Short message sending method and device based on depth prediction model, equipment and medium

Abstract

The invention belongs to the field of artificial intelligence, and relates to a short message sending method based on a depth prediction model, and the method comprises the steps: monitoring an early warning system in real time, collecting the early warning information when the early warning system generates the early warning information, and obtaining a to-be-sent early warning information queue; obtaining an available short message modem set from a short message modem cluster, and performing performance prediction on each available short message modem in the available short message modem set through a pre-trained evaluation model to obtain the sending performance of each available short message modem; selecting a preset number of target short message modems from the available short message modem set according to the sending performance; and sequentially sending the to-be-sent early warning information queue to corresponding operation and maintenance personnel through a preset number of target short message modems. In addition, the invention also relates to a block chain technology, and early warning information and the like can be stored in a block chain. The sending efficiency and the sending success rate of the early warning information can be improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and in particular to a method for sending text messages based on a deep prediction model. Background Art

[0002] In the field of operation and maintenance warning, operation and maintenance warning through SMS has long been a common means of IT information system operation and maintenance. According to the warning information, the status changes of system operation can be obtained in time, which is a powerful tool for operation and maintenance personnel to respond to faults in time and ensure the stable operation of the system.

[0003] However, in the traditional SMS warning method, the performance differences of SMS modems are not taken into consideration when sending SMS warning messages through SMS modems, which may result in the use of SMS modems with poor performance to send warning information, thereby causing problems such as low SMS sending efficiency and low sending success rate. Summary of the invention

[0004] The purpose of the embodiments of the present application is to propose a method, device, equipment and medium for sending text messages based on a deep prediction model, the main purpose of which is to improve the efficiency and success rate of sending early warning information.

[0005] In order to solve the above technical problems, the embodiment of the present application provides a method for sending short messages based on a deep prediction model, which adopts the following technical solution:

[0006] Real-time monitoring of the early warning system, when the early warning system generates early warning information, collecting the early warning information to obtain a queue of early warning information to be sent;

[0007] Obtaining an available SMS modem set from an SMS modem cluster, and performing performance prediction on each available SMS modem in the available SMS modem set using a pre-trained evaluation model to obtain a sending performance of each available SMS modem;

[0008] According to the sending performance, selecting a preset number of target SMS modems from the set of available SMS modems;

[0009] The queue of warning messages to be sent is sent to corresponding operation and maintenance personnel in sequence through a preset number of target SMS modems.

[0010] Furthermore, after collecting the warning information, the method further includes:

[0011] Convert the warning information into a vector representation by using a preset vector conversion technology to obtain a warning vector;

[0012] Extracting features from the warning vector using a pre-trained priority extraction model to obtain warning features;

[0013] The warning features are classified by a preset classifier to obtain the priority corresponding to the warning information.

[0014] Further, the collecting of the warning information to obtain a queue of warning information to be sent includes:

[0015] According to the priority of the warning information, the warning information is sorted in descending order according to the priority to obtain an initial warning information queue;

[0016] The warning information of the same priority in the initial warning information queue is sorted according to the timestamp of the warning information to obtain the warning information queue.

[0017] Furthermore, before predicting the available SMS modems by using the pre-trained evaluation model, the method further includes:

[0018] Collecting historical data information sets of multiple SMS modems in each period, preprocessing the historical data information sets to obtain data sample sets;

[0019] Based on the historical data information set of each period, the data sample set is divided into a training set and a test set according to a preset ratio;

[0020] Constructing an initial evaluation model, and training the initial evaluation model using the training set to obtain an intermediate evaluation model;

[0021] Testing the intermediate evaluation model using the test set to obtain an evaluation loss value;

[0022] When the evaluation loss value is less than or equal to a preset evaluation loss threshold, obtaining the pre-trained evaluation model;

[0023] When the evaluation loss value is greater than the preset evaluation loss threshold, the initial evaluation model is continuously trained using the training set until the evaluation loss value is less than or equal to the preset evaluation loss threshold, and the training is completed to obtain the pre-trained evaluation model.

[0024] Furthermore, the pre-trained evaluation model is used to perform performance prediction on each available SMS modem in the available SMS modem set to obtain the sending performance of each available SMS modem, including:

[0025] Acquire current data information of each available SMS modem, wherein the current data information includes current cycle information and current sent quantity;

[0026] Using a pre-trained evaluation model, extract the features of the current cycle information and the current number of sent messages and combine them to obtain current data features;

[0027] A predicted value of each of the available SMS modems is predicted according to the current data feature, and the predicted value is used as the sending performance of each of the available SMS modems, wherein the sending performance includes a sending rate and a sending success rate.

[0028] Further, selecting a preset number of target SMS modems from the set of available SMS modems according to the sending performance includes:

[0029] Remove available SMS modems whose sending rates in the sending performance are less than a preset rate threshold from the available SMS modem set to obtain a first available SMS modem set;

[0030] Remove the available SMS modems whose sending performance has a sending success rate lower than a preset sending success rate threshold from the first available SMS modem set, to obtain a second available SMS modem set;

[0031] Performing a weighted operation on the sending performance of each available SMS modem in the second available SMS modem set to obtain a comprehensive performance of each available SMS modem;

[0032] Sorting is performed according to the comprehensive performance, and according to the number of the queues of warning messages to be sent, a preset number of available SMS modems are selected from the sorted second available SMS modem set as target SMS modems.

[0033] Furthermore, after sending the queue of warning information to be sent to corresponding operation and maintenance personnel in sequence, the method further includes:

[0034] Monitor the sending status of each warning message in the warning message queue;

[0035] Obtain the warning information that the sending status is failed, and obtain the warning information that was not sent;

[0036] The priority of the unsent warning information is modified to the highest priority, and the unsent warning information is put back into the queue of warning information to be sent.

