Method and device for predicting trend of gap index of switch machine

Abstract

The invention provides a prediction method and device for a switch machine gap index trend, and the method comprises the steps: obtaining to-be-predicted data which comprises the operation data of a switch machine gap in a first preset time range; performing data feature processing on the to-be-predicted data; and based on the processed to-be-predicted data and a pre-trained LSTM prediction model, obtaining a prediction result of the switch machine gap index trend. By adopting the prediction method for the gap index trend of the switch machine, the change trend of the gap index of the switch machine can be accurately captured, problems possibly existing in the operation of the switch machine can be found in advance, the operation reliability of the switch machine is improved, and a data basis can be provided for the work of railway signal operation and maintenance personnel.

Description

Technical Field

[0001] One or more embodiments of the present specification relate to the field of railway signal systems, and in particular, to a method and device for predicting a trend of a switch gap index. Background Art

[0002] The railway signal system is a key technical facility to ensure the safety of train operation. As an important equipment in the railway signal system, the operating status of the switch has a significant impact on the safety and efficiency of railway transportation. The switch gap directly reflects the current operating status of the switch. The traditional switch gap index prediction method is mainly based on statistical analysis and empirical judgment, which makes it difficult to accurately capture the changing trend of the switch gap index. Summary of the invention

[0003] The present invention aims to solve at least one of the technical problems existing in the prior art or related art.

[0004] In view of this, one or more embodiments of the present specification provide a method and device for predicting the trend of a switch gap index, which can accurately capture the changing trend of the switch gap index.

[0005] According to a first aspect of one or more embodiments of this specification, a method for predicting a trend of a switch gap index is proposed, the method comprising:

[0006] Acquiring data to be predicted, wherein the data to be predicted includes operation data of the switch machine gap within a first preset time range;

[0007] Performing data feature processing on the data to be predicted;

[0008] Based on the processed data to be predicted and the pre-trained LSTM prediction model, the prediction results of the switch machine gap index trend are obtained.

[0009] In some optional embodiments, the performing data feature processing on the data to be predicted includes:

[0010] Constructing a data set to be predicted based on the data to be predicted;

[0011] The data set to be predicted is normalized.

[0012] In some optional embodiments, obtaining the prediction result of the switch machine gap index trend includes:

[0013] Obtaining the output result of the pre-trained LSTM prediction model;

[0014] The output result is subjected to a denormalization process to obtain a prediction result of the switch machine gap index trend.

[0015] In some optional embodiments, the training process of the pre-trained LSTM prediction model includes:

[0016] Build LSTM prediction model;

[0017] Acquiring historical data, wherein the historical data includes operation data of the switch machine gap within a second preset time range;

[0018] Performing data feature processing on the historical data to obtain a training data set and a verification data set;

[0019] Training the LSTM prediction model based on the training data set to obtain a trained LSTM prediction model;

[0020] The trained LSTM prediction model is adjusted and optimized based on the verification data set and the preset optimizer to obtain the trained LSTM prediction model.

[0021] In some optional embodiments, the adjusting and optimizing the trained LSTM prediction model based on the verification data set and a preset optimizer to obtain the trained LSTM prediction model includes:

[0022] Input the verification data set into the trained LSTM prediction model to obtain a prediction result;

[0023] Determining whether the prediction result meets the preset prediction accuracy;

[0024] If not, optimizing and updating the trained LSTM prediction model based on the preset optimizer, using the updated LSTM prediction model as the trained LSTM prediction model and adjusting and optimizing it;

[0025] If it meets the requirements, the currently trained LSTM prediction model is saved as the trained LSTM prediction model.

[0026] In some optional embodiments, constructing an LSTM prediction model includes:

[0027] Build the LSTM prediction model framework, including input layer, hidden layer and output layer;

[0028] Set the parameters of each layer;

[0029] Among them, the input layer includes a number of neurons, the number of which is equal to the number of data features of the data to be predicted; the number of hidden layers is 1 to 3, each layer includes 50-200 neurons, the activation function is Relu, and the output uses linear; the output layer includes 1 neuron with a linear activation function; the optimizer is Adam with a learning rate of 0.001.

[0030] In some optional embodiments, the method further comprises:

[0031] The prediction results of the switch machine gap index trend are visualized.

[0032] According to a second aspect of one or more embodiments of this specification, a device for predicting a trend of a switch gap index is provided, the device comprising:

[0033] A data acquisition module, used for acquiring data to be predicted, wherein the data to be predicted includes operation data of the switch machine gap within a first preset time range;

[0034] A data processing module, used for performing data feature processing on the data to be predicted;

[0035] The result acquisition module is used to obtain the prediction result of the switch machine gap index trend based on the processed data to be predicted and the pre-trained LSTM prediction model.

