A method, device and computer equipment for predicting respiratory signals

Abstract

The present application relates to a method, device and computer equipment for predicting a respiratory signal. The method comprises: obtaining a target respiratory signal to be predicted; selecting samples whose similarity with the target respiratory signal to be predicted is greater than a preset similarity threshold from the sample library according to the key time sequence element features of each sample in a preset respiratory signal sample library and the target respiratory signal to be predicted, and obtaining an adaptation signal set; selecting samples with the highest similarity with the target respiratory signal to be predicted from the adaptation signal set according to the trend component similarity and detail component similarity between each sample in the adaptation signal set and the target respiratory signal to be predicted, as adaptation samples; using the optimal complete prediction model corresponding to the adaptation sample as the target prediction model; and using the target prediction model to perform respiratory prediction on the target respiratory signal to be predicted. The use of this method can solve the problem that the prior art introduces a certain delay and wastes computing resources.

Description

Technical Field

[0001] The present application relates to the technical field of signal matching, and in particular to a method, device and computer equipment for predicting respiratory signals. Background Art

[0002] During radiotherapy, chest and abdominal tumors will change with respiratory movement. If the treatment is intervened according to the real-time measurement position of the target area, there will be a delay of hundreds of milliseconds, resulting in missed irradiation of the target area, affecting the treatment effect. In addition, after obtaining the position of the tumor target area, the treatment system needs a certain amount of time to respond. Therefore, it is necessary to predict the position of the target area after the delay by predicting the respiratory movement, adjust the target area position in advance, and improve the accuracy of radiation irradiation. The respiratory signal is a direct feedback signal of the respiratory movement. Therefore, accurate prediction of the respiratory signal is very important.

[0003] The existing technology mainly uses a preset respiratory signal prediction model to predict the respiratory signal to be predicted. If the preset respiratory signal prediction model predicts an inaccurate respiratory signal, it is replaced with another preset respiratory signal prediction model to predict the respiratory signal to be predicted until the accuracy of the prediction result of the adopted respiratory signal prediction model reaches the preset accuracy requirement. This method will introduce a certain delay, and the amount of calculation is large, and there is a problem of wasting computing resources.

[0004] There is no effective solution to the problem that existing technologies introduce certain delays and waste computing resources. Summary of the invention

[0005] Based on this, it is necessary to provide a method, device and computer equipment for predicting respiratory signals in response to the above technical problems.

[0006] In a first aspect, the present application provides a method for predicting a respiratory signal. The method comprises:

[0007] Acquire a target breathing signal to be predicted;

[0008] According to the key timing element features of each breathing signal sample in the preset breathing signal sample library and the key timing element features of the target breathing signal to be predicted, all breathing signal samples whose similarity with the target breathing signal to be predicted is greater than a preset similarity threshold are selected from the preset breathing signal sample library to obtain an adaptation signal set;

[0009] According to the similarities of each trend component and each detail component between each respiratory signal sample in the adaptation signal set and the target respiratory signal to be predicted, selecting a respiratory signal sample having the highest similarity with the target respiratory signal to be predicted from the adaptation signal set as an adaptation sample;

[0010] The optimal complete prediction model corresponding to the adapted sample is used as the target prediction model matched with the target respiratory signal to be predicted; the optimal complete prediction model corresponding to the adapted sample is the complete prediction model corresponding to the optimal prediction result obtained when the adapted sample is predicted by using multiple complete prediction models corresponding to the adapted sample respectively;

[0011] The target prediction model is used to perform respiration prediction on the target respiration signal to be predicted.

[0012] In one embodiment, according to the key timing element features of each breathing signal sample in the preset breathing signal sample library and the key timing element features of the target breathing signal to be predicted, a breathing signal sample having a similarity with the target breathing signal to be predicted greater than a preset similarity threshold is selected from the preset breathing signal sample library to obtain an adaptation signal set, including:

[0013] Determine each first similarity result between each of the breathing signal samples in the preset breathing signal sample library and the target breathing signal to be predicted according to the key timing element feature of each of the breathing signal samples in the preset breathing signal sample library, the key timing element feature of the target breathing signal to be predicted, and the similarity characterization coefficient of each of the key timing element features;

[0014] Based on each of the first similarity results between each of the breathing signal samples in the preset breathing signal sample library and the target breathing signal to be predicted, a breathing signal sample whose similarity result with the target breathing signal to be predicted is greater than a preset similarity threshold is selected from the preset breathing signal sample library to obtain the adaptation signal set.