[0037] In order to solve the above technical problems, the embodiment of the present application also provides a text message sending device based on a deep prediction model, which adopts the following technical solution:

[0038] A monitoring module is used to monitor the early warning system in real time. When the early warning system generates early warning information, the early warning information is collected to obtain a queue of early warning information to be sent;

[0039] An evaluation module is used to obtain an available SMS modem set from an SMS modem cluster, and to perform performance prediction on each available SMS modem in the available SMS modem set using a pre-trained evaluation model to obtain the sending performance of each available SMS modem;

[0040] A selection module, configured to select a preset number of target SMS modems from the set of available SMS modems according to the sending performance;

[0041] The sending module is used to send the queue of warning messages to be sent to the corresponding operation and maintenance personnel in sequence through a preset number of target SMS modems.

[0042] In order to solve the above technical problems, an embodiment of the present application also provides a computer device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the SMS sending method based on the deep prediction model as described above.

[0043] In order to solve the above technical problems, an embodiment of the present application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the SMS sending method based on the deep prediction model as described above.

[0044] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0045] By real-time monitoring of the early warning system, when the early warning system generates early warning information, the early warning information is collected to obtain a queue of early warning information to be sent, so that the early warning information can be captured in time and the duration of the fault can be reduced; by aggregating the early warning information into an early warning information queue, processing resources can be allocated more effectively, the efficiency of sending subsequent early warning information can be improved, and the fault situation can be ensured to be handled in a timely manner, thereby ensuring the stability of the system.

[0046] The pre-trained evaluation model can be used to accurately predict the sending performance of each available SMS modem, so that the warning information can be sent according to the sending performance of each available SMS modem in the subsequent process, thereby improving the sending success rate and efficiency of the warning information.

[0047] By selecting the target SMS modem based on the sending performance (such as sending rate and sending success rate), it can be ensured that the selected SMS modem has high working efficiency and reliability, thereby improving the sending efficiency and sending success rate of the warning information; by selecting the available SMS modem with better performance, the system resources can be used more effectively, avoiding the allocation of resources to SMS modems with poor performance, which can reduce resource waste and improve the overall system operation efficiency.

[0048] By using a preset number of target SMS cats, the queue of warning information to be sent can be quickly sent to the operation and maintenance personnel, which improves the efficiency and success rate of sending warning information; sending warning information through target SMS cats can ensure the accuracy and completeness of the warning information and avoid decision-making errors caused by information distortion or omissions. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0050] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0051] Figure 2 A flowchart of an embodiment of a method for sending text messages based on a deep prediction model according to the present application;

[0052] Figure 3 It is a structural schematic diagram of an embodiment of a text message sending device based on a deep prediction model according to the present application;

[0053] Figure 4 It is a structural schematic diagram of an embodiment of a device according to the present application. DETAILED DESCRIPTION

[0054] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0055] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0056] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0057] like Figure 1 As shown, the system architecture 100 may include a terminal device 101, a network 102 and a server 103. The terminal device 101 may be a laptop 1011, a tablet computer 1012 or a mobile phone 1013. The network 102 is used to provide a medium for a communication link between the terminal device 101 and the server 103. The network 102 may include various connection types, such as wired, wireless communication links or optical fiber cables.

[0058] The user can use the terminal device 101 to interact with the server 103 through the network 102 to receive or send messages, etc. Various communication client applications can be installed on the terminal device 101, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0059] The terminal device 101 can be any electronic device with a display screen and supporting web browsing. In addition to a laptop computer 1011, a tablet computer 1012 or a mobile phone 1013, the terminal device 101 can also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 (Moving Picture Experts Group Audio Layer IV), a laptop computer, a desktop computer, etc.

[0060] The server 103 may be a server that provides various services, such as a background server that provides support for a web page displayed on the terminal device 101 .

[0061] It should be noted that the SMS sending method based on the deep prediction model provided in the embodiment of the present application is generally executed by a server / terminal device, and accordingly, the SMS sending device based on the deep prediction model is generally arranged in the server / terminal device.

[0062] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to the implementation requirements.

[0063] Continue to refer Figure 2, showing a flowchart of an embodiment of a method for sending text messages based on a deep prediction model according to the present application. According to different requirements, the order of the steps in the flowchart can be changed, and some steps can be omitted. The method for sending text messages based on a deep prediction model provided in the embodiment of the present application can be applied to any scenario where text messages based on a deep prediction model are required to be sent, and the method for sending text messages based on a deep prediction model can be applied to products in these scenarios. The method for sending text messages based on a deep prediction model comprises the following steps:

[0064] Step S201: monitor the early warning system in real time. When the early warning system generates early warning information, collect the early warning information to obtain a queue of early warning information to be sent.

[0065] In this embodiment, the method for sending a short message based on a deep prediction model is executed on an electronic device (eg Figure 1 The server / terminal device shown in the figure) can obtain the collection warning information through a wired connection or a wireless connection. It should be pointed out that the above-mentioned wireless connection method may include but is not limited to 3G / 4G / 5G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other wireless connection methods currently known or to be developed in the future. The executor of the present application may be an evaluation model, or simply referred to as a model. The current data information of the above-mentioned SMS cat usually has the same format as the historical data used in the training of the evaluation model, but the training data contains other noise, which depends on the specific situation.

[0066] In this embodiment, the above-mentioned early warning system refers to a system for monitoring abnormal conditions of various components in the system and generating early warning information based on the abnormal conditions. The components include but are not limited to the CPU, memory, hard disk, etc. in the operating system, and the abnormal conditions include but are not limited to the occupancy of the components, the temperature of the components, memory leakage, etc. For example, the occupancy of the memory, CPU, hard disk, etc. in the operating system is monitored. When the occupancy rate of the memory, CPU, hard disk, etc. in the operating system exceeds a preset threshold, an early warning information is generated. For example, if the CPU occupancy rate exceeds the preset threshold of 95%, an early warning information of the CPU occupancy warning will be generated.