[0036] According to a third aspect of one or more embodiments of this specification, an electronic device is provided, the electronic device comprising:

[0037] processor;

[0038] a memory for storing processor-executable instructions;

[0039] The processor implements the method as described in the first aspect by running the executable instructions.

[0040] According to a fourth aspect of one or more embodiments of the present specification, a computer-readable storage medium is provided, on which computer instructions are stored, and when the instructions are executed by a processor, the steps of the method described in the first aspect are implemented.

[0041] It can be seen from the above technical solutions that in one or more embodiments of the present specification, the operating data of the switch gap within the first preset time range is obtained as the data to be predicted; data feature processing is performed on the data to be predicted; and the prediction result of the switch gap index trend is obtained based on the processed data to be predicted and the pre-trained LSTM prediction model. In the technical solution provided in the present application, a pre-trained LSTM prediction model is used to perform trend prediction on the collected data to be predicted, so as to achieve an accurate prediction of the future trend of the switch gap index. The use of this technical solution helps to discover possible problems in the operation of the switch in advance, improve the operational reliability of the switch, and provide a data basis for the work of railway signal operation and maintenance personnel.

[0042] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present application. Also, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0044] Figure 1 It is a flow chart of a method for predicting a trend of a switch machine gap index provided by an exemplary embodiment;

[0045] Figure 2 is a flowchart of a training LSTM prediction model provided by an exemplary embodiment;

[0046] Figure 3 is a flowchart of an optimized trained LSTM prediction model provided by an exemplary embodiment;

[0047] Figure 4 It is a schematic diagram of visual display of prediction results of switch machine gap index trend provided by an exemplary embodiment;

[0048] Figure 5 is a schematic structural diagram of an electronic device provided by an exemplary embodiment;

[0049] Figure 6 It is a block diagram of a device for predicting a trend of a switch machine gap index provided by an exemplary embodiment. DETAILED DESCRIPTION

[0050] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The specific methods described in the following exemplary embodiments do not represent all schemes consistent with one or more embodiments of this specification. Instead, they are only examples of devices and methods consistent with some aspects of one or more embodiments of this specification as detailed in the attached claims.

[0051] It should be noted that: in other embodiments, the steps of the corresponding method are not necessarily performed in the order shown and described in this specification. In some other embodiments, the steps included in the method may be more or less than those described in this specification. In addition, a single step described in this specification may be decomposed into multiple steps for description in other embodiments; and multiple steps described in this specification may be combined into a single step for description in other embodiments.

[0052] The railway signal system is a key technical facility to ensure the safety of train operation. The operating status of the switch as a railway signal equipment is related to the driving safety. The switch gap index is closely related to the operating status of the switch. The switch gap prediction is an important topic.

[0053] In the related technologies, linear regression, logistic regression and decision tree methods are mainly used to predict the switch gap index, but these three schemes have their defects and shortcomings, and cannot meet the requirements of the railway signal system in terms of the accuracy and stability of the prediction index:

[0054] For the linear regression method, linear regression assumes that there is a linear relationship between the independent variable and the dependent variable. If the actual relationship is nonlinear, the prediction result may be inaccurate. It is sensitive to outliers, and the presence of outliers may significantly affect the fitting effect of the model. However, most railway signal equipment indicators are nonlinear, that is, there will be a large deviation between the predicted value and the actual value;

[0055] As for the logistic regression method, it is less effective when processing nonlinearly separable data. It is sensitive to outliers, and outliers in the data may cause the model performance to deteriorate. In the field of railway signaling, when troubleshooting, the indicator value is often changed, which will mislead the logistic regression model;

[0056] As for the decision tree method, the stability of the decision tree is relatively poor, and different sample sets may lead to tree structures with large differences. That is, it has high requirements for railway signals and is not suitable for the scenario of railway signal-equipment-operating condition.

[0057] Therefore, developing a method that can accurately predict the trend of switch gap indicators is of great significance to improving the safety and efficiency of railway transportation.

[0058] Next, one or more embodiments of this specification are further described in conjunction with the following embodiments:

[0059] Figure 1 A flowchart of a method for predicting a switch gap index trend provided by an exemplary embodiment is provided. Figure 1 As shown, the method may include the following steps:

[0060] Step 101, obtaining data to be predicted, wherein the data to be predicted includes operation data of a switch machine gap within a first preset time range.