[0015] In one embodiment, before determining each first similarity result between each of the breathing signal samples in the preset breathing signal sample library and the target breathing signal according to the key timing element feature of each of the breathing signal samples in the preset breathing signal sample library, the key timing element feature of the target breathing signal, and the similarity characterization coefficient of each of the key timing element features, the method includes:

[0016] Determine the key timing element feature of each of the respiratory signal samples in the preset respiratory signal sample library, and determine the key timing element feature of the target respiratory signal to be predicted;

[0017] Based on the total absolute error of the key timing element feature for all the breathing signal samples in the preset breathing signal sample library, a similarity characterization coefficient of each of the key timing element features is determined.

[0018] In one embodiment, before determining the key timing element feature of each of the respiratory signal samples in the preset respiratory signal sample library and determining the key timing element feature of the target respiratory signal to be predicted, the method includes:

[0019] Acquire the global time sequence element feature of each of the respiratory signal samples in the preset respiratory signal sample library;

[0020] Based on the global time-series meta-features of the respective breathing signal samples in the preset breathing signal sample library, determining the total absolute error of each time-series meta-feature in the global time-series meta-features for all breathing signal samples in the preset breathing signal sample library;

[0021] According to the total absolute error of each timing meta-feature in the global timing meta-feature for all the breathing signal samples in the preset breathing signal sample library, a preset number of timing meta-features with the smallest total absolute error for all the breathing signal samples in the preset breathing signal sample library are selected from the global timing meta-features and determined as key timing meta-features.

[0022] In one embodiment, determining the total absolute error of each timing meta-feature in the global timing meta-feature for all the breathing signal samples in the preset breathing signal sample library based on the global timing meta-feature of each breathing signal sample in the preset breathing signal sample library comprises:

[0023] For each of the timing element features in the global timing element features, determining the similarity between each of the breathing signal samples in the preset breathing signal sample library and other breathing signal samples in the preset breathing signal sample library according to the timing element features;

[0024] Based on the similarity between each of the breathing signal samples in the preset breathing signal sample library and other breathing signal samples in the preset breathing signal sample library, taking the breathing signal sample with the highest similarity to each of the breathing signal samples in the preset breathing signal sample library as each matching sample of each of the breathing signal samples in the preset breathing signal sample library;

[0025] Using the optimal complete prediction model corresponding to each matching sample of each breathing signal sample in the preset breathing signal sample library, predict each breathing signal sample in the breathing signal sample library to obtain the prediction error of each breathing signal sample;

[0026] The sum of the prediction errors of all the breathing signal samples is used as the total absolute error of the time sequence element feature for all the breathing signal samples in the preset breathing signal sample library;

[0027] All the time-series meta-features in the global time-series meta-features are traversed to obtain the total absolute error of each of the time-series meta-features for all the breathing signal samples in the preset breathing signal sample library.

[0028] In one of the embodiments, before selecting, from the adaptation signal set, a breathing signal sample having the highest similarity to the target breathing signal to be predicted as an adaptation sample according to the similarities of each trend component and each detail component between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted, the method comprises:

[0029] Decomposing each respiratory signal sample in the adapted signal set and each intrinsic mode function of the target respiratory signal to be predicted, and obtaining an intrinsic mode function result of each respiratory signal sample in the adapted signal set and an intrinsic mode function result of the target respiratory signal to be predicted;

[0030] Based on the intrinsic mode function results of each of the breathing signal samples in the adapted signal set, the intrinsic mode function result of the target breathing signal to be predicted, the intrinsic mode function results of each of the breathing signal samples in the adapted signal set and the intrinsic mode function result of the target breathing signal to be predicted, obtain the multi-scale permutation entropy of each of the breathing signal samples in the adapted signal set and the multi-scale permutation entropy of the target breathing signal to be predicted;

[0031] Determine the trend component and detail component of each of the respiratory signal samples in the adapted signal set and the trend component and detail component of the target respiratory signal to be predicted based on the multi-scale permutation entropy of each of the respiratory signal samples in the adapted signal set and the multi-scale permutation entropy of the target respiratory signal to be predicted;

[0032] Based on the trend component and the detail component of each of the breathing signal samples in the adaptation signal set, and the trend component and the detail component of the target breathing signal to be predicted, the similarities of the trend components and the similarities of the detail components between each of the breathing signal samples in the adaptation signal set and the target breathing signal to be predicted are determined.