[0067] In this embodiment, the early warning system monitors the abnormal conditions of various components in the system in real time. When an abnormal condition is detected in any component, the early warning system will generate corresponding early warning information based on the abnormal condition. Since the early warning system is monitored in real time, when the early warning system generates early warning information, the early warning information is collected through a preset collection method. The above-mentioned preset collection method includes but is not limited to using scripts to collect early warning information, collecting early warning information through TCP / IP interface calls, and directly collecting early warning information through a direct connection to the SMS queue database, and collecting all early warning information to generate a queue of early warning information to be sent.

[0068] In this embodiment, the above-mentioned queue of warning information to be sent is saved in a database, and the queue of warning information is stored in a database table to facilitate the subsequent sending of the queue of warning information to be sent.

[0069] In one embodiment, after collecting the warning information, the method further includes:

[0070] Convert the warning information into a vector representation by using a preset vector conversion technology to obtain a warning vector;

[0071] Extracting features from the warning vector using a pre-trained priority extraction model to obtain warning features;

[0072] The warning features are classified by a preset classifier to obtain the priority corresponding to the warning information.

[0073] In this embodiment, after the warning information is collected, the warning information is processed by a priority classification model to obtain the priority corresponding to each warning information, wherein the priority classification model refers to a convolutional neural network, including an output layer, a convolution layer, a pooling layer, a fully connected layer and an output layer, wherein the fully connected layer refers to a preset classifier.

[0074] Specifically, after the warning information is collected, each warning information is converted into a vector representation through a preset vector conversion technology (such as Word2Vec, BERT, etc.) to obtain a warning vector, and each warning vector is input into a pre-trained priority classification model. The local features of the warning vector are extracted through multiple convolution kernels of the convolution layer of the pre-trained priority classification model, and then the local features output by the convolution layer are downsampled through the pooling layer to reduce the dimension and calculation amount of the data while retaining important features to obtain the warning features. The warning features are mapped to the space of priority categories through a fully connected layer, and the priority probability distribution of the warning information is calculated through the softmax function. The priority of the warning information is selected as the priority of the warning information, and the priority of the warning information is output through the output layer.

[0075] In this embodiment, the priority of each warning information is output through the priority classification model, so that subsequent high-priority warning information can be sent first, avoiding low-priority warning information occupying the sending channel, resulting in high-priority warning information not being pushed to the equipment of the operation and maintenance personnel in time, causing system crashes, thereby improving system stability.

[0076] In another embodiment, before extracting features from the warning vector using a pre-trained priority extraction model to obtain warning features, the method further includes:

[0077] Obtaining a labeled early warning information data set, and dividing the labeled early warning information data set into a training set and a test set according to a preset ratio;

[0078] Constructing an initial priority classification model, and training the initial priority classification model using the training set to obtain an intermediate priority classification model;

[0079] Testing the intermediate priority classification model through a test set to obtain a classification loss value;

[0080] When the classification loss value is less than or equal to a preset classification loss threshold, obtaining the pre-trained priority classification model;

[0081] When the classification loss value is greater than the preset classification loss threshold, the initial priority classification model is continued to be trained using the training set until the classification loss value is less than or equal to the preset classification loss threshold, and the training is completed to obtain the pre-trained priority classification model.

[0082] In this embodiment, an initial priority classification model is constructed to obtain a large number of annotated warning information data sets, for example, the CPU usage in the warning information data set is marked as high priority, etc. A large number of annotated warning information data sets are split into training sets and test sets according to a preset ratio; the initial priority classification model is trained with the training set to obtain an intermediate priority classification model; the intermediate priority classification model of the test set is used for testing to obtain a classification loss value, and it is determined whether the classification loss value is less than or equal to a preset classification loss threshold. When the classification loss value is less than or equal to the preset classification loss threshold, the training is completed to obtain a pre-trained priority classification model. When the classification loss value is greater than the preset classification loss threshold, the initial priority classification model is continuously trained with the training set until the classification loss value is less than or equal to the preset classification loss threshold, and the training is terminated to obtain a pre-trained priority classification model.

[0083] In one embodiment, the collecting of the warning information to obtain a queue of warning information to be sent includes:

[0084] According to the priority of the warning information, the warning information is sorted in descending order according to the priority to obtain an initial warning information queue;

[0085] The warning information of the same priority in the initial warning information queue is sorted according to the timestamp of the warning information to obtain the warning information queue.

[0086] In this embodiment, each warning message has a corresponding priority (such as high, medium, and low). The above priorities are only examples. In actual situations, the priorities may also include ultra-high, higher, lower, etc.; according to the priority of each warning message, each warning message is sorted from high to low in order of priority to obtain an initial warning message queue; since each warning message will carry a timestamp when it is generated, for example, the timestamp generated by warning message A is (06 minutes and 06 seconds on December 12, 2024), the above example is only an example. In different situations, the timestamp can be accurate to milliseconds, etc.; after obtaining the initial warning message queue, the warning messages of the same priority in the initial warning message queue are sorted again according to the timestamp of each warning message, and finally the warning message queue is obtained.

[0087] In this embodiment, the warning information is sorted by priority and timestamp sorting, which can ensure that high-priority warning information can be processed in a timely manner, thereby ensuring the stability of the system; and through priority sorting, processing resources can be effectively allocated, high-priority warning information is processed first, and low-priority warning information is processed gradually, thereby improving resource utilization efficiency and avoiding resource waste.

[0088] In this embodiment, by real-time monitoring of the early warning system, when the early warning system generates early warning information, the early warning information is collected to obtain a queue of early warning information to be sent, so that the early warning information can be captured in time and the duration of the fault can be reduced; by aggregating the early warning information into an early warning information queue, processing resources can be allocated more effectively, the efficiency of sending subsequent early warning information can be improved, and the fault situation can be ensured to be handled in a timely manner, thereby ensuring the stability of the system.

[0089] Step S202: obtaining an available SMS modem set from the SMS modem cluster, and performing performance prediction on each available SMS modem in the available SMS modem set using a pre-trained evaluation model to obtain the sending performance of each available SMS modem.