[0061] In this embodiment, the first preset time range is set to 10 days, the operating data of the switch gap is the static positioning data of the switch, and the data to be predicted is: the static positioning data of the switch in the past 10 days starting from the current date.

[0062] Specifically, a data acquisition device may be provided on the switch machine, which can collect and save operating data during the operation of the switch machine. In this embodiment, the required data to be predicted (static positioning data of the switch machine for the past 10 days starting from the current date) can be extracted from the operating data saved by the acquisition device.

[0063] In some other embodiments, all the operation data within the first preset time range can be extracted from the operation data stored in the data acquisition device provided on the switch machine, and then the static positioning data of the switch machine can be screened out from all the operation data; or, the data acquisition device provided on the switch machine can be connected to one or more data storage databases, and the required data to be predicted can be obtained from the corresponding data storage databases; or, the switch machine is provided with multiple data acquisition devices, and each data acquisition device only collects a certain type of data, and the required data to be predicted can be obtained from the corresponding data acquisition device. One or more embodiments of this specification do not further limit the specific method of obtaining the data to be predicted.

[0064] Step 103: performing data feature processing on the data to be predicted.

[0065] In this embodiment, after obtaining the data to be predicted, data feature processing can be performed on the data to be predicted according to the characteristics of the switch gap and prediction requirements, that is, features related to the switch gap index trend are extracted from the data to be predicted to construct a feature vector.

[0066] Specifically, in this embodiment, according to the characteristic that the switch gap is relatively stable when the natural environment is stable and the need to predict the future value of the gap when time and environment change, the maximum value, minimum value and average value are extracted from the switch operation data.

[0067] In one embodiment, when executing step 103, the following methods may be used but are not limited to:

[0068] Constructing a data set to be predicted based on the data to be predicted;

[0069] The data set to be predicted is normalized.

[0070] Continuing with the above example, after obtaining the static positioning data of the switch machine in the past 10 days, the maximum value, minimum value and average value of each day in the 10 days are extracted from the static positioning data of the switch machine in the past 10 days, and the extracted data is used as the data set to be predicted and then normalized. Here, the maximum and minimum value normalization (MinMaxScaler) is used. Among them, the normalization process is to eliminate the dimensional differences between the data in the data set to be predicted, and obtain pure data, which is convenient for subsequent prediction based on the data to be predicted and improves the prediction accuracy.

[0071] Step 105, based on the processed data to be predicted and the pre-trained LSTM prediction model, obtain the prediction result of the switch machine gap index trend.

[0072] The advantages of the long short-term memory (LSTM) neural network algorithm are mainly reflected in its ability to process long sequence data and good learning ability. In one or more embodiments of this specification, a LSTM prediction model is constructed based on the long short-term memory (LSTM) neural network algorithm, and the trained LSTM prediction model is used to predict the trend of the switch machine gap index, which improves the accuracy and stability of the prediction compared with the traditional prediction method.

[0073] In this embodiment, a trained LSTM prediction model is pre-established, namely, the pre-trained LSTM prediction model. After the data to be predicted is processed for data features, the processed data is input into the pre-trained LSTM prediction model, so that the pre-trained LSTM prediction model predicts the input data, and outputs a prediction result after the prediction is completed, which is the prediction result of the switch gap index trend.

[0074] Continuing with the above example, the normalized data set to be predicted is input into the pre-trained LSTM prediction model. After receiving the input data, the prediction model will make predictions based on the data and output the prediction results. In this embodiment, because the input data is normalized, the output prediction results are also normalized prediction results. The output results need to be denormalized to obtain the prediction results of the switch gap index trend; wherein, the denormalization uses the parameters of the previous step normalization to perform direction calculation.

[0075] In one embodiment, if Figure 2 As shown, the training process of the pre-trained LSTM prediction model may adopt but is not limited to the following methods and steps:

[0076] Step 201, constructing an LSTM prediction model.

[0077] When step 201 is executed, the LSTM prediction model framework may be constructed first, and then the framework may be initialized to complete the construction of the LSTM prediction model.

[0078] In this embodiment, an LSTM prediction model framework including an input layer, a hidden layer and an output layer is constructed, and then the initial parameters of each layer in the framework are set; wherein the input layer includes a number of neurons, and the number of neurons is equal to the number of data features of the data to be predicted; the number of hidden layers is 1 to 3, each layer includes 50-200 neurons, the activation function is Relu, and the output uses linear; the output layer includes 1 neuron with a linear activation function.