[0033] In one embodiment, the selecting, from the adaptation signal set, a breathing signal sample having the highest similarity to the target breathing signal to be predicted as an adaptation sample according to the similarities of each trend component and each detail component between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted, comprises:

[0034] Determining each trend similarity weight and each detail similarity weight between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted based on the number of trend components and the number of detail components between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted;

[0035] Based on each trend similarity weight and each detail similarity weight between each breathing signal sample in the adapted signal set and the target breathing signal to be predicted, weighted summation is performed on the trend component similarity and the detail component similarity between each breathing signal sample in the adapted signal set and the target breathing signal to be predicted, to obtain each second similarity result between each breathing signal sample in the adapted signal set and the target breathing signal to be predicted;

[0036] Based on each second similarity result between each of the breathing signal samples in the adaptation signal set and the target breathing signal to be predicted, a breathing signal sample having the highest similarity to the target breathing signal to be predicted is selected from the adaptation signal set as the adaptation sample.

[0037] In one embodiment, the step of using the target prediction model to perform respiration prediction on the target respiration signal to be predicted includes:

[0038] Using the target breathing signal to be predicted, the target prediction model is trained to obtain a complete target prediction model;

[0039] The target respiratory signal to be predicted is predicted using the complete target prediction model.

[0040] In a second aspect, the present application also provides a device for predicting a respiratory signal. The device comprises:

[0041] A signal acquisition module, used to acquire a target respiratory signal to be predicted;

[0042] A signal set determination module, configured to select, from the preset respiratory signal sample library, all respiratory signal samples whose similarity to the target respiratory signal to be predicted is greater than a preset similarity threshold, according to key timing element features of each respiratory signal sample in the preset respiratory signal sample library and key timing element features of the target respiratory signal to be predicted, to obtain an adaptation signal set;

[0043] A sample determination module, configured to select, from the adaptation signal set, a breathing signal sample having the highest similarity to the target breathing signal to be predicted, as an adaptation sample, according to the similarities of each trend component and each detail component between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted;

[0044] A model determination module is used to use the optimal complete prediction model corresponding to the adapted sample as the target prediction model matched with the target respiratory signal to be predicted; the optimal complete prediction model corresponding to the adapted sample is a complete prediction model corresponding to the optimal prediction result obtained by respectively predicting the adapted sample using multiple complete prediction models corresponding to the adapted sample;

[0045] and a prediction module, which is used to perform respiration prediction on the target respiration signal to be predicted by using the target prediction model.

[0046] In a third aspect, the present application further provides a computer device, wherein the computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method for predicting a respiratory signal described in the first aspect is implemented.

[0047] The above-mentioned respiratory signal prediction method, device and computer equipment pre-match the trained optimal complete respiratory signal model for each respiratory signal sample in a preset respiratory signal sample library, and through secondary matching, that is, the first level matching is performed according to the key time sequence element characteristics, and the second level matching is performed according to the trend component similarity and the detail component similarity, to quickly and accurately match the adaptation sample with the highest similarity to the target respiratory signal to be predicted in the preset respiratory signal sample library. Because the adaptation sample has the highest similarity with the respiratory signal to be predicted, the optimal complete respiratory signal model of the adaptation sample is used to predict the respiratory signal to be predicted, which can ensure the accuracy of the respiratory signal prediction, and the prediction model does not need to be trained in the entire model determination process, and no delay is introduced. The secondary matching process has a small amount of calculation, which solves the problem that the prior art introduces a certain delay and wastes computing resources.

[0048] Details of one or more embodiments of the present application are set forth in the following drawings and description to make other features, objects, and advantages of the present application more readily apparent. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0050] Figure 1 A hardware structure block diagram of a terminal for a method for predicting a respiratory signal provided in an embodiment of the present application;

[0051] Figure 2 A flowchart of a method for predicting a respiratory signal provided in one embodiment of the present application;

[0052] Figure 3 A flowchart of a method for predicting a respiratory signal provided in a preferred embodiment of the