[0090] In this embodiment, the above-mentioned SMS cat cluster refers to a collection composed of multiple SMS cats, which are connected to the SMS sending server through a serial port or a communication interface to realize the sending and receiving functions of SMS, wherein each SMS cat has the ability to send SMS, and the SMS cat cluster improves the SMS sending capacity and reliability of the entire system by combining the resources of multiple SMS cats; the above-mentioned SMS sending server is a classic network communication model, adopting the C / S mode (Client / Server, client / server architecture), which is used to receive warning information from the warning information queue and send warning information from the warning information queue. The client and the server communicate through the network to realize data exchange and resource sharing. The above-mentioned sending of the warning information in the warning information queue through the SMS cat can be realized through concurrent lightweight, with a small amount of code and little resource occupation.

[0091] In this embodiment, the above-mentioned pre-trained evaluation model refers to a multilayer perceptron model (MLP) in deep learning, including an input layer, several hidden layers and an output layer, wherein the input layer and the output layer are both linear layers, and the number of hidden layers can be adjusted according to the specific situation of the model training.

[0092] In this embodiment, the sending performance refers to the number of messages sent, the sending rate, the sending success rate, etc. of the available SMS cats within a cycle. The state of each SMS cat in the SMS cat cluster is checked, and the state includes idle, offline, busy (sending SMS messages), faulty, etc. SMS cats with idle states are selected from the SMS cat cluster, and these idle SMS cats are used as available SMS cats to obtain an available SMS cat set; based on each available SMS cat in the available SMS cat set, the sending performance of each available SMS cat is predicted by a pre-trained evaluation model.

[0093] Judging from the historical sending data of the SMS cat, the sending performance of the SMS cat is not fixed. If only a fixed value or a similar value is taken to evaluate the sending performance of the SMS cat, the sending performance of the SMS cat will be less accurate. After analysis, the sending performance of the SMS cat is different in each cycle. Based on this, the sending performance of the SMS cat in each cycle is analyzed separately.

[0094] In one embodiment, before predicting the available SMS modem by using the pre-trained evaluation model, the method further includes:

[0095] Collecting historical data information sets of multiple SMS modems in each period, preprocessing the historical data information sets to obtain data sample sets;

[0096] Based on the historical data information set of each period, the data sample set is divided into a training set and a test set according to a preset ratio;

[0097] Constructing an initial evaluation model, and training the initial evaluation model using the training set to obtain an intermediate evaluation model;

[0098] Testing the intermediate evaluation model using the test set to obtain an evaluation loss value;

[0099] When the evaluation loss value is less than or equal to a preset evaluation loss threshold, obtaining the pre-trained evaluation model;

[0100] When the evaluation loss value is greater than the preset evaluation loss threshold, the initial evaluation model is continuously trained using the training set until the evaluation loss value is less than or equal to the preset evaluation loss threshold, and the training is completed to obtain the pre-trained evaluation model.

[0101] In this embodiment, historical data information sets of multiple SMS cats in each period are collected from a historical database, and the historical data information sets include but are not limited to data information such as the number of SMS cats sent, the sending rate, and the sending success rate in each period. The number of sent is used as input, and the sending rate and the sending success rate are used as annotation information; the above-mentioned each period refers to a predefined time period as a period, such as weekly, monthly, quarterly, etc.; a preprocessing operation is performed on the historical data information set, and the preprocessing operation includes but is not limited to a normalization operation, etc., to obtain a data sample set, wherein the normalization operation is to compress the historical data information set between [0.1] through normalization processing, each sample is a multidimensional vector with a dimension of m, and if there are n samples, the data sample set is an n*m matrix; according to a preset ratio (such as 7: 3) Split the data sample set into a training set and a test set, and construct an initial evaluation model. The initial evaluation model is consistent with the model architecture of the evaluation model mentioned above. The initial evaluation model is trained with the training set. After the training is completed, an intermediate evaluation model is obtained. The intermediate evaluation model is verified with the test set. A preset loss function (such as MSEloss) is used to measure the difference between the predicted value and the actual value to obtain an evaluation loss value. The evaluation loss value is compared with a preset evaluation loss threshold. When the evaluation loss value is less than or equal to the preset evaluation loss threshold, a trained evaluation model is obtained. When the evaluation loss value is greater than the preset evaluation loss threshold, the initial evaluation model is continued to be trained with the training set until the evaluation loss value is less than or equal to the preset evaluation loss threshold to obtain a trained evaluation model.

[0102] Specifically, back propagation is performed based on the loss value, the gradient is calculated, and the model parameters are updated using the optimizer. K-fold cross validation is used to evaluate the performance of the model after the model parameters are updated, and the best hyperparameter combination (such as batch size, learning rate, number of network layers, number of neurons, etc.) is determined. According to the results of cross validation, the hyperparameters and model architecture are continuously adjusted to reduce the loss rate on the test set and improve the accuracy. Among them, weight decay is used. The methods of weight decay and dropout are used to assist the training process of the initial evaluation model. Specifically, a weight decay term is added to the optimizer, and the weight decay term is added to the loss function as a regularization term, so that the model parameters shrink to smaller values ​​during the training process to limit the complexity of the initial evaluation model parameters and prevent the overfitting of the initial evaluation model. A dropout layer is added to each hidden layer of the initial evaluation model to randomly drop a part of the neurons and improve the generalization ability of the model. Specifically, during the training process, the dropout layer randomly drops the output of the hidden layer neurons with a certain probability (such as 0.5); during the test or evaluation process, the dropout layer will not drop the output of any neuron, but will multiply the output of all neurons by a scaling factor related to the drop probability to maintain the consistency of the output.

[0103] In this embodiment, by acquiring historical data information sets of SMS modems in each period and training the initial evaluation model, a trained evaluation model is obtained, which can accurately evaluate the sending performance of each SMS modem in each period, thereby improving the sending efficiency of warning information.

