth data prediction results

Abstract

An embodiment of the present invention provides a method and device for generating bandwidth data prediction results, which obtains historical bandwidth usage data and generates a timestamp corresponding to the historical bandwidth usage data; generates multiple historical data time series for the historical bandwidth usage data based on the timestamp; divides the historical bandwidth usage data into multiple bandwidth fuzzy clusters through the multiple historical data time series; establishes a prediction data cluster attribution model based on the multiple bandwidth fuzzy clusters; determines neural network initialization parameters, and generates a bandwidth prediction model through the prediction data cluster attribution model and the neural network initialization parameters; and generates a bandwidth prediction result based on the bandwidth prediction model, thereby improving the efficiency and accuracy of bandwidth data prediction, reducing the maintenance cost of underlying equipment, and further improving the normal brightness rate of pass codes.

Description

Technical Field

[0001] The present invention relates to the technical field of bandwidth data prediction result generation, and in particular to a bandwidth data prediction result generation method, a bandwidth data prediction result generation device, an electronic device, and a computer-readable storage medium. Background Art

[0002] Passcodes are an important tool in daily life. The huge number of visitors undoubtedly puts tremendous pressure on the stable operation of the system. In order to ensure the sustainable and normal operation of passcodes, a large amount of manpower and material resources are required to monitor and maintain the equipment of the passcode business. However, since the usage changes of passcodes are affected by many factors, if the usage of passcodes cannot be accurately predicted, it is still impossible to adjust the underlying equipment in time according to the usage to ensure the normal use of passcodes. Once the passcode system is abnormal, it is very easy to cause queues and congestion in areas with high traffic volume, such as subway stations and hospitals, causing inconvenience and even chaos.

[0003] Therefore, how to predict the usage of the pass code becomes a problem that those skilled in the art need to overcome. Summary of the Invention

[0004] The embodiments of the present invention provide a method, device, electronic device and computer-readable storage medium for generating bandwidth data prediction results, so as to solve the problem of how to predict the usage of a pass code.

[0005] An embodiment of the present invention discloses a method for generating bandwidth data prediction results, which may include:

[0006] Acquire historical bandwidth usage data and generate a timestamp corresponding to the historical bandwidth usage data;

[0007] generating a plurality of historical data time series for the historical bandwidth usage data based on the timestamp;

[0008] Dividing the historical bandwidth usage data into a plurality of bandwidth fuzzy clusters according to a plurality of the historical data time series;

[0009] Establishing a prediction data clustering attribution model based on a plurality of bandwidth fuzzy clusters;

[0010] Determining neural network initialization parameters, and generating a bandwidth prediction model using the predicted data clustering attribution model and the neural network initialization parameters;

[0011] A bandwidth prediction result is generated based on the bandwidth prediction model.

[0012] Optionally, the step of dividing the historical bandwidth usage data into a plurality of bandwidth fuzzy clusters using a plurality of the historical data time series may include:

[0013] Determining a category number parameter, a first fuzzy membership parameter, and a first fuzzy weight index parameter;

[0014] Determining corresponding initial cluster centers from the plurality of historical data time series based on the first fuzzy membership parameter and the first fuzzy weight index parameter according to a first preset algorithm;

[0015] Determining a second fuzzy weight index based on the category number parameter, the initial cluster center, and the first fuzzy weight index according to a second preset algorithm;

[0016] Determining a second fuzzy membership parameter based on the second fuzzy weight index, the initial cluster center, and the category number parameter according to a third preset algorithm;

[0017] Determining a target cluster center based on the initial cluster center, the second fuzzy weight index, the second fuzzy membership parameter, and the category number parameter; the target cluster center has a corresponding historical data time series;

[0018] The historical data time series is divided into a plurality of bandwidth fuzzy clusters according to the target cluster center and the category number parameter.

[0019] Optionally, the step of establishing a prediction data clustering attribution model based on the multiple bandwidth fuzzy clusters may include:

[0020] Determining a predicted bandwidth cluster center sequence and a plurality of other bandwidth cluster center sequences associated with the predicted bandwidth cluster center sequence from the bandwidth fuzzy clustering;

[0021] Calculating the Euclidean distance between the predicted bandwidth cluster center sequence and the plurality of other bandwidth cluster center sequences; the Euclidean distance has a corresponding Euclidean distance parameter; the Euclidean distance parameter corresponds one-to-one to the other bandwidth cluster center sequences;

[0022] Selecting a minimum Euclidean distance parameter from the Euclidean distance parameters, and determining the other bandwidth cluster center sequence corresponding to the minimum Euclidean distance parameter as the target bandwidth cluster center sequence;

[0023] Determining target bandwidth fuzzy clustering based on the target bandwidth cluster center sequence;

[0024] The predicted data clustering attribution model is established based on the target bandwidth fuzzy clustering.

[0025] Optionally, the predicted data clustering attribution model has corresponding predicted data clustering attribution data, and the step of generating a bandwidth prediction model using the predicted data clustering attribution model and the neural network initialization parameters may include:

[0026] Generating an initialization matrix for the clustering belonging data of the predicted data through a mapping formula; the initialization matrix includes a plurality of genes for expressing the bandwidth fuzzy clustering;

[0027] Determining a group fitness value through the initialization matrix, and determining an initial optimal gene from the genes;

[0028] Determining a maximum deviation distance parameter based on the initialization matrix, and generating a similarity time point number parameter and a maximum deviation time point number parameter according to the maximum deviation distance parameter in accordance with a preset standard;

[0029] Determine the target optimal gene based on the initial optimal gene, the neural network initialization parameter, the similarity time point number parameter, the maximum deviation time point number parameter and the predicted data clustering attribution data;

[0030] Determining initial weight parameters and threshold position parameters for a neural network based on the target optimal gene;

[0031] A neural network is trained according to the initial weight parameters, the threshold position parameters and the predicted data clustering attribution data to generate a bandwidth prediction model.

[0032] Optionally, the neural network initialization parameter may include an input layer number parameter, and the input layer number parameter is determined according to the number of timestamps.

[0033] Optionally, the step of generating a bandwidth prediction result based on the bandwidth prediction model may include:

[0034] Normalize the source data to be predicted to generate the data to be predicted;

[0035] Training the data to be predicted based on the bandwidth prediction model to obtain predicted value data;

[0036] Denormalization is performed on the predicted value data to generate a bandwidth prediction result.

[0037] The embodiment of the present invention further discloses a device for generating bandwidth data prediction results, which may include:

[0038] A historical bandwidth usage data acquisition module is used to acquire historical bandwidth usage data and generate a timestamp corresponding to the historical bandwidth usage data;

[0039] A historical data time series generating module, configured to generate a plurality of historical data time series for the historical bandwidth usage data based on the timestamp;

[0040] A bandwidth fuzzy clustering division module is used to divide the historical bandwidth usage data into a plurality of bandwidth fuzzy clusters according to a plurality of historical data time series;

[0041] A clustering attribution model establishment module, configured to establish a clustering attribution model for prediction data based on a plurality of bandwidth fuzzy clusters;

[0042] A bandwidth prediction model generation module is used to determine the neural network initialization parameters and generate a bandwidth prediction model through the predicted data clustering attribution model and the neural network initialization parameters;

[0043] The bandwidth prediction result generating module is used to generate a bandwidth prediction result based on the bandwidth prediction model.

[0044] Optionally, the bandwidth fuzzy clustering partitioning module may include:

[0045] A first parameter determination submodule, configured to determine a category number parameter, a first fuzzy membership parameter, and a first fuzzy weight index parameter;

[0046] an initial cluster center determination submodule, configured to determine corresponding initial cluster centers from a plurality of the historical data time series based on the first fuzzy membership parameter and the first fuzzy weight index parameter according to a first preset algorithm;

[0047] A second fuzzy weight index determination submodule, configured to determine a second fuzzy weight index based on the category number parameter, the initial cluster center, and the first fuzzy weight index according to a second preset algorithm;

[0048] a second fuzzy membership parameter determination submodule, configured to determine a second fuzzy membership parameter based on the second fuzzy weight index, the initial cluster center, and the category number parameter according to a third preset algorithm;

[0049] a target cluster center determination submodule, configured to determine a target cluster center based on the initial cluster center, the second fuzzy weight index, the second fuzzy membership parameter, and the category number parameter; the target cluster center having a corresponding historical data time series;

[0050] The bandwidth fuzzy clustering division submodule is used to divide the historical data time series into multiple bandwidth fuzzy clusters according to the target cluster center and the category number parameter.

[0051] Optionally, the clustering attribution model building module may include:

[0052] a bandwidth cluster center sequence determination submodule, configured to determine a predicted bandwidth cluster center sequence and a plurality of other bandwidth cluster center sequences associated with the predicted bandwidth cluster center sequence from the bandwidth fuzzy clustering;

[0053] A Euclidean distance calculation submodule, configured to calculate the Euclidean distance between the predicted bandwidth cluster center sequence and a plurality of the other bandwidth cluster center sequences; the Euclidean distance has a corresponding Euclidean distance parameter; the Euclidean distance parameter corresponds one-to-one to the other bandwidth cluster center sequences;

[0054] a target bandwidth cluster center sequence determination submodule, configured to select a minimum Euclidean distance parameter from the Euclidean distance parameters, and determine the other bandwidth cluster center sequence corresponding to the minimum Euclidean distance parameter as the target bandwidth cluster center sequence;

[0055] a target bandwidth fuzzy clustering determination submodule, configured to determine target bandwidth fuzzy clustering based on the target bandwidth cluster center sequence;

[0056] The prediction data clustering attribution model establishment submodule is used to establish the prediction data clustering attribution model based on the target bandwidth fuzzy clustering.