[0079] Specifically, in this embodiment, the input layer includes 3 neurons, corresponding to the maximum value, the minimum value and the average value; the number of hidden layers is 2, each layer includes 50 neurons; the batch size is 32, that is, the number of samples used in one iteration during model training is 32; the number of cycles (the number of times the entire data set used for training is traversed during the training process) ranges from 50 to 100, which can be adjusted according to the training results during the training process; Dropout is added between the input layer and the first hidden layer, and the dropout ratio is set to 0.2 to 0.5 to avoid overfitting.

[0080] Preferably, in this example, before training the LSTM prediction model, that is, during initialization, a parameter adjustment strategy such as network search or random search can be used to determine the best parameter combination, and the combination is used as the initial parameters of the LSTM prediction model.

[0081] Step 203, obtaining historical data, wherein the historical data includes operation data of the switch machine gap within a second preset time range.

[0082] In this embodiment, the second preset time range is set to 180 days, the operating data of the switch gap is the static positioning data of the switch, and the historical data is: the static positioning data of the switch for nearly 180 days starting from the current date.

[0083] The specific method of obtaining historical data can refer to the aforementioned method of obtaining the data to be predicted, which will not be repeated here.

[0084] Step 205: perform data feature processing on the historical data to obtain a training data set and a verification data set.

[0085] Continuing with the above example, after obtaining the static positioning data of the switch machine for nearly 180 days, the maximum value, minimum value and average value of each day in the 180 days are extracted from the static positioning data of the switch machine for nearly 180 days, and 80% of the extracted data are used as the training data set, and the remaining 20% ​​are used as the verification data set. The training data set and the verification data set are subjected to maximum and minimum value normalization processing (MinMaxScaler) to obtain the normalized training data set V and the normalized verification data set T; then, continuous data with a length of m are extracted from the normalized training data set V in sequence as the training set X train , extract the m+1th data as the training set X train The training comparison result Y train At the same time, sequentially extract continuous data of length m from the normalized validation data set T as the validation set X test , extract the m+1th data as the validation set X test Verification comparison result Y test In this embodiment, m is 3.

[0086] In other feasible embodiments, m may also take other values, such as 5.

[0087] In other feasible embodiments, the process of obtaining the training data set and the verification data set can also be that after extracting the maximum value, minimum value and average value of 180 days, the data set is divided into groups of four to obtain 45 data pairs, and the first three data in each group are marked as training data, and the fourth data is marked as comparison data; then the original data set is divided into a training data set and a verification data set in a ratio of 8:2; finally, normalization is performed to obtain a normalized training data set V and a normalized verification data set T.

[0088] Step 207: Train the LSTM prediction model based on the training data set to obtain a trained LSTM prediction model.

[0089] In this embodiment, the training set X train Input into the LSTM prediction model to get the prediction result Y train-pre ; According to the training prediction result Y train-pre Compare with the training result Y train Fine-tune the LSTM prediction model. For example, you can adjust the initial preset learning rate and dropout ratio to update the LSTM prediction model and obtain the trained LSTM prediction model.

[0090] Step 209: adjust and optimize the trained LSTM prediction model based on the verification data set and the preset optimizer to obtain the trained LSTM prediction model.

[0091] In this embodiment, if Figure 3 As shown, the process of adjusting and optimizing the trained LSTM prediction model may be performed in the following manners and steps, but not limited to:

[0092] Step 2091, input the verification data set into the trained LSTM prediction model to obtain the prediction result.

[0093] Continuing with the above example, before adjusting and optimizing the trained LSTM prediction model, we can set the prediction accuracy Acc of the LSTM prediction model and the optimizer to Adam with a learning rate of 0.001; test Input into the trained LSTM prediction model to obtain the verified prediction result Y test-pre .

[0094] It should be noted that, in other embodiments, the optimizer may also use other optimizers other than Adam, such as SGD, RMSprop, etc., or still use Adam, but its learning rate is not 0.001. This application does not make specific limitations on this.

[0095] Step 2093, determining whether the prediction result meets the preset prediction accuracy.

[0096] Continuing with the above example, according to the verification prediction result Y test-pre Verify the comparison result Y test Calculate the error Acc of the currently trained LSTM prediction model test For example, using R 2 Error calculation method; Compare the error Acc test and prediction accuracy Acc (preset prediction accuracy), if the error Acc test is less than or equal to the prediction accuracy Acc, then it meets the preset prediction accuracy. If the error Acc test If it is greater than the prediction accuracy Acc, it does not meet the preset prediction accuracy.