[0104] In one embodiment, the step of performing performance prediction on each available SMS modem in the set of available SMS modems by using a pre-trained evaluation model to obtain the sending performance of each available SMS modem includes:

[0105] Acquire current data information of each available SMS modem, wherein the current data information includes current cycle information and current sent quantity;

[0106] Using a pre-trained evaluation model, extract the features of the current cycle information and the current number of sent messages and combine them to obtain current data features;

[0107] A predicted value of each of the available SMS modems is predicted according to the current data feature, and the predicted value is used as the sending performance of each of the available SMS modems, wherein the sending performance includes a sending rate and a sending success rate.

[0108] Since the available SMS cats in each cycle and the sending performance corresponding to the current data situation of each available SMS cat are different, in this embodiment, the current data information of each available SMS cat in the available SMS cat set is obtained, and the current data information includes the current cycle information and the current sent quantity. The current data information refers to the cycle in which the available SMS cat is, such as any season of spring, summer, autumn and winter, and the time period of any day in the season (such as morning or evening), etc.; the above-mentioned sent quantity refers to the number of SMS messages that have been sent since the available SMS cat was reset. The above-mentioned pre-trained evaluation model is used to process based on the current data information of the available SMS cat, extract the features of the current cycle information and the current sent quantity, obtain the current cycle feature and the current sent quantity feature, and combine the current cycle feature and the current sent quantity feature to obtain the current data feature; the current data feature is calculated by the parameters and logic learned by the evaluation model during training to obtain the predicted value of the available SMS cat, and the predicted value is used as the sending performance, wherein the sending performance includes the sending rate and the sending success rate.

[0109] In this embodiment, the sending performance of each available SMS modem is accurately predicted through a pre-trained evaluation model, so that the warning information is subsequently sent according to the sending performance of each available SMS modem, thereby improving the sending success rate and sending efficiency of the warning information.

[0110] Step S203: selecting a preset number of target SMS modems from the set of available SMS modems according to the sending performance.

[0111] In this embodiment, according to the sending performance of each available SMS cat and the number of queues of warning messages to be sent obtained in step S201, the preset number of target SMS cats is determined according to the number of queues of warning messages to be sent, and a preset number of target SMS cats are selected from the set of available SMS cats according to the sending performance and the preset number; for example, if the number of queues of warning messages to be sent is 100, the number of target SMS cats is determined to be 5.

[0112] In one embodiment, selecting a preset number of target SMS modems from the set of available SMS modems according to the sending performance includes:

[0113] Remove available SMS modems whose sending rates in the sending performance are less than a preset rate threshold from the available SMS modem set to obtain a first available SMS modem set;

[0114] Remove the available SMS modems whose sending performance has a sending success rate lower than a preset sending success rate threshold from the first available SMS modem set, to obtain a second available SMS modem set;

[0115] Performing a weighted operation on the sending performance of each available SMS modem in the second available SMS modem set to obtain a comprehensive performance of each available SMS modem;

[0116] Sorting is performed according to the comprehensive performance, and according to the number of the queues of warning messages to be sent, a preset number of available SMS modems are selected from the sorted second available SMS modem set as target SMS modems.

[0117] In this embodiment, in order to prevent interference from extreme values ​​in the sending performance of available SMS cats, it is necessary to set a preset rate threshold and a preset sending success rate threshold, remove available SMS cats with a sending rate efficiency of the preset rate threshold from the set of available SMS cats to obtain a first set of available SMS cats, and then remove available SMS cats with a sending success rate less than the preset sending success rate threshold from the first set of available SMS cats to obtain a second set of available SMS cats, and based on the second set of available SMS cats after removing the extreme values, calculate the comprehensive performance of each available SMS cat, specifically using a weighted average or weighted sum operation to calculate, to obtain the comprehensive performance of each available SMS cat, and then determine a preset number based on the number of queues of warning messages to be sent obtained in step S201, and select a preset number of available SMS cats from the second set of available SMS cats as target SMS cats according to the preset number to obtain a preset number of target SMS cats.

[0118] In this embodiment, by setting a preset rate threshold and a preset sending success rate threshold, interference from extreme values ​​in the set of available SMS modems is eliminated, thereby improving the reliability of target SMS modem selection; the number of target SMS modems is determined according to the number of queues of warning information to be sent, ensuring that resources are reasonably allocated; and by comprehensively considering the comprehensive performance of the available SMS modems, the stability of the system when the SMS modems are sending is enhanced.

[0119] In this embodiment, if the set of available SMS modems does not contain a preset number of target SMS modems, the process waits until an available SMS modem is found, and the available SMS modem is used as the target SMS modem.

[0120] In this embodiment, by selecting the target SMS modem based on the sending performance (such as sending rate and sending success rate), it can be ensured that the selected SMS modem has high working efficiency and reliability, thereby improving the sending efficiency and sending success rate of the warning information; by selecting the available SMS modem with better performance, the system resources can be used more effectively, avoiding the allocation of resources to SMS modems with poor performance, which can reduce resource waste and improve the operating efficiency of the overall system.

[0121] Step S204: The queue of warning messages to be sent is sent in sequence to corresponding operation and maintenance personnel through a preset number of target SMS modems.

[0122] In this embodiment, in the SMS sending task, each target SMS modem needs a certain amount of time to send a SMS message. If a single-threaded method is adopted, that is, the SMS modems are allowed to send SMS messages one by one, then when a certain SMS modem is sending, other SMS modems will be idle. This processing method is not only inefficient, but also cannot fully utilize the resources of the SMS modems.

[0123] In order to improve the throughput and efficiency of SMS sending, we adopted a multi-threaded programming method; multi-threading allows multiple tasks to run simultaneously, that is, parallel processing. In this scenario, we can assign an independent thread to each target SMS modem, and each thread is responsible for controlling the corresponding SMS modem to send SMS messages.