[0057] Optionally, the predicted data cluster attribution model may have corresponding predicted data cluster attribution data, and the bandwidth prediction model generation module includes:

[0058] An initialization matrix generation submodule is used to generate an initialization matrix for the clustering attribution data of the prediction data through a mapping formula; the initialization matrix contains a plurality of genes for expressing the bandwidth fuzzy clustering;

[0059] An initial optimal gene determination submodule, configured to determine a group fitness value through the initialization matrix and determine an initial optimal gene from the genes;

[0060] a maximum deviation distance parameter determination submodule, configured to determine a maximum deviation distance parameter based on the initialization matrix, and generate a similarity time point number parameter and a maximum deviation time point number parameter according to the maximum deviation distance parameter in accordance with a preset standard;

[0061] A target optimal gene determination submodule is used to determine the target optimal gene based on the initial optimal gene, the neural network initialization parameter, the similar time point number parameter, the maximum deviation time point number parameter and the predicted data clustering attribution data;

[0062] A second parameter determination submodule is used to determine initial weight parameters and threshold position parameters for a neural network based on the target optimal gene;

[0063] The bandwidth prediction model generation submodule is used to train a neural network according to the initial weight parameters, the threshold position parameters and the prediction data clustering attribution data to generate a bandwidth prediction model.

[0064] Optionally, the neural network initialization parameter may include an input layer number parameter, and the input layer number parameter is determined according to the number of timestamps.

[0065] Optionally, the bandwidth prediction result generation module may include:

[0066] The submodule for generating data to be predicted is used to normalize the source data to be predicted and generate the data to be predicted;

[0067] A prediction value data generation submodule, configured to train the data to be predicted based on the bandwidth prediction model to obtain prediction value data;

[0068] The bandwidth prediction result generating submodule is used to perform denormalization processing on the predicted value data to generate a bandwidth prediction result.

[0069] An embodiment of the present invention further discloses an electronic device, comprising a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus;

[0070] The memory is used to store computer programs;

[0071] The processor is configured to implement the method described in the embodiment of the present invention when executing the program stored in the memory.

[0072] An embodiment of the present invention further discloses a computer-readable storage medium having instructions stored thereon. When executed by one or more processors, the processors are enabled to execute the method according to the embodiment of the present invention.

[0073] The embodiments of the present invention include the following advantages:

[0074] An embodiment of the present invention obtains historical bandwidth usage data and generates a timestamp corresponding to the historical bandwidth usage data; generates multiple historical data time series for the historical bandwidth usage data based on the timestamp; divides the historical bandwidth usage data into multiple bandwidth fuzzy clusters through the multiple historical data time series; establishes a prediction data cluster attribution model based on the multiple bandwidth fuzzy clusters; determines neural network initialization parameters, and generates a bandwidth prediction model through the prediction data cluster attribution model and the neural network initialization parameters; generates a bandwidth prediction result based on the bandwidth prediction model, thereby improving the efficiency and accuracy of bandwidth data prediction, reducing the maintenance cost of underlying equipment, and further improving the normal brightness rate of pass codes. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1This is a flowchart of a method for generating bandwidth data prediction results provided in an embodiment of the present invention;

[0076] Figure 2 is a flowchart of another method for generating bandwidth data prediction results provided in an embodiment of the present invention;

[0077] Figure 3 This is a structural block diagram of a bandwidth data prediction result generating device provided in an embodiment of the present invention;

[0078] Figure 4 This is a hardware structure block diagram of an electronic device provided in each embodiment of the present invention. DETAILED DESCRIPTION

[0079] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0080] The pass code has been promoted in many cities and counties, so the sustainable and normal operation of the pass code will play an important role in maintaining social order. In the normal operation of the pass code business, it is inseparable from the cloud device server and the underlying storage device, and the normal operation of the underlying storage device depends on bandwidth data. Bandwidth data is an important indicator for regulating the underlying storage device. It determines whether the underlying storage device can operate normally under different circumstances. The bandwidth data size is positively correlated with the number of real-time visitors. The number of visitors will vary due to factors such as weather, holidays, and time periods (such as rush hour). For example, rush hour is the peak period for showing the code, and the number of visitors will increase. Early morning is the low period for showing the code, and the number of visitors will decrease. Therefore, during the peak period for showing the code, higher bandwidth is required to support the normal operation of the pass code business. During the low period for showing the code, lower bandwidth can meet the demand. Since the usage of pass codes is affected by many factors, operators and operation and maintenance personnel need to invest a lot of manpower and material resources to monitor the bandwidth of underlying equipment in real time. However, since the factors affecting the usage of pass code bandwidth are complex and changeable, manual maintenance methods require more manpower to accurately predict and process bandwidth usage, which will also incur more costs. In addition, due to the limited computing power of humans, it is still difficult to accurately predict the usage of pass code bandwidth. In an embodiment of the present invention, a neural network algorithm and a clustering attribution model are combined to predict bandwidth data to improve the efficiency and accuracy of bandwidth data prediction.

[0081] Reference Figure 1 , shows a flowchart of a method for generating bandwidth data prediction results provided in an embodiment of the present invention, which may specifically include the following steps:

[0082] Step 101: Acquire historical bandwidth usage data and generate a timestamp corresponding to the historical bandwidth usage data;

[0083] Step 102: Generate multiple historical data time series for the historical bandwidth usage data based on the timestamp;

[0084] Step 103: dividing the historical bandwidth usage data into a plurality of bandwidth fuzzy clusters based on a plurality of the historical data time series;

[0085] Step 104: establishing a prediction data clustering attribution model based on the plurality of bandwidth fuzzy clusters;

[0086] Step 105, determining neural network initialization parameters, and generating a bandwidth prediction model using the predicted data clustering attribution model and the neural network initialization parameters;

[0087] Step 106: Generate a bandwidth prediction result based on the bandwidth prediction model.

[0088] In practical applications, bandwidth refers to a computer network's ability to transmit data between devices or across the internet within a specific timeframe. It can be understood as the maximum data transmission rate over any given path. Fluctuations in bandwidth demand reflect changes in the number of users using passcodes and also provide a basis for maintaining and adjusting the underlying passcode equipment. Therefore, accurate bandwidth forecasting can effectively enhance foresight of passcode usage, enabling accurate equipment adjustments and maintenance, reducing equipment failure rates and avoiding peak traffic congestion.

[0089] In a specific implementation, an embodiment of the present invention can obtain historical bandwidth usage data and generate a timestamp corresponding to the historical bandwidth usage data, and generate multiple historical data time series for the historical bandwidth usage data based on the timestamp. For example, the historical bandwidth usage data may include the average pass code usage bandwidth e1, the maximum pass code usage bandwidth e2, the minimum pass code usage bandwidth e3, the daily pass code usage bandwidth e4, the ambient temperature e5, the ambient humidity e6, the holiday conditions e7, the weather conditions e8 and other data information two years before the prediction date and the prediction day. The average pass code usage bandwidth e1, the maximum pass code usage bandwidth e2, the minimum pass code usage bandwidth e3, and the daily pass code usage bandwidth e4 two years before the prediction date and the prediction day can be obtained based on the pass code node database, and the ambient temperature Data information such as e5, ambient humidity e6, holiday conditions e7, and weather conditions e8 can be obtained from the weather forecast website. Then, a time resolution can be set for the obtained historical bandwidth usage data, wherein the time resolution can be a scale in time units, used to generate a timestamp corresponding to the historical bandwidth usage data, and the time unit can be a unit time such as microseconds, milliseconds, centiseconds, minute seconds, seconds, minutes, hours, days, weeks, months, and years. For example, the time resolution for the historical bandwidth usage data can be set to 5 minutes to form historical bandwidth usage data with a time interval of 5 minutes. The historical bandwidth usage data of one day can generate 292 timestamps, wherein each timestamp has corresponding historical bandwidth usage data, which can include the average pass code usage for two years before the forecast day and the forecast day. Maximum passcode usage Minimum passcode usage Daily passcode usage Data information such as ambient temperature e5, ambient humidity e6, holiday conditions e7, and weather conditions e8 can then be sorted in chronological order from early to late according to timestamps to generate multiple sequences containing historical bandwidth usage data as historical data time series.

[0090] Of course, the above examples are only for illustration. Those skilled in the art may obtain historical bandwidth usage data from other local and / or Internet databases, and the embodiments of the present invention do not limit this. Those skilled in the art may use any time unit as the time resolution to generate a timestamp, and the embodiments of the present invention do not limit this.

[0091] In an embodiment of the present invention, historical bandwidth usage data is acquired, timestamps corresponding to the historical bandwidth usage data are generated, and multiple historical data time series for the historical bandwidth usage data are generated based on the timestamps. This achieves associating and segmenting the historical bandwidth usage data with time, assigning a time dimension identifier to the historical bandwidth usage data, and sorting the historical bandwidth usage data according to certain rules based on the timestamps to generate multiple historical data time series for the historical bandwidth usage data, thereby further facilitating subsequent data processing.

[0092] In a specific implementation, the embodiment of the present invention can divide the historical bandwidth usage data into multiple bandwidth fuzzy clusters through multiple historical data time series.

[0093] In practical applications, fuzzy clustering is an analytical method that clusters objective objects based on their characteristics, closeness, and similarity by establishing fuzzy similarity relationships. Fuzzy clustering can effectively improve data accuracy and reduce data processing volume. Embodiments of the present invention can use fuzzy clustering to divide historical bandwidth usage data through multiple historical data time series to determine multiple bandwidth fuzzy clusters.