[0097] In this embodiment, R 2 Error calculation method to obtain the error Acc test , not only the size of the prediction error is taken into account, but also the model's ability to explain the fluctuations in the true value, so it can more comprehensively evaluate the performance of the model.

[0098] Step 2095: If it does not meet the requirements, the trained LSTM prediction model is optimized and updated based on the preset optimizer, and the updated LSTM prediction model is used as the trained LSTM prediction model and adjusted and optimized.

[0099] Continuing with the above example, if in step 2093, it is determined that the prediction accuracy of the currently trained LSTM prediction model does not meet the preset prediction accuracy, the Adam with an initial learning rate of 0.001 set when the LSTM prediction model was constructed or other optimizers newly set in step 207 are used to optimize and update the model, and the updated model is used as the trained LSTM prediction model, and steps 2091-2093 are re-executed.

[0100] In a feasible embodiment, when executing step 2095, before or after using the optimizer to adjust the LSTM prediction model, the aforementioned steps 201 to 209 can be executed again to train the currently trained LSTM prediction model again.

[0101] Step 2097, if it meets the requirements, save the currently trained LSTM prediction model as the trained LSTM prediction model.

[0102] Continuing with the above example, if in step 2093, it is determined that the prediction accuracy of the currently trained LSTM prediction model meets the preset prediction accuracy, it means that the model optimization is completed, and the currently trained LSTM prediction model is used as the pre-trained LSTM prediction model.

[0103] In one feasible embodiment, when executing step 105, if Figure 4 As shown, in order to facilitate the operation and maintenance personnel to intuitively understand the changing trend of the switch machine gap index, the obtained prediction results of the switch machine gap index trend can also be visualized through visualization technology.

[0104] In the above embodiment, the operating data of the switch gap within the first preset time range is obtained as the data to be predicted; data feature processing is performed on the data to be predicted; and the prediction result of the switch gap index trend is obtained based on the processed data to be predicted and the pre-trained LSTM prediction model. In the technical solution provided in the present application, a pre-trained LSTM prediction model is used to perform trend prediction on the collected data to be predicted, so as to achieve an accurate prediction of the future trend of the switch gap index. The use of this technical solution helps to discover possible problems in the operation of the switch in advance, improve the operational reliability of the switch, and provide a data basis for the work of railway signal operation and maintenance personnel.

[0105] Corresponding to the aforementioned embodiment of a method, the present application also provides an embodiment of a device for predicting a trend of a switch gap index.

[0106] An embodiment of a device for predicting a trend of a switch gap index in the present application can be applied to electronic devices. Figure 4is a schematic structural diagram of an electronic device in an exemplary embodiment. Figure 5 At the hardware level, the electronic device includes a processor 502, an internal bus 504, a network interface 506, a memory 508, and a non-volatile memory 510, and may also include hardware required for other services. The processor 502 reads the corresponding computer program from the non-volatile memory 510 into the memory and then runs it, forming a synchronization device for the user login status at the logical level. One or more embodiments of this specification can be implemented based on software, such as the processor 502 reading the corresponding computer program from the non-volatile memory 510 into the memory 508 and then running it. Of course, in addition to the software implementation method, this application does not exclude other implementation methods, such as logic devices or a combination of software and hardware, etc., that is, the execution subject of the following processing flow is not limited to each logical unit, but can also be hardware or logic devices.

[0107] Please refer to Figure 6 , Figure 6 is a block diagram of a prediction device for a switch machine gap index trend in an exemplary embodiment, such as Figure 6 As shown, the device may include:

[0108] A data acquisition module 1 is used to acquire data to be predicted, wherein the data to be predicted includes operation data of the switch gap within a first preset time range;

[0109] Data processing module 2, used for performing data feature processing on the data to be predicted;

[0110] The result acquisition module 3 is used to obtain the prediction result of the switch machine gap index trend based on the processed data to be predicted and the pre-trained LSTM prediction model.

[0111] Optionally, the data processing module 2 is specifically used for:

[0112] Constructing a data set to be predicted based on the data to be predicted;

[0113] The data set to be predicted is normalized.

[0114] Optionally, the result acquisition module 3 is specifically used for:

[0115] Obtaining the output result of the pre-trained LSTM prediction model;

[0116] The output result is subjected to a denormalization process to obtain a prediction result of the switch machine gap index trend.

[0117] Optionally, the result acquisition module 3 can also be used to: visualize the prediction results of the switch machine gap index trend.