[0124] By sending SMS commands in parallel through multithreading, we can enable multiple SMS modems to work simultaneously, thus significantly improving the throughput of SMS sending. This parallel processing method makes the SMS sending speed almost reach the maximum value allowed by the operator, maximizing the use of SMS modem resources.

[0125] In this embodiment, the target SMS cat uses a multi-threaded programming method to send the queue of warning messages to be sent to the corresponding operation and maintenance personnel in sequence. After the target SMS cat completes sending the warning messages in the queue of warning messages to be sent, the state of the target SMS cat is modified to an idle state and the thread is destroyed.

[0126] In the process of sending the warning information through the target SMS modem, there may be reasons for the failure of SMS sending. Therefore, in one embodiment, after sending the queue of warning information to be sent to the corresponding operation and maintenance personnel in sequence, the method further includes:

[0127] Monitor the sending status of each warning message in the warning message queue;

[0128] Obtain the warning information that the sending status is failed, and obtain the warning information that was not sent;

[0129] The priority of the unsent warning information is modified to the highest priority, and the unsent warning information is put back into the queue of warning information to be sent.

[0130] In this embodiment, the sending success rate of the target SMS cat is not 100%, and there may be situations where the warning information fails to be sent. By monitoring the sending status of each warning message in the warning message queue in real time or periodically, ignoring the warning message with a successful sending status, obtaining the warning message with a failed sending status, obtaining the unsent warning message, and modifying the priority of the unsent warning message to the highest priority, ensuring that the warning message can be sent in time when the sending fails. After modifying the priority of the unsent warning message, put the unsent warning message back into the queue of warning messages to be sent for sending, and re-enter the SMS sending process until the sending status of all warning messages in the warning SMS queue is successful, and the sending is terminated.

[0131] In this embodiment, by monitoring the sending status of the warning information and giving the highest priority to the warning information that failed to be sent, it is ensured that the unsent warning information can be sent in time, thereby improving the timeliness of the warning information sending and the reliability of the system; by adjusting the priority of the warning information that failed to be sent, the fault tolerance of the system is enhanced and the stability of the system is guaranteed.

[0132] In this embodiment, the queue of warning information to be sent can be quickly sent to the operation and maintenance personnel through a preset number of target SMS cats, thereby improving the efficiency and success rate of sending warning information; sending warning information through target SMS cats can ensure the accuracy and completeness of the warning information, and avoid decision-making errors caused by information distortion or omissions.

[0133] It should be emphasized that in order to further ensure the privacy and security of the above-mentioned SMS sending results based on the deep prediction model, the above-mentioned warning information can also be stored in a blockchain node.

[0134] The blockchain referred to in this application is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.

[0135] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0136] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0137] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0138] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0139] Further references Figure 3 , as a response to the above Figure 2 The present application provides an embodiment of a device for sending short messages based on a deep prediction model. Figure 2 Corresponding to the method embodiment shown, the apparatus can be specifically applied to various computer devices.

[0140] like Figure 3 As shown, the SMS sending device 300 based on the deep prediction model described in this embodiment includes: a monitoring module 301, an evaluation module 302, a selection module 303, and a sending module 304. Among them:

[0141] The monitoring module 301 is used to monitor the early warning system in real time. When the early warning system generates early warning information, the early warning information is collected to obtain a queue of early warning information to be sent;

[0142] In one embodiment, the apparatus further comprises:

[0143] A vector conversion module, used to convert the warning information into a vector representation through a preset vector conversion technology to obtain a warning vector;

[0144] A first feature extraction module, used to extract features from the warning vector using a pre-trained priority extraction model to obtain a warning feature;

[0145] The classification module is used to classify the warning features through a preset classifier to obtain the priority corresponding to the warning information.

[0146] In one embodiment, the monitoring module 301 includes:

[0147] A first sorting submodule is used to sort the warning information according to the priority of the warning information from high to low, so as to obtain an initial warning information queue;

[0148] The second sorting submodule is used to sort the warning information of the same priority in the initial warning information queue according to the timestamp of the warning information to obtain the warning information queue.

[0149] An evaluation module 302 is used to obtain an available SMS modem set from an SMS modem cluster, and to perform performance prediction on each available SMS modem in the available SMS modem set using a pre-trained evaluation model to obtain a sending performance of each available SMS modem;

[0150] In one embodiment, the apparatus further comprises:

[0151] A collection module, used for collecting historical data information sets of multiple SMS modems in each period, and preprocessing the historical data information sets to obtain data sample sets;

[0152] A splitting module is used to divide the data sample set into a training set and a test set according to a preset ratio based on the historical data information set of each period;

[0153] A construction module is used to construct an initial evaluation model, and train the initial evaluation model through the training set to obtain an intermediate evaluation model;

[0154] A testing module, used to test the intermediate evaluation model through the test set to obtain an evaluation loss value;

[0155] A first comparison module, used for obtaining the pre-trained evaluation model when the evaluation loss value is less than or equal to a preset evaluation loss threshold;

[0156] The second comparison module is used to continue training the initial evaluation model using the training set when the evaluation loss value is greater than the preset evaluation loss threshold, until the evaluation loss value is less than or equal to the preset evaluation loss threshold, and the training is completed to obtain the pre-trained evaluation model.

[0157] In one embodiment, the evaluation module 302 includes:

[0158] An acquisition submodule is used to acquire the current data information of each available SMS modem, wherein the current data information includes current cycle information and current sent quantity;

[0159] An extraction submodule, used to extract and combine the features of the current cycle information and the current number of sent data using a pre-trained evaluation model to obtain current data features;

[0160] The prediction submodule is used to predict the prediction value of each of the available SMS modems according to the current data characteristics, and use the prediction value as the sending performance of each of the available SMS modems, wherein the sending performance includes a sending rate and a sending success rate.