[0094] The embodiment of the present invention divides historical bandwidth usage data into multiple bandwidth fuzzy clusters through multiple historical data time series, thereby achieving efficient clustering of historical bandwidth usage data, thereby reducing the data processing volume and improving data accuracy, laying a data foundation for improving the efficiency of subsequent data analysis and the accuracy of generating bandwidth prediction results.

[0095] In a specific implementation, the embodiment of the present invention can determine the neural network initialization parameters, and generate a bandwidth prediction model by predicting the data clustering attribution model and the neural network initialization parameters, and then generate a bandwidth prediction result based on the bandwidth prediction model.

[0096] In practical applications, the Elman neural network is a typical local regression network (Global FeedForward Local Recurrent), which can be regarded as a recursive neural network with local memory units and local feedback connections. It is based on the basic structure of the BP network (Back Propagation) and adds a follow-up layer to the hidden layer as a one-step delay operator to achieve the purpose of memory, thereby enabling the system to adapt to time-varying characteristics and enhance the global stability of the network. It has stronger computing power than feedforward neural networks and can also be used to solve fast optimization problems.

[0097] For example, after obtaining multiple bandwidth fuzzy clusters based on fuzzy clustering, the bandwidth fuzzy cluster with the smallest Euclidean distance can be selected to establish a cluster attribution model, where the Euclidean distance is also called the Euclidean distance or Euclidean metric, which is a commonly used distance definition and is the true distance between two points in m-dimensional space. Then, the neural network initialization parameters of the Elman neural network can be determined based on the timestamp corresponding to the historical bandwidth data used in the bandwidth fuzzy clustering, and then a bandwidth prediction model is generated using the cluster attribution model and the neural network initialization parameters. Finally, a bandwidth prediction result is generated based on the bandwidth prediction model.

[0098] An embodiment of the present invention obtains historical bandwidth usage data and generates a timestamp corresponding to the historical bandwidth usage data; generates multiple historical data time series for the historical bandwidth usage data based on the timestamp; divides the historical bandwidth usage data into multiple bandwidth fuzzy clusters through the multiple historical data time series; establishes a prediction data cluster attribution model based on the multiple bandwidth fuzzy clusters; determines neural network initialization parameters, and generates a bandwidth prediction model through the prediction data cluster attribution model and the neural network initialization parameters; generates a bandwidth prediction result based on the bandwidth prediction model, thereby improving the efficiency and accuracy of bandwidth data prediction, reducing the maintenance cost of underlying equipment, and further improving the normal brightness rate of pass codes.

[0099] Based on the above embodiment, a modified embodiment of the above embodiment is proposed. It should be noted that, in order to simplify the description, only the differences from the above embodiment are described in the modified embodiment.

[0100] In an optional embodiment of the present invention, the step of dividing the historical bandwidth usage data into a plurality of bandwidth fuzzy clusters using a plurality of the historical data time series includes:

[0101] Determining a category number parameter, a first fuzzy membership parameter, and a first fuzzy weight index parameter;

[0102] Determining corresponding initial cluster centers from the plurality of historical data time series based on the first fuzzy membership parameter and the first fuzzy weight index parameter according to a first preset algorithm;

[0103] Determining a second fuzzy weight index based on the category number parameter, the initial cluster center, and the first fuzzy weight index according to a second preset algorithm;

[0104] Determining a second fuzzy membership parameter based on the second fuzzy weight index, the initial cluster center, and the category number parameter according to a third preset algorithm;

[0105] Determining a target cluster center based on the initial cluster center, the second fuzzy weight index, the second fuzzy membership parameter, and the category number parameter; the target cluster center has a corresponding historical data time series;

[0106] The historical data time series is divided into a plurality of bandwidth fuzzy clusters according to the target cluster center and the category number parameter.

[0107] In a specific implementation, the embodiment of the present invention can determine a category number parameter, a first fuzzy membership parameter, and a first fuzzy weight index parameter, and determine the corresponding initial cluster center from multiple historical data time series based on the first fuzzy membership parameter and the first fuzzy weight index parameter according to a first preset algorithm. For example, the category number parameter can be c, the first fuzzy membership parameter can be u ij , the first fuzzy weight index parameter is m, the initial cluster center is v j , where u ij is the historical data time series x i The fuzzy membership parameter of the jth class, v j is the initial cluster center of the jth class, multiple historical data time series x i The J category can be formed by the following formula:

[0108] Formula 1:

[0109]

[0110] Among them, based on the first fuzzy membership parameter u ij and the first fuzzy weight index parameter m to determine the corresponding initial cluster center v from multiple historical data time series j The calculation formula is:

[0111] Formula 2:

[0112]

[0113] In a specific implementation, the embodiment of the present invention can determine the second fuzzy weight index based on the category number parameter, the initial cluster center and the first fuzzy weight index according to the second preset algorithm. For example, when the category number parameter is c and the initial cluster center is v j , when the first fuzzy weight index is m, the second fuzzy weight index m can be determined by the following formula:

[0114] Formula 3:

[0115]

[0116] In a specific implementation, the embodiment of the present invention can determine the second fuzzy membership parameter based on the second fuzzy weight index, the initial cluster center and the category number parameter according to the third preset algorithm; for example, when the second fuzzy weight index is m and the initial cluster center is v j When the number of categories parameter is c, the second fuzzy membership parameter u can be determined by the following formula ij :

[0117] Formula 4:

[0118]

[0119] For example, the initial cluster centers can be updated simultaneously according to the following formula:

[0120] Formula 5:

[0121] In a specific implementation, the embodiment of the present invention can be based on the initial cluster center, the second fuzzy weight index

[0122]

[0123] The target cluster center is determined by the number, the second fuzzy membership parameter and the number of categories parameter; the target cluster center has a corresponding historical data time series. For example, when the initial cluster center is v j , the historical data time series is x i , the second fuzzy weight index is m, the second fuzzy membership parameter is u ij , when the number of categories parameter is c, the target cluster center can be determined by the following formula Among them, c i For the class center of class c:

[0124] Formula 6:

[0125]

[0126] For example, if the fuzzy membership parameter u ij satisfy And / or, when the preset number of iterations is reached, the iteration ends and c categories are obtained. If the above conditions are not met, the second fuzzy weight index can be determined again according to the second preset algorithm based on the category number parameter, the initial cluster center and the first fuzzy weight index; the second fuzzy membership parameter can be determined according to the third preset algorithm based on the second fuzzy weight index, the initial cluster center and the category number parameter until the target cluster center is generated. Among them, the target cluster center has a corresponding historical data time series.

[0127] In a specific implementation, the embodiment of the present invention can divide the historical data time series into multiple bandwidth fuzzy clusters according to the target cluster center and the number of categories. For example, when the target cluster center is The number of categories is c, then the target cluster center can be and the number of categories parameter c will correspond to the target cluster center The historical data time series are divided into bandwidth fuzzy clustering.

[0128] In an embodiment of the present invention, by determining a category number parameter, a first fuzzy membership parameter and a first fuzzy weight index parameter; determining a corresponding initial cluster center from a plurality of historical data time series based on the first fuzzy membership parameter and the first fuzzy weight index parameter according to a first preset algorithm; determining a second fuzzy weight index based on the category number parameter, the initial cluster center and the first fuzzy weight index according to a second preset algorithm; determining a second fuzzy membership parameter based on the second fuzzy weight index, the initial cluster center and the category number parameter according to a third preset algorithm; determining a target cluster center based on the initial cluster center, the second fuzzy weight index, the second fuzzy membership parameter and the category number parameter; the target cluster center has a corresponding historical data time series; and dividing the historical data time series into a plurality of bandwidth fuzzy clusters according to the target cluster center and the category number parameter. The historical data time series can be divided and clustered through fuzzy clustering, which improves the efficiency of clustering. At the same time, by calculating the fuzzy membership parameters and fuzzy weight index, the problem of traditional fuzzy clustering algorithms having no definite clustering results and needing to be re-analyzed in combination with relevant data is solved, which improves the accuracy and efficiency of clustering and further provides data support for subsequent improvements in bandwidth prediction efficiency and accuracy.

[0129] In an optional embodiment of the present invention, the step of establishing a prediction data cluster attribution model based on the plurality of bandwidth fuzzy clusters includes:

[0130] Determining a predicted bandwidth cluster center sequence and a plurality of other bandwidth cluster center sequences associated with the predicted bandwidth cluster center sequence from the bandwidth fuzzy clustering;

[0131] Calculating the Euclidean distance between the predicted bandwidth cluster center sequence and the plurality of other bandwidth cluster center sequences; the Euclidean distance has a corresponding Euclidean distance parameter; the Euclidean distance parameter corresponds one-to-one to the other bandwidth cluster center sequences;

[0132] Selecting a minimum Euclidean distance parameter from the Euclidean distance parameters, and determining the other bandwidth cluster center sequence corresponding to the minimum Euclidean distance parameter as the target bandwidth cluster center sequence;

[0133] Determining target bandwidth fuzzy clustering based on the target bandwidth cluster center sequence;

[0134] The predicted data clustering attribution model is established based on the target bandwidth fuzzy clustering.