[0118] The implementation process of the functions and effects of each unit in the above-mentioned device is specifically described in the implementation process of the corresponding steps in the above-mentioned method, and will not be repeated here.

[0119] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can refer to the partial description of the method embodiments. The device embodiments described above are only schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of at least one embodiment of the present disclosure. A person of ordinary skill in the art can understand and implement them without paying creative labor.

[0120] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer, which may be in the form of a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email transceiver, a game console, a tablet computer, a wearable device or a combination of any of these devices.

[0121] In a typical configuration, a computer includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0122] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0123] Computer-readable media include permanent and non-permanent, removable and non-removable media that can be used to store information by any method or technology. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, disk storage, quantum memory, graphene-based storage media or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include temporary computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0124] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0125] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0126] The terms used in one or more embodiments of this specification are only for the purpose of describing specific embodiments, and are not intended to limit one or more embodiments of this specification. The singular forms of "a", "said" and "the" used in one or more embodiments of this specification and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0127] It should be understood that although the terms first, second, third, etc. may be used to describe various information in one or more embodiments of this specification, these information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of one or more embodiments of this specification, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0128] The above description is merely a preferred embodiment of one or more embodiments of the present specification and is not intended to limit one or more embodiments of the present specification. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of one or more embodiments of the present specification shall be included in the scope of protection of one or more embodiments of the present specification.

Claims

1. A method for predicting the trend of switch gap index, characterized in that: The method comprises: Acquiring data to be predicted, wherein the data to be predicted includes operation data of the switch machine gap within a first preset time range; Performing data feature processing on the data to be predicted; Based on the processed data to be predicted and the pre-trained LSTM prediction model, the prediction results of the switch machine gap index trend are obtained.

2. The method according to claim 1, characterized in that The performing data feature processing on the data to be predicted includes: Constructing a data set to be predicted based on the data to be predicted; The data set to be predicted is normalized.

3. The method according to claim 2, characterized in that Obtaining the prediction result of the switch machine gap index trend, including: obtaining the output result of the pre-trained LSTM prediction model; The output result is subjected to a denormalization process to obtain a prediction result of the switch machine gap index trend.

4. The method according to claim 1, characterized in that: The training process of the pre-trained LSTM prediction model includes: Build LSTM prediction model; Acquiring historical data, wherein the historical data includes operation data of the switch machine gap within a second preset time range; Performing data feature processing on the historical data to obtain a training data set and a verification data set; Training the LSTM prediction model based on the training data set to obtain a trained LSTM prediction model; The trained LSTM prediction model is adjusted and optimized based on the verification data set and the preset optimizer to obtain the trained LSTM prediction model.

5. The method according to claim 4, characterized in that The step of adjusting and optimizing the trained LSTM prediction model based on the verification data set and the preset optimizer to obtain the trained LSTM prediction model includes: Input the verification data set into the trained LSTM prediction model to obtain a prediction result; Determining whether the prediction result meets the preset prediction accuracy; If not, optimizing and updating the trained LSTM prediction model based on the preset optimizer, using the updated LSTM prediction model as the trained LSTM prediction model and adjusting and optimizing it; If it meets the requirements, the currently trained LSTM prediction model is saved as the trained LSTM prediction model.

6. The method according to claim 4 or 5, characterized in that: Build an LSTM prediction model, including: Build the LSTM prediction model framework, including input layer, hidden layer and output layer; Set the parameters of each layer; Among them, the input layer includes a number of neurons, the number of which is equal to the number of data features of the data to be predicted; the number of hidden layers is 1 to 3, each layer includes 50-200 neurons, the activation function is Relu, and the output uses linear; the output layer includes 1 neuron with a linear activation function; the optimizer is Adam with a learning rate of 0.

001.

7. The method according to any one of claims 1 to 6, characterized in that: The method further comprises: The prediction results of the switch machine gap index trend are visualized.

8. A device for predicting the trend of switch machine gap index, characterized in that: The device comprises: A data acquisition module, used for acquiring data to be predicted, wherein the data to be predicted includes operation data of the switch machine gap within a first preset time range; A data processing module, used for performing data feature processing on the data to be predicted; The result acquisition module is used to obtain the prediction result of the switch machine gap index trend based on the processed data to be predicted and the pre-trained LSTM prediction model.

9. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor implements the method according to any one of claims 1 to 7 by running the executable instructions.

10. A computer-readable storage medium having computer instructions stored thereon, characterized in that: When the instruction is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images