[0161] A selection module 303, configured to select a preset number of target SMS modems from the set of available SMS modems according to the sending performance;

[0162] In one embodiment, the selection module 303 includes:

[0163] A first removal submodule is used to remove the available SMS modems whose sending rates in the sending performance are less than a preset rate threshold from the available SMS modem set, so as to obtain a first available SMS modem set;

[0164] A second removal submodule is configured to remove the available SMS modems whose sending performance has a sending success rate lower than a preset sending success rate threshold from the first available SMS modem set, to obtain a second available SMS modem set;

[0165] A weighting submodule, used for performing a weighting operation on the sending performance of each available SMS modem in the second available SMS modem set to obtain the comprehensive performance of each available SMS modem;

[0166] The third sorting submodule is used to sort according to the comprehensive performance, and select a preset number of available SMS modems as target SMS modems from the sorted second available SMS modem set according to the number of the queues of warning messages to be sent.

[0167] The sending module 304 is used to send the queue of warning messages to be sent to corresponding operation and maintenance personnel in sequence through a preset number of target SMS modems.

[0168] In one embodiment, the apparatus further comprises:

[0169] A monitoring module, used for monitoring the sending status of each warning message in the warning message queue;

[0170] An extraction module, used for obtaining the warning information whose sending status is failure, and obtaining the warning information that has not been sent;

[0171] The modification module is used to modify the priority of the unsent warning information to the highest priority, and put the unsent warning information back into the queue of warning information to be sent.

[0172] In this embodiment, by real-time monitoring of the early warning system, when the early warning system generates early warning information, the early warning information is collected to obtain a queue of early warning information to be sent, so that the early warning information can be captured in time and the duration of the fault can be reduced; by aggregating the early warning information into an early warning information queue, processing resources can be allocated more effectively, the efficiency of sending subsequent early warning information can be improved, and the fault situation can be ensured to be handled in a timely manner, thereby ensuring the stability of the system.

[0173] The pre-trained evaluation model can be used to accurately predict the sending performance of each available SMS modem, so that the warning information can be sent according to the sending performance of each available SMS modem in the subsequent process, thereby improving the sending success rate and efficiency of the warning information.

[0174] By selecting the target SMS modem based on the sending performance (such as sending rate and sending success rate), it can be ensured that the selected SMS modem has high working efficiency and reliability, thereby improving the sending efficiency and sending success rate of the warning information; by selecting the available SMS modem with better performance, the system resources can be used more effectively, avoiding the allocation of resources to SMS modems with poor performance, which can reduce resource waste and improve the overall system operation efficiency.

[0175] By using a preset number of target SMS cats, the queue of warning information to be sent can be quickly sent to the operation and maintenance personnel, which improves the efficiency and success rate of sending warning information; sending warning information through target SMS cats can ensure the accuracy and completeness of the warning information and avoid decision-making errors caused by information distortion or omissions.

[0176] In order to solve the above technical problems, the present application embodiment also provides a device (computer device). Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0177] The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that are interconnected and communicated through a system bus. It should be noted that the figure only shows a computer device 4 with a memory 41, a processor 42, and a network interface 43, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0178] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0179] The memory 41 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 41 can be an internal storage unit of the computer device 4, such as a hard disk or memory of the computer device 4. In other embodiments, the memory 41 can also be an external storage device of the computer device 4, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. equipped on the computer device 4. Of course, the memory 41 can also include both the internal storage unit of the computer device 4 and its external storage device. In this embodiment, the memory 41 is generally used to store the operating system and various application software installed on the computer device 4, such as computer-readable instructions of a method for sending short messages based on a depth prediction model, etc. In addition, the memory 41 can also be used to temporarily store various types of data that have been output or are to be output.

[0180] The processor 42 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip in some embodiments. The processor 42 is generally used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is used to run computer-readable instructions or process data stored in the memory 41, such as computer-readable instructions for running the text message sending method based on the depth prediction model.

[0181] The network interface 43 may include a wireless network interface or a wired network interface. The network interface 43 is generally used to establish a communication connection between the computer device 4 and other electronic devices.

[0182] During the implementation of the electronic device of the present application, by real-time monitoring of the early warning system, when the early warning system generates early warning information, the early warning information is collected to obtain a queue of early warning information to be sent, so as to capture the early warning information in time and reduce the duration of the fault; by aggregating the early warning information into an early warning information queue, processing resources can be allocated more effectively, the efficiency of sending subsequent early warning information can be improved, and the fault situation can be ensured to be handled in time, thereby ensuring the stability of the system.

[0183] The pre-trained evaluation model can be used to accurately predict the sending performance of each available SMS modem, so that the warning information can be sent according to the sending performance of each available SMS modem in the subsequent process, thereby improving the sending success rate and efficiency of the warning information.

[0184] By selecting the target SMS modem based on the sending performance (such as sending rate and sending success rate), it can be ensured that the selected SMS modem has high working efficiency and reliability, thereby improving the sending efficiency and sending success rate of the warning information; by selecting the available SMS modem with better performance, the system resources can be used more effectively, avoiding the allocation of resources to SMS modems with poor performance, which can reduce resource waste and improve the overall system operation efficiency.

[0185] By using a preset number of target SMS cats, the queue of warning information to be sent can be quickly sent to the operation and maintenance personnel, which improves the efficiency and success rate of sending warning information; sending warning information through target SMS cats can ensure the accuracy and completeness of the warning information and avoid decision-making errors caused by information distortion or omissions.

[0186] The present application also provides another implementation, namely, providing a medium (computer-readable storage medium), wherein the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the SMS sending method based on the deep prediction model as described above.

[0187] During the implementation process, the computer-readable storage medium of the present application monitors the early warning system in real time. When the early warning system generates early warning information, the early warning information is collected to obtain a queue of early warning information to be sent, so as to capture the early warning information in time and reduce the duration of the fault. By aggregating the early warning information into an early warning information queue, processing resources can be allocated more effectively, the efficiency of sending subsequent early warning information can be improved, and the fault situation can be ensured to be handled in time, thereby ensuring the stability of the system.