[0135] In a specific implementation, the embodiment of the present invention can determine the predicted bandwidth cluster center sequence and multiple other bandwidth cluster center sequences associated with the predicted bandwidth cluster center sequence from bandwidth fuzzy clustering, and calculate the Euclidean distance between the predicted bandwidth cluster center sequence and the multiple other bandwidth cluster center sequences; the Euclidean distance has a corresponding Euclidean distance parameter; the Euclidean distance parameter has a one-to-one correspondence with other bandwidth cluster center sequences. For example, it can be determined that the pass code uses bandwidth data x i To predict the bandwidth cluster center sequence, other passcodes use bandwidth data x j For other bandwidth cluster center sequences, the predicted bandwidth cluster center sequence x can be calculated by the following formula i and multiple other bandwidth cluster center sequences x j The Euclidean distance x ijc :

[0136] Formula 7:

[0137] x ijk =||x ik -x jk ||

[0138] Where i, j = 1, 2, ..., n, k = 1, 2, ..., m; x ik is the value of the i-th predicted data at the k-th time point; x jk is the value of the jth cluster center at the kth time point, and the Euclidean distance x ijc With the corresponding Euclidean distance parameter x ijc , and with other bandwidth cluster center sequences x j One to one correspondence.

[0139] In a specific implementation, the embodiment of the present invention can select the minimum Euclidean distance parameter from the Euclidean distance parameters, determine other bandwidth cluster center sequences corresponding to the minimum Euclidean distance parameter as the target bandwidth cluster center sequence; determine the target bandwidth fuzzy clustering based on the target bandwidth cluster center sequence; establish a prediction data clustering attribution model based on the target bandwidth fuzzy clustering, illustratively, by comparing the Euclidean distance x ijc , select the minimum Euclidean distance x ijc , which will correspond to the minimum Euclidean distance x ijc The other bandwidth cluster center sequences c are used as target bandwidth fuzzy clusters, and a prediction data clustering attribution model is established based on the target bandwidth fuzzy clusters c.

[0140] The embodiment of the present invention determines a predicted bandwidth cluster center sequence and multiple other bandwidth cluster center sequences associated with the predicted bandwidth cluster center sequence from the bandwidth fuzzy clustering; calculates the Euclidean distance between the predicted bandwidth cluster center sequence and the multiple other bandwidth cluster center sequences; the Euclidean distance has a corresponding Euclidean distance parameter; the Euclidean distance parameter corresponds one-to-one with the other bandwidth cluster center sequences; selects the minimum Euclidean distance parameter from the Euclidean distance parameters, determines the other bandwidth cluster center sequence corresponding to the minimum Euclidean distance parameter as the target bandwidth cluster center sequence; determines the target bandwidth fuzzy clustering based on the target bandwidth cluster center sequence; and establishes the predicted data cluster attribution model based on the target bandwidth fuzzy clustering. This achieves the establishment of the predicted data cluster attribution model and establishes a data foundation for subsequent data calculation and bandwidth data prediction.

[0141] In an optional embodiment of the present invention, the step of generating a bandwidth prediction model by using the prediction data clustering attribution model and the neural network initialization parameters includes:

[0142] Generating an initialization matrix for the clustering belonging data of the predicted data through a mapping formula; the initialization matrix includes a plurality of genes for expressing the bandwidth fuzzy clustering;

[0143] Determining a group fitness value through the initialization matrix, and determining an initial optimal gene from the genes;

[0144] Determining a maximum deviation distance parameter based on the initialization matrix, and generating a similarity time point number parameter and a maximum deviation time point number parameter according to the maximum deviation distance parameter in accordance with a preset standard;

[0145] Determine the target optimal gene based on the initial optimal gene, the neural network initialization parameter, the similarity time point number parameter, the maximum deviation time point number parameter and the predicted data clustering attribution data;

[0146] Determining initial weight parameters and threshold position parameters for a neural network based on the target optimal gene;

[0147] A neural network is trained according to the initial weight parameters, the threshold position parameters and the predicted data clustering attribution data to generate a bandwidth prediction model.

[0148] In an optional embodiment of the present invention, the neural network initialization parameter includes an input layer number parameter, and the input layer number parameter is determined according to the number of timestamps.

[0149] In practical applications, the traditional Elman neural network has a technical defect that it is easy to fall into the local optimal solution during the calculation process. That is, when there is a local optimal solution, if the search step size is small, all solutions may point to the optimal solution method during this local solution, which will cause the Elman neural network training to stagnate. In addition, the Elman neural network also has shortcomings such as slow convergence speed, low prediction accuracy, and long training time. Therefore, this problem should be considered when using the Elman neural network to train data.

[0150] In a specific implementation, the prediction data clustering attribution model may have corresponding prediction data clustering attribution data, the neural network initial parameters may include an input layer number parameter, a hidden layer number parameter, an output layer number parameter, and a convergence error parameter. The neural network initialization parameters may be determined based on the timestamp. For example, when the number of timestamps is 292, the input layer number parameter of the neural network initialization parameter may be determined to be 292, and the corresponding hidden layer parameter may be determined to be 25, the output layer parameter may be determined to be 288, and the convergence error parameter may be determined to be 10*-6. At the same time, the genetic algorithm parameters may be initialized, wherein the genetic algorithm parameters may include a maximum number of iterations parameter, a population size parameter, a mutation probability parameter, a maximum deviation coefficient parameter, a maximum error parameter, and the like. For example, the maximum number of iterations parameter may be set to 10,000, the population size parameter may be set to 20, the mutation probability parameter may be set to 0.09, and the maximum deviation coefficient parameter may be set to 0.00001. When the prediction data clustering attribution model is C, the prediction data clustering attribution model has corresponding prediction data clustering attribution data X. ij For example, a chaotic sequence can be generated by a mapping formula and processed by a chaotic algorithm to predict the data clustering belonging to the data X ij Mapping to the solution space to construct an initialization matrix, where the solution space refers to the set of all solutions of the homogeneous linear equations forming a vector space, that is, a set. The chaos algorithm refers to the chaotic sequence encryption algorithm. The algorithm first uses a one-way hash function to hash the key into the iterative initial value of the chaotic map. The chaotic sequence is not used until several iterations have passed. Then, the iteratively generated chaotic sequence value is mapped to ASCII code and then XORed with the map data byte by byte. Taking into account the limited precision effect in actual calculations, the chaotic mapping parameters are changed with the step size. Using actual map data, the chaos algorithm can quickly find the global coverage optimal value. For example, a 1*D-dimensional random matrix can be randomly generated in the interval (0, 1) as particle P1. The cubic mapping formula can be used as the mapping formula. The cubic mapping is used to generate a chaotic sequence for each number in particle P1 to obtain N initialized particles. The mapping formula is as follows:

[0151] Formula 8:

[0152] x n+1 =4(x n )3 -3x n

[0153] The predicted data can be clustered into the data X by the following formula ij The chaotic sequence composed is mapped to the solution space to construct the initialization matrix:

[0154] Formula 9:

[0155]

[0156] x i =z i (x imax -x imin )+x imin

[0157]

[0158] The initialization matrix can be an N*D matrix B:

[0159] Among them, x ij It can be used as the genes in the initialization matrix for clustering IOPS data, x ij Represents the j-th dimension value of the i-th gene, i=1...n, j=1...d.

[0160] The population fitness value is calculated based on the matrix B. The population fitness refers to the relative ability of an individual with a known genotype to pass its genes to the gene pool of its offspring under certain environmental conditions. It is a measure of the individual's chances of survival and reproduction. By calculating the population fitness value, the optimal gene can be selected. Then, a selection operation can be performed. For example, the roulette method can be used to calculate the selection probability P of each individual i based on the population fitness value and the fitness value ratio. i The formula is:

[0161] Formula 10:

[0162] f i =k / F i

[0163]

[0164] Among them, Fi is the fitness value of individual i, k is the coefficient, and N is the number of individuals in the population.

[0165] Then, the mutation operation can be performed using the following formula:

[0166] Formula 11:

[0167]

[0168] Among them, amax Gene a ij The upper bound of a min Gene a ij The lower bound of f(g) = r2(1-g / G max ) 2 ; r2 is a random number; g is the current number of iterations; G max is the maximum number of evolutions; r is a random number between [0, 1].

[0169] Then, the maximum deviation distance parameter λ can be determined based on the initialization matrix B, as follows:

[0170] Formula 12:

[0171]

[0172] Among them, ω i is the weight of the i-th individual, x imax is the largest value among the i-th individual; x imin is the minimum value among the i-th individual; is the average value of the i-th individual, x jmax is the maximum value among the global optimal individuals, x jmin is the global optimal individual minimum value, is the average value of the global optimal individual; Φ i is the deviation, which is a random number in (0, 0.1).

[0173] Then, the similarity time point number parameter n can be generated according to the preset standard based on the maximum deviation distance parameter λ ij and the maximum deviation point number parameter m ij , when it is known that the initialization matrix B after the selection operation and mutation operation is:

[0174]

[0175] The maximum deviation distance between each pair of x in the initialization matrix B can be calculated based on the maximum deviation distance parameter λ. If X ijk The number of ≦γ is n ij , suppose that X ijk >The number of γ is m ij , it can be calculated according to the following formula:

[0176] Formula 13:

[0177]

[0178] Among them, i, j = 1, 2, ..., n; k = 1, 2, ..., m.

[0179] After that, we can judge whether the gene meets the maximum deviation similarity criterion. We can encode the population according to the maximum deviation similarity criterion. If the criterion is met, it is coded as 1, and if it is not met, it is coded as 0. The criteria for judging whether the gene meets the maximum deviation similarity criterion are as follows:

[0180]

[0181] Where n0 is [α×β], α (0≤α≤1); m0 is [β×m], β (0≤β≤1-α); for example, α can be set to 0.9 and β can be set to 0.1. If the gene meets the above conditions, the gene is subjected to chaotic processing.

[0182] Then, it can be determined whether the stopping criterion is met. If so, stop. If not, repeat the self-generation of similar time point number parameters and maximum deviation time point number parameters until it is determined again whether the stopping criterion is met. The calculated optimal gene is used as the target optimal gene.

[0183] After the iteration is completed, the initial weight parameters and threshold position parameters for the neural network are determined based on the target optimal gene. For example, the optimal gene can be split to generate initial weight parameters and threshold position parameters. The initial weight parameters and threshold position parameters can be the initial weights and initial thresholds for the Elman neural network. Then, for example, the Elman neural network can be trained based on the initial weight parameters, threshold position parameters, and the predicted data cluster attribution data, and the generated model can be used as a bandwidth prediction model.

[0184] The embodiment of the present invention generates an initialization matrix for the predicted data clustering belonging data through a mapping formula; the initialization matrix includes multiple genes for expressing the bandwidth fuzzy clustering; a group fitness value is determined through the initialization matrix, and an initial optimal gene is determined from the genes; a maximum deviation distance parameter is determined based on the initialization matrix, and a similarity time point number parameter and a maximum deviation time point number parameter are generated according to the maximum deviation distance parameter according to a preset standard; a target optimal gene is determined based on the initial optimal gene, the neural network initialization parameter, the similarity time point number parameter, the maximum deviation time point number parameter, and the predicted data clustering belonging data; an initial weight parameter and a threshold position parameter for the neural network are determined based on the target optimal gene; the neural network is trained according to the initial weight parameter, the threshold position parameter, and the predicted data clustering belonging data to generate a bandwidth prediction model, wherein the neural network initialization parameter includes an input layer number parameter, and the input layer number parameter is determined according to the number of timestamps, thereby solving the problems of the Elman neural network easily falling into a local optimum point, slow convergence speed, low prediction accuracy, and long training time during prediction, and further improving the feasibility and accuracy of bandwidth prediction.

[0185] In an optional embodiment of the present invention, the step of generating a bandwidth prediction result based on the bandwidth prediction model includes:

[0186] Normalize the source data to be predicted to generate the data to be predicted;

[0187] Training the data to be predicted based on the bandwidth prediction model to obtain predicted value data;

[0188] Denormalization is performed on the predicted value data to generate a bandwidth prediction result.

[0189] In a specific implementation, the bandwidth usage data of the pass code on the day to be predicted can be used as the source data to be predicted, the bandwidth usage data of the pass code on the day to be predicted can be normalized, the normalized bandwidth usage data of the pass code can be used as the data to be predicted, the normalized bandwidth usage data of the pass code can be put into the prediction model, and the obtained predicted value can be used as the predicted value data. The predicted value data can be denormalized and the actual predicted value of the bandwidth usage of the pass code can be used as the bandwidth prediction result.

[0190] The embodiment of the present invention generates the data to be predicted by normalizing the source data to be predicted; training the data to be predicted based on a bandwidth prediction model to obtain predicted value data; and performing denormalization processing on the predicted value data to generate a bandwidth prediction result, thereby achieving the generation of a bandwidth prediction result based on the bandwidth prediction model, thereby improving the accuracy and efficiency of predicting the bandwidth data.

[0191] In order to enable those skilled in the art to better understand the embodiments of the present invention, the embodiments of the present invention are described below using a complete example.

[0192] The widespread use of passcodes has put increasing pressure on servers, hard drives, and other devices that handle passcode services. The status of the underlying devices directly affects passcode usage. Within the underlying hardware, bandwidth usage is a key indicator of device performance. If passcode operations and maintenance personnel can adjust bandwidth before peak passcode usage and adjust bandwidth based on passcode usage during different time periods, they can avoid peak passcode congestion. Therefore, accurate passcode bandwidth usage prediction can greatly improve the perception of passcode usage and effectively avoid peak congestion. At the same time, passcode bandwidth usage prediction can also allow operations and maintenance personnel to perceive the current status of the device, make predictions in advance, reduce the device's failure rate, and thereby increase the normal passcode brightness rate.

[0193] As the speed of human mobility increases, the volume of passcode traffic is growing. The factors that influence passcode bandwidth prediction are also diverse, including social, political, weather, and even economic factors. Therefore, it is necessary to conduct accurate and reasonable research on passcode bandwidth prediction.

[0194] However, accurate and rapid advance prediction of passcode bandwidth is currently not possible. Furthermore, the traditional Elman neural network algorithm is often used for prediction, which suffers from local minima, slow convergence, low prediction accuracy, and long training times.

[0195] refer to Figure 2 , Figure 2 is a flowchart of another method for generating bandwidth data prediction results provided in an embodiment of the present invention;

[0196] To achieve the above object, the present invention provides the following technical solution: a method for predicting and calculating the bandwidth used by a pass code using an improved Elman neural network, comprising the following steps:

[0197] Step S1, constructing a pass code bandwidth usage dataset;

[0198] When constructing the passcode bandwidth usage dataset, a cluster dataset and a forecast day dataset are constructed using historical passcode bandwidth usage data. The historical passcode bandwidth usage data includes the average passcode bandwidth e1, maximum passcode bandwidth e2, minimum passcode bandwidth e3, daily passcode bandwidth e4, ambient temperature e5, ambient humidity e6, holiday conditions e7, and weather conditions e8 for the two years prior to the forecast day and the forecast day. The cluster dataset includes historical passcode bandwidth usage data other than the forecast day data. The forecast day dataset contains the forecast day data to be predicted.

[0199] Specifically, in this embodiment of the present invention, historical passcode bandwidth usage data is obtained from the passcode node database, while ambient temperature, humidity, holidays, and weather conditions are obtained from local weather forecast websites. The time resolution of the passcode bandwidth usage data is set to 5 minutes, forming a passcode bandwidth usage curve with a 5-minute interval. Holidays are categorized as ordinary weekends, major holidays, and the day after a holiday, and weather conditions are categorized as sunny, rainy, and cloudy. Therefore, a single-day passcode bandwidth usage data curve consists of 292 data points (each data point includes a time point and the corresponding passcode bandwidth usage).

[0200] Step S2, constructing the pass code using bandwidth fuzzy c-means clustering center;

[0201] Specifically, the step of constructing the pass code using bandwidth fuzzy c-means clustering center may include:

[0202] Construct the fuzzy clustering objective function.

[0203] Specifically,

[0204] Among them, u ij For individual x i The fuzzy membership degree of the jth class; m is the fuzzy weight index; v j is the cluster center of the jth class; u ij and v j The calculation formula is:

[0205]

[0206] Initialize the cluster center, set the number of categories, and the fuzzy weight index.

[0207] Specifically, the number of categories c and the fuzzy weight index m are set. In this example, the number of categories c is set to 5, and the calculation formula of the fuzzy weight index m is as follows:

[0208]

[0209] Compute the fuzzy membership matrix.

[0210] Among them, the fuzzy membership matrix formula is calculated as follows:

[0211]

[0212] Computing Center.

[0213] The calculation of the class center is as follows:

[0214]

[0215] Calculate the fuzzy clustering target value and determine whether the end condition is met. If so, the algorithm terminates. Otherwise, return to the step to calculate the fuzzy membership matrix.

[0216] Specifically, if Or if the number of iterations is reached, the iteration ends and c categories are obtained, that is, the bandwidth data of historical pass codes are divided into c categories. Otherwise, the process returns to the step of calculating the fuzzy membership matrix.

[0217] Calculate the cluster centers.

[0218] Specifically, the cluster center is calculated as follows:

[0219]

[0220] Where c i It is the class center of class c.

[0221] Step S3, constructing a clustering attribution model for prediction data;

[0222] Specifically, the steps of constructing a prediction data clustering attribution model may include:

[0223] Constructing the predicted passcode uses the Euclidean distance between the bandwidth data and the corresponding cluster center.

[0224] Specifically, the predicted passcode is constructed using bandwidth data x i With each cluster center x j The corresponding Euclidean distance x ijc The calculation formula of Euclidean distance is as follows:

[0225] x ijk =||x ik -x jk ||

[0226] Where i, j = 1, 2, ..., n; k = 1, 2, ..., m; x ik is the value of the i-th predicted data at the k-th time point; x jk is the value of the j-th cluster center at the k-th time point.

[0227] Compare the Euclidean distances and select the smallest corresponding cluster c to classify the predicted data as class c data.

[0228] Specifically, compare the Euclidean distance x ijc , select the minimum x ijc The corresponding cluster c classifies the predicted data as c-type data.

[0229] Step S4, constructing a pass code bandwidth prediction model;

[0230] Specifically, the step of constructing a pass code bandwidth prediction model may include:

[0231] Initialize the Elman neural network parameters, initialize the genetic algorithm parameters, and initialize the maximum deviation similarity criterion parameters.

[0232] Specifically, the Elman neural network parameters are initialized. These parameters include the number of input layers, the number of hidden layers, the number of output layers, and the convergence error. In this example, the number of input layers is set to 292, the number of hidden layers is set to 25, the number of output layers is set to 288, and the convergence error is set to 10*-6.

[0233] Specifically, initialize the genetic algorithm parameters. These include the maximum number of iterations, population size, mutation probability, maximum deviation coefficient, and maximum error. In this case, the maximum number of iterations is set to 10,000, the population size is set to 50, the mutation probability is set to 0.09, and the maximum error is set to 0.00001.