[0188] The pre-trained evaluation model can be used to accurately predict the sending performance of each available SMS modem, so that the warning information can be sent according to the sending performance of each available SMS modem in the subsequent process, thereby improving the sending success rate and efficiency of the warning information.

[0189] By selecting the target SMS modem based on the sending performance (such as sending rate and sending success rate), it can be ensured that the selected SMS modem has high working efficiency and reliability, thereby improving the sending efficiency and sending success rate of the warning information; by selecting the available SMS modem with better performance, the system resources can be used more effectively, avoiding the allocation of resources to SMS modems with poor performance, which can reduce resource waste and improve the overall system operation efficiency.

[0190] By using a preset number of target SMS cats, the queue of warning information to be sent can be quickly sent to the operation and maintenance personnel, which improves the efficiency and success rate of sending warning information; sending warning information through target SMS cats can ensure the accuracy and completeness of the warning information and avoid decision-making errors caused by information distortion or omissions.

[0191] The non-Company software tools or components appearing in the embodiments of this application are merely examples and do not represent actual use.

[0192] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, 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, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0193] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A method for sending short messages based on a deep prediction model, characterized in that: The steps include: Real-time monitoring of the early warning system, when the early warning system generates early warning information, collecting the early warning information to obtain a queue of early warning information to be sent; Obtaining an available SMS modem set from an SMS modem cluster, and performing performance prediction on each available SMS modem in the available SMS modem set using a pre-trained evaluation model to obtain a sending performance of each available SMS modem; According to the sending performance, selecting a preset number of target SMS modems from the set of available SMS modems; The queue of warning messages to be sent is sent to corresponding operation and maintenance personnel in sequence through a preset number of target SMS modems.

2. The method for sending short messages based on a deep prediction model according to claim 1, wherein: After collecting the warning information, the method further includes: Convert the warning information into a vector representation by using a preset vector conversion technology to obtain a warning vector; Extracting features from the warning vector using a pre-trained priority extraction model to obtain warning features; The warning features are classified by a preset classifier to obtain the priority corresponding to the warning information.

3. The method for sending short messages based on a deep prediction model as claimed in claim 2, characterized in that: The collecting of the warning information to obtain a queue of warning information to be sent includes: According to the priority of the warning information, the warning information is sorted in descending order according to the priority to obtain an initial warning information queue; The warning information of the same priority in the initial warning information queue is sorted according to the timestamp of the warning information to obtain the warning information queue.

4. The method for sending short messages based on a deep prediction model according to claim 1, wherein: Before predicting the available SMS modem by using the pre-trained evaluation model, the method further includes: Collecting historical data information sets of multiple SMS modems in each period, preprocessing the historical data information sets to obtain data sample sets; Based on the historical data information set of each period, the data sample set is divided into a training set and a test set according to a preset ratio; Constructing an initial evaluation model, and training the initial evaluation model using the training set to obtain an intermediate evaluation model; Testing the intermediate evaluation model using the test set to obtain an evaluation loss value; When the evaluation loss value is less than or equal to a preset evaluation loss threshold, obtaining the pre-trained evaluation model; When the evaluation loss value is greater than the preset evaluation loss threshold, the initial evaluation model is continuously trained using the training set until the evaluation loss value is less than or equal to the preset evaluation loss threshold, and the training is completed to obtain the pre-trained evaluation model.

5. The method for sending short messages based on a deep prediction model according to claim 1, wherein: The step of performing performance prediction on each available SMS modem in the set of available SMS modems by using a pre-trained evaluation model to obtain the sending performance of each available SMS modem includes: Acquire current data information of each available SMS modem, wherein the current data information includes current cycle information and current sent quantity; Using a pre-trained evaluation model, extract the features of the current cycle information and the current number of sent messages and combine them to obtain current data features; A predicted value of each of the available SMS modems is predicted according to the current data feature, and the predicted value is used as the sending performance of each of the available SMS modems, wherein the sending performance includes a sending rate and a sending success rate.

6. The method for sending short messages based on a deep prediction model according to claim 1, wherein: The step of selecting a preset number of target SMS modems from the set of available SMS modems according to the sending performance includes: Remove available SMS modems whose sending rates in the sending performance are less than a preset rate threshold from the available SMS modem set to obtain a first available SMS modem set; Remove the available SMS modems whose sending performance has a sending success rate lower than a preset sending success rate threshold from the first available SMS modem set, to obtain a second available SMS modem set; Performing a weighted operation on the sending performance of each available SMS modem in the second available SMS modem set to obtain a comprehensive performance of each available SMS modem; Sorting is performed according to the comprehensive performance, and according to the number of the queues of warning messages to be sent, a preset number of available SMS modems are selected from the sorted second available SMS modem set as target SMS modems.

7. The method for sending short messages based on a deep prediction model according to claim 1, characterized in that: After sending the queue of early warning information to be sent in sequence to corresponding operation and maintenance personnel, the method further includes: Monitor the sending status of each warning message in the warning message queue; Obtain the warning information that the sending status is failed, and obtain the warning information that was not sent; The priority of the unsent warning information is modified to the highest priority, and the unsent warning information is put back into the queue of warning information to be sent.

8. A text message sending device based on a deep prediction model, characterized in that: The device comprises: A monitoring module is used to monitor the early warning system in real time. When the early warning system generates early warning information, the early warning information is collected to obtain a queue of early warning information to be sent; An evaluation module is used to obtain an available SMS modem set from an SMS modem cluster, and to perform performance prediction on each available SMS modem in the available SMS modem set using a pre-trained evaluation model to obtain the sending performance of each available SMS modem; A selection module, configured to select a preset number of target SMS modems from the set of available SMS modems according to the sending performance; The sending module is used to send the queue of warning messages to be sent to the corresponding operation and maintenance personnel in sequence through a preset number of target SMS modems.

9. A computer device, characterized in that: The computer device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the SMS sending method based on the deep prediction model as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the text message sending method based on the deep prediction model as described in any one of claims 1 to 7 is implemented.

Images