[0234] The initial genes are processed into chaos. Then, the genes are restored to the solution range through the mapping formula to form the initialization matrix of the improved genetic algorithm, which is represented by matrix B:

[0235]

[0236] where x ij Represents the j-th dimension value of the i-th gene, i=1...n, j=1...d.

[0237] Specifically, the initial population is mapped to the solution space through chaos processing.

[0238] 1) Randomly generate a 1*D dimensional random matrix in the interval (0, 1) as particle P1.

[0239] 2) Generate a chaotic sequence using cubic mapping for each data in particle P1, and obtain N initialized particles.

[0240] The formula for cubic mapping is:

[0241] x n+1 =4(x n ) 3 -3x n

[0242] 3) The chaotic space is mapped to the solution space using the following formula.

[0243] The formula is:

[0244]

[0245] x i =z i (x imax -x imin )+x imin

[0246] 4) The resulting solution space can be represented by an N*D matrix B:

[0247]

[0248] Among them, x ij is the j-th dimension value of the i-th particle.

[0249] Calculate the population fitness value.

[0250] Specifically, the best individual is recorded.

[0251] Choose an action.

[0252] Specifically, in this case, the roulette method is selected based on the fitness ratio selection strategy, and the selection probability of each individual i is P iAs shown in the following formula:

[0253] f i =k / F i

[0254]

[0255] Where, F i is the fitness value of individual i, k is the coefficient, and N is the number of individuals in the population.

[0256] Mutation operation.

[0257] Specifically, the mutation operation adopts the formula:

[0258]

[0259] Among them, amax is the upper bound of gene aij; amin is the lower bound of gene aij; f(g) = r2(1-g / Gmax)2; r2 is a random number; g is the current number of iterations; Gmax is the maximum number of evolutions; r is a random number between [0,1].

[0260] Construct the maximum deviation distance λ.

[0261]

[0262] Specifically, Where, ω i is the weight of the i-th individual, x imax is the largest value among the i-th individual; x imin is the minimum value among the i-th individual; is the average value of the i-th individual, x jmax is the maximum value among the global optimal individuals, x jmin is the global optimal individual minimum value, is the average value of the global optimal individual; φi is the deviation, which is a random number in (0, 0.1).

[0263] Construct the number of similar time points n ij and the maximum deviation point number m ij .

[0264] Specifically, the population after selection, mutation, and operation is:

[0265]

[0266] According to the maximum deviation distance λ calculation formula in the step of constructing the maximum deviation distance λ, the maximum deviation distance between two genes in the population is calculated, assuming that X ijk The number of ≦γ is n ij , set to satisfy Xijk >The number of γ is m ij , the formula is as follows:

[0267]

[0268] Wherein, i, j = 1, 2, ..., n; k = 1, 2, ..., m.

[0269] Determine whether the maximum deviation similarity criterion is met. If so, perform chaos processing on the gene.

[0270] Specifically, the population is coded according to the maximum deviation similarity criterion, and the population that meets the criterion is coded as 1, and the population that does not meet the criterion is coded as 0.

[0271] The criteria for determining whether the maximum deviation similarity criterion is met are as follows: if the following formula is satisfied, the gene is determined to meet the maximum deviation similarity criterion; otherwise, the gene is determined not to meet the maximum deviation similarity criterion:

[0272]

[0273] Where n0 is [α×β], α (0≤α≤1); m0 is [β×m], β (0≤β≤1-α); in this case, α is set to 0.9 and β is set to 0.1.

[0274] Among them, the encoded population can be represented by an N*D dimensional matrix A:

[0275]

[0276] Determine whether the stopping criteria are met. If so, proceed to the next step. If not, return to the step to construct the number of similar time points n. ij and the maximum deviation point number m ij ; The calculated optimal individual is the optimal gene.

[0277] At the end of the iteration, the optimal gene is split and input into the initial weight and threshold position corresponding to the Elman neural network as the initial weight and initial threshold of the Elman neural network.

[0278] Train the Elman neural network to obtain the passcode bandwidth prediction model.

[0279] Specifically, the calculated prediction model attribution data is used as a training data set of the Elman neural network to train the Elman neural network to obtain the prediction data cluster attribution model.

[0280] Step S5: predicting the bandwidth used by the pass code.

[0281] Specifically, the step of predicting the bandwidth used by the pass code may include:

[0282] When predicting the pass code bandwidth usage, the pass code bandwidth usage data of the predicted day is put into the prediction model, the pass code bandwidth usage of the prediction day is predicted, and the obtained prediction value is denormalized to obtain the actual prediction value of the pass code bandwidth usage.

[0283] (1) Normalize the forecast day data.

[0284] (2) The normalized pass code bandwidth is put into the obtained prediction model for training.

[0285] (3) Predict the bandwidth used by the passcode on the predicted day, and perform denormalization on the predicted value to obtain the actual predicted value of the bandwidth used by the passcode.

[0286] An embodiment of the present invention further provides a device for generating bandwidth data prediction results, comprising:

[0287] A construction unit is used to construct a pass code bandwidth usage data set; further used to construct a fuzzy c-means cluster center of the pass code bandwidth usage; further used to construct a prediction data clustering attribution model; further used to construct a pass code bandwidth usage prediction model;

[0288] The prediction unit is used to predict the bandwidth used by the pass code.

[0289] The above method realizes the rapid and accurate prediction of the bandwidth used by the pass code, so as to assist the operation and maintenance personnel of the pass code to adjust and control it before the peak of the pass code usage, and adjust the bandwidth according to the usage of the pass code in different time periods to avoid the congestion of the pass code during the peak period. At the same time, the accurate prediction of the bandwidth used by the pass code can greatly improve the usage perception of the pass code and effectively avoid peak congestion. The prediction of the bandwidth used by the pass code can also enable the operation and maintenance personnel to perceive the current status of the equipment, make predictions in advance, reduce the failure rate of the equipment, and thus improve the normal lighting rate of the pass code. In addition, it also solves the problems of the traditional Elman neural network algorithm in the calculation process, such as local minimum points, slow convergence speed, low prediction accuracy, and long training time; the historical data is classified by the fuzzy C-means clustering algorithm, and the category with the minimum distance between the predicted data and the cluster center is selected as the category to which the predicted data belongs, so as to improve the accuracy of the data and reduce the amount of data processing.

[0290] It should be noted that for the sake of simplicity, the method embodiments are described as a series of actions. However, those skilled in the art should be aware that the embodiments of the present invention are not limited by the order of the actions described, because according to the embodiments of the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in this specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

[0291] Reference Figure 3 , shows a structural block diagram of a bandwidth data prediction result generating device provided in an embodiment of the present invention, which may specifically include the following modules:

[0292] The historical bandwidth usage data acquisition module 301 is used to acquire historical bandwidth usage data and generate a timestamp corresponding to the historical bandwidth usage data;

[0293] A historical data time series generating module 302 is configured to generate a plurality of historical data time series for the historical bandwidth usage data based on the timestamp;

[0294] A bandwidth fuzzy clustering division module 303 is configured to divide the historical bandwidth usage data into a plurality of bandwidth fuzzy clusters based on a plurality of the historical data time series;

[0295] A clustering attribution model building module 304 is configured to build a prediction data clustering attribution model based on the plurality of bandwidth fuzzy clusters;

[0296] The bandwidth prediction model generation module 305 is used to determine the neural network initialization parameters and generate a bandwidth prediction model through the predicted data clustering attribution model and the neural network initialization parameters;

[0297] The bandwidth prediction result generating module 306 is configured to generate a bandwidth prediction result based on the bandwidth prediction model.

[0298] Optionally, the bandwidth fuzzy clustering partitioning module may include:

[0299] A first parameter determination submodule, configured to determine a category number parameter, a first fuzzy membership parameter, and a first fuzzy weight index parameter;

[0300] an initial cluster center determination submodule, configured to determine corresponding initial cluster centers from a plurality of the historical data time series based on the first fuzzy membership parameter and the first fuzzy weight index parameter according to a first preset algorithm;

[0301] A second fuzzy weight index determination submodule, configured to determine a second fuzzy weight index based on the category number parameter, the initial cluster center, and the first fuzzy weight index according to a second preset algorithm;

[0302] a second fuzzy membership parameter determination submodule, configured to determine a second fuzzy membership parameter based on the second fuzzy weight index, the initial cluster center, and the category number parameter according to a third preset algorithm;

[0303] a target cluster center determination submodule, configured to determine a target cluster center based on the initial cluster center, the second fuzzy weight index, the second fuzzy membership parameter, and the category number parameter; the target cluster center having a corresponding historical data time series;

[0304] The bandwidth fuzzy clustering division submodule is used to divide the historical data time series into multiple bandwidth fuzzy clusters according to the target cluster center and the category number parameter.

[0305] Optionally, the clustering attribution model building module may include:

[0306] a bandwidth cluster center sequence determination submodule, configured to determine a predicted bandwidth cluster center sequence and a plurality of other bandwidth cluster center sequences associated with the predicted bandwidth cluster center sequence from the bandwidth fuzzy clustering;

[0307] A Euclidean distance calculation submodule, configured to calculate the Euclidean distance between the predicted bandwidth cluster center sequence and a plurality of the other bandwidth cluster center sequences; the Euclidean distance has a corresponding Euclidean distance parameter; the Euclidean distance parameter corresponds one-to-one to the other bandwidth cluster center sequences;

[0308] a target bandwidth cluster center sequence determination submodule, configured to select a minimum Euclidean distance parameter from the Euclidean distance parameters, and determine the other bandwidth cluster center sequence corresponding to the minimum Euclidean distance parameter as the target bandwidth cluster center sequence;

[0309] a target bandwidth fuzzy clustering determination submodule, configured to determine target bandwidth fuzzy clustering based on the target bandwidth cluster center sequence;

[0310] The prediction data clustering attribution model establishment submodule is used to establish the prediction data clustering attribution model based on the target bandwidth fuzzy clustering.

[0311] Optionally, the predicted data cluster attribution model may have corresponding predicted data cluster attribution data, and the bandwidth prediction model generation module includes:

[0312] An initialization matrix generation submodule is used to generate an initialization matrix for the clustering attribution data of the prediction data through a mapping formula; the initialization matrix contains a plurality of genes for expressing the bandwidth fuzzy clustering;

[0313] An initial optimal gene determination submodule, configured to determine a group fitness value through the initialization matrix and determine an initial optimal gene from the genes;

[0314] a maximum deviation distance parameter determination submodule, configured to determine a maximum deviation distance parameter based on the initialization matrix, and generate a similarity time point number parameter and a maximum deviation time point number parameter according to the maximum deviation distance parameter in accordance with a preset standard;

[0315] A target optimal gene determination submodule is used to determine the target optimal gene based on the initial optimal gene, the neural network initialization parameter, the similar time point number parameter, the maximum deviation time point number parameter and the predicted data clustering attribution data;

[0316] A second parameter determination submodule is used to determine initial weight parameters and threshold position parameters for a neural network based on the target optimal gene;

[0317] The bandwidth prediction model generation submodule is used to train a neural network according to the initial weight parameters, the threshold position parameters and the prediction data clustering attribution data to generate a bandwidth prediction model.

[0318] Optionally, the neural network initialization parameter may include an input layer number parameter, and the input layer number parameter is determined according to the number of timestamps.

[0319] Optionally, the bandwidth prediction result generation module may include:

[0320] The submodule for generating data to be predicted is used to normalize the source data to be predicted and generate the data to be predicted;

[0321] A prediction value data generation submodule, configured to train the data to be predicted based on the bandwidth prediction model to obtain prediction value data;

[0322] The bandwidth prediction result generating submodule is used to perform denormalization processing on the predicted value data to generate a bandwidth prediction result.

[0323] As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0324] In addition, an embodiment of the present invention further provides an electronic device, comprising: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, the various processes of the above-mentioned bandwidth data prediction result generating method embodiment are implemented, and the same technical effects can be achieved. To avoid repetition, they will not be described here.

[0325] An embodiment of the present invention further provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the various processes of the aforementioned bandwidth data prediction result generation method embodiment and achieves the same technical effects. To avoid repetition, the description is omitted here. The computer-readable storage medium may be, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0326] Figure 4 A schematic diagram of the hardware structure of an electronic device for implementing various embodiments of the present invention.

[0327] The electronic device 400 includes but is not limited to: a radio frequency unit 401, a network module 402, an audio output unit 403, an input unit 404, a sensor 405, a display unit 406, a user input unit 407, an interface unit 408, a memory 409, a processor 410, and a power supply 411. It will be understood by those skilled in the art that Figure 4 The electronic device structure shown in the figure does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown, or may combine certain components or arrange the components differently. In the embodiments of the present invention, the electronic device includes but is not limited to a mobile phone, a tablet computer, a laptop computer, a PDA, an in-vehicle terminal, a wearable device, and a pedometer.

[0328] It should be understood that in this embodiment of the present invention, the RF unit 401 can be used to receive and transmit signals during information transmission or calls. Specifically, it receives downlink data from the base station and transmits it to the processor 410 for processing; in addition, it transmits uplink data to the base station. Typically, the RF unit 401 includes, but is not limited to, an antenna, at least one amplifier, a transceiver, a coupler, a low-noise amplifier, a duplexer, and the like. Furthermore, the RF unit 401 can communicate with the network and other devices via a wireless communication system.

[0329] The electronic device provides users with wireless broadband Internet access through the network module 402, such as helping users to send and receive emails, browse web pages, and access streaming media.

[0330] The audio output unit 403 can convert audio data received by the RF unit 401 or the network module 402 or stored in the memory 409 into an audio signal and output it as sound. In addition, the audio output unit 403 can also provide audio output related to a specific function performed by the electronic device 400 (for example, a call signal reception sound, a message reception sound, etc.). The audio output unit 403 includes a speaker, a buzzer, a receiver, etc.

[0331] The input unit 404 is used to receive audio or video signals. The input unit 404 may include a graphics processing unit (GPU) 4041 and a microphone 4042. The graphics processor 4041 processes image data of a still picture or video obtained by an image capture device (such as a camera) in a video capture mode or an image capture mode. The processed image frames can be displayed on the display unit 406. The image frames processed by the graphics processor 4041 can be stored in the memory 409 (or other storage medium) or transmitted via the radio frequency unit 401 or the network module 402. The microphone 4042 can receive sound and process such sound into audio data. The processed audio data can be converted into a format that can be sent to a mobile communication base station via the radio frequency unit 401 in the case of a telephone call mode.

[0332] The electronic device 400 also includes at least one sensor 405, such as a light sensor, a motion sensor, and other sensors. Specifically, the light sensor includes an ambient light sensor and a proximity sensor, wherein the ambient light sensor can adjust the brightness of the display panel 4061 according to the brightness of the ambient light, and the proximity sensor can turn off the display panel 4061 and / or the backlight when the electronic device 400 is moved to the ear. As a type of motion sensor, the accelerometer sensor can detect the magnitude of acceleration in all directions (generally three axes), and can detect the magnitude and direction of gravity when stationary. It can be used to identify the posture of the electronic device (such as horizontal and vertical screen switching, related games, magnetometer posture calibration), vibration recognition related functions (such as pedometer, tapping), etc.; the sensor 405 can also include a fingerprint sensor, a pressure sensor, an iris sensor, a molecular sensor, a gyroscope, a barometer, a hygrometer, a thermometer, an infrared sensor, etc., which will not be repeated here.

[0333] The display unit 406 is used to display information input by the user or information provided to the user. The display unit 406 may include a display panel 4061, which may be configured in the form of a liquid crystal display (LCD), an organic light-emitting diode (OLED), or the like.

[0334] The user input unit 407 can be used to receive input digital or character information, and to generate key signal input related to the user settings and function control of the electronic device. Specifically, the user input unit 407 includes a touch panel 4071 and other input devices 4072. The touch panel 4071, also known as a touch screen, can collect user touch operations on or near it (such as operations performed by the user using any suitable object or accessory such as a finger, stylus, etc. on or near the touch panel 4071). The touch panel 4071 may include two parts: a touch detection device and a touch controller. Among them, the touch detection device detects the user's touch direction and detects the signal caused by the touch operation, and transmits the signal to the touch controller; the touch controller receives the touch information from the touch detection device and converts it into touch point coordinates, which are then sent to the processor 410, which receives and executes the command sent by the processor 410. In addition, the touch panel 4071 can be implemented using various types such as resistive, capacitive, infrared and surface acoustic wave. In addition to the touch panel 4071, the user input unit 407 may also include other input devices 4072. Specifically, other input devices 4072 may include but are not limited to a physical keyboard, function keys (such as volume control keys, switch keys, etc.), a trackball, a mouse, and a joystick, which will not be described in detail here.

[0335] Furthermore, the touch panel 4071 may be overlaid on the display panel 4061. When the touch panel 4071 detects a touch operation on or near it, it transmits the information to the processor 410 to determine the type of touch event. Subsequently, the processor 410 provides corresponding visual output on the display panel 4061 according to the type of touch event. Figure 4 In the figure, the touch panel 4071 and the display panel 4061 are two independent components to realize the input and output functions of the electronic device. However, in some embodiments, the touch panel 4071 and the display panel 4061 can be integrated to realize the input and output functions of the electronic device, which is not limited here.

[0336] The interface unit 408 is an interface for connecting external devices to the electronic device 400. For example, the external devices may include a wired or wireless headset port, an external power supply (or battery charger) port, a wired or wireless data port, a memory card port, a port for connecting a device with an identification module, an audio input / output (I / O) port, a video I / O port, a headphone port, etc. The interface unit 408 may be used to receive input (e.g., data information, power, etc.) from the external device and transmit the received input to one or more elements within the electronic device 400, or may be used to transmit data between the electronic device 400 and the external device.

[0337] Memory 409 can be used to store software programs and various data. Memory 409 may primarily include a program storage area and a data storage area. The program storage area may store an operating system and at least one application required for a function (such as a sound playback function or an image playback function); the data storage area may store data generated based on the use of the mobile phone (such as audio data, a phone book, etc.). Furthermore, memory 409 may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0338] Processor 410 is the control center of the electronic device. It connects all parts of the electronic device using various interfaces and circuits. By running or executing software programs and / or modules stored in memory 409 and accessing data stored in memory 409, it performs various functions of the electronic device and processes data, thereby monitoring the electronic device as a whole. Processor 410 may include one or more processing units; preferably, processor 410 may integrate an application processor and a modem processor, wherein the application processor primarily processes the operating system, user interface, and application programs, while the modem processor primarily handles wireless communications. It is understood that the modem processor may not be integrated into processor 410.

[0339] The electronic device 400 may also include a power supply 411 (such as a battery) to supply power to each component. Preferably, the power supply 411 may be logically connected to the processor 410 through a power management system, thereby enabling the power management system to manage charging, discharging, and power consumption.

[0340] In addition, the electronic device 400 includes some functional modules not shown, which will not be described here.

[0341] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or apparatus comprising the element.

[0342] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better embodiment. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for enabling a terminal (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0343] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are protected by the present invention.

[0344] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the embodiments of the present invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0345] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0346] In the embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0347] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of the solution of this embodiment according to actual needs.

[0348] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0349] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, ROM, RAM, a magnetic disk, or an optical disk.

[0350] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A method for generating bandwidth data prediction results, characterized in that: include: Acquire historical bandwidth usage data and generate a timestamp corresponding to the historical bandwidth usage data; generating a plurality of historical data time series for the historical bandwidth usage data based on the timestamp; Dividing the historical bandwidth usage data into a plurality of bandwidth fuzzy clusters according to a plurality of the historical data time series; Establishing a prediction data clustering attribution model based on a plurality of bandwidth fuzzy clusters; Determining neural network initialization parameters, and generating a bandwidth prediction model using the predicted data clustering attribution model and the neural network initialization parameters; the neural network initialization parameters include an input layer number parameter, the input layer number parameter is determined according to the number of timestamps; the predicted data clustering attribution model has corresponding predicted data clustering attribution data; generating a bandwidth prediction result based on the bandwidth prediction model; The step of establishing a prediction data clustering attribution model based on a plurality of bandwidth fuzzy clusters comprises: Determining a predicted bandwidth cluster center sequence and a plurality of other bandwidth cluster center sequences associated with the predicted bandwidth cluster center sequence from the bandwidth fuzzy clustering; Calculating the Euclidean distance between the predicted bandwidth cluster center sequence and the plurality of other bandwidth cluster center sequences; the Euclidean distance has a corresponding Euclidean distance parameter; the Euclidean distance parameter corresponds one-to-one to the other bandwidth cluster center sequences; Selecting a minimum Euclidean distance parameter from the Euclidean distance parameters, and determining the other bandwidth cluster center sequence corresponding to the minimum Euclidean distance parameter as the target bandwidth cluster center sequence; Determining target bandwidth fuzzy clustering based on the target bandwidth cluster center sequence; The predicted data clustering attribution model is established based on the target bandwidth fuzzy clustering.

2. The method according to claim 1, characterized in that The step of dividing the historical bandwidth usage data into a plurality of bandwidth fuzzy clusters by using a plurality of the historical data time series comprises: Determining a category number parameter, a first fuzzy membership parameter, and a first fuzzy weight index parameter; Determining corresponding initial cluster centers from the plurality of historical data time series based on the first fuzzy membership parameter and the first fuzzy weight index parameter according to a first preset algorithm; Determining a second fuzzy weight index based on the category number parameter, the initial cluster center, and the first fuzzy weight index according to a second preset algorithm; Determining a second fuzzy membership parameter based on the second fuzzy weight index, the initial cluster center, and the category number parameter according to a third preset algorithm; Determining a target cluster center based on the initial cluster center, the second fuzzy weight index, the second fuzzy membership parameter, and the category number parameter; the target cluster center has a corresponding historical data time series; The historical data time series is divided into a plurality of bandwidth fuzzy clusters according to the target cluster center and the category number parameter.

3. The method according to claim 1, characterized in that The step of generating a bandwidth prediction model by using the prediction data clustering attribution model and the neural network initialization parameters includes: Generating an initialization matrix for the clustering belonging data of the predicted data through a mapping formula; the initialization matrix includes a plurality of genes for expressing the bandwidth fuzzy clustering; Determining a group fitness value through the initialization matrix, and determining an initial optimal gene from the genes; Determining a maximum deviation distance parameter based on the initialization matrix, and generating a similarity time point number parameter and a maximum deviation time point number parameter according to the maximum deviation distance parameter in accordance with a preset standard; Determine the target optimal gene based on the initial optimal gene, the neural network initialization parameter, the similarity time point number parameter, the maximum deviation time point number parameter and the predicted data clustering attribution data; Determining initial weight parameters and threshold position parameters for a neural network based on the target optimal gene; A neural network is trained according to the initial weight parameters, the threshold position parameters and the predicted data clustering attribution data to generate a bandwidth prediction model.

4. The method according to claim 1, wherein The step of generating a bandwidth prediction result based on the bandwidth prediction model includes: Normalize the source data to be predicted to generate the data to be predicted; Training the data to be predicted based on the bandwidth prediction model to obtain predicted value data; Denormalization is performed on the predicted value data to generate a bandwidth prediction result.

5. A bandwidth data prediction result generating device, characterized in that: include: A historical bandwidth usage data acquisition module is used to acquire historical bandwidth usage data and generate a timestamp corresponding to the historical bandwidth usage data; A historical data time series generating module, configured to generate a plurality of historical data time series for the historical bandwidth usage data based on the timestamp; A bandwidth fuzzy clustering division module is used to divide the historical bandwidth usage data into a plurality of bandwidth fuzzy clusters according to a plurality of historical data time series; A clustering attribution model establishment module, configured to establish a clustering attribution model for prediction data based on a plurality of bandwidth fuzzy clusters; a bandwidth prediction model generation module, configured to determine neural network initialization parameters and generate a bandwidth prediction model using the predicted data cluster attribution model and the neural network initialization parameters; the neural network initialization parameters including an input layer number parameter, the input layer number parameter being determined based on the number of timestamps; the predicted data cluster attribution model having corresponding predicted data cluster attribution data; A bandwidth prediction result generating module, configured to generate a bandwidth prediction result based on the bandwidth prediction model; The clustering attribution model building module includes: a bandwidth cluster center sequence determination submodule, configured to determine a predicted bandwidth cluster center sequence and a plurality of other bandwidth cluster center sequences associated with the predicted bandwidth cluster center sequence from the bandwidth fuzzy clustering; A Euclidean distance calculation submodule, configured to calculate the Euclidean distance between the predicted bandwidth cluster center sequence and a plurality of the other bandwidth cluster center sequences; the Euclidean distance has a corresponding Euclidean distance parameter; the Euclidean distance parameter corresponds one-to-one to the other bandwidth cluster center sequences; a target bandwidth cluster center sequence determination submodule, configured to select a minimum Euclidean distance parameter from the Euclidean distance parameters, and determine the other bandwidth cluster center sequence corresponding to the minimum Euclidean distance parameter as the target bandwidth cluster center sequence; a target bandwidth fuzzy clustering determination submodule, configured to determine target bandwidth fuzzy clustering based on the target bandwidth cluster center sequence; The prediction data clustering attribution model establishment submodule is used to establish the prediction data clustering attribution model based on the target bandwidth fuzzy clustering.

6. The device according to claim 5, characterized in that The bandwidth fuzzy clustering partitioning module includes: A first parameter determination submodule, configured to determine a category number parameter, a first fuzzy membership parameter, and a first fuzzy weight index parameter; an initial cluster center determination submodule, configured to determine corresponding initial cluster centers from a plurality of the historical data time series based on the first fuzzy membership parameter and the first fuzzy weight index parameter according to a first preset algorithm; A second fuzzy weight index determination submodule, configured to determine a second fuzzy weight index based on the category number parameter, the initial cluster center, and the first fuzzy weight index according to a second preset algorithm; a second fuzzy membership parameter determination submodule, configured to determine a second fuzzy membership parameter based on the second fuzzy weight index, the initial cluster center, and the category number parameter according to a third preset algorithm; a target cluster center determination submodule, configured to determine a target cluster center based on the initial cluster center, the second fuzzy weight index, the second fuzzy membership parameter, and the category number parameter; the target cluster center having a corresponding historical data time series; The bandwidth fuzzy clustering division submodule is used to divide the historical data time series into multiple bandwidth fuzzy clusters according to the target cluster center and the category number parameter.

7. The device according to claim 5, characterized in that The bandwidth prediction model generation module includes: An initialization matrix generation submodule is used to generate an initialization matrix for the clustering attribution data of the prediction data through a mapping formula; the initialization matrix contains a plurality of genes for expressing the bandwidth fuzzy clustering; An initial optimal gene determination submodule, configured to determine a group fitness value through the initialization matrix and determine an initial optimal gene from the genes; a maximum deviation distance parameter determination submodule, configured to determine a maximum deviation distance parameter based on the initialization matrix, and generate a similarity time point number parameter and a maximum deviation time point number parameter according to the maximum deviation distance parameter in accordance with a preset standard; A target optimal gene determination submodule is used to determine the target optimal gene based on the initial optimal gene, the neural network initialization parameter, the similar time point number parameter, the maximum deviation time point number parameter and the predicted data clustering attribution data; A second parameter determination submodule is used to determine initial weight parameters and threshold position parameters for a neural network based on the target optimal gene; The bandwidth prediction model generation submodule is used to train a neural network according to the initial weight parameters, the threshold position parameters and the prediction data clustering attribution data to generate a bandwidth prediction model.

8. The device according to claim 5, characterized in that The bandwidth prediction result generation module includes: The submodule for generating data to be predicted is used to normalize the source data to be predicted and generate the data to be predicted; A prediction value data generation submodule, configured to train the data to be predicted based on the bandwidth prediction model to obtain prediction value data; The bandwidth prediction result generating submodule is used to perform denormalization processing on the predicted value data to generate a bandwidth prediction result.

9. An electronic device, characterized in that: comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus; The memory is used to store computer programs; The processor is configured to implement the method according to any one of claims 1 to 4 when executing a program stored in the memory.

10. A computer-readable storage medium having instructions stored thereon, which, when executed by one or more processors, cause the processors to perform the method according to any one of claims 1 to 4.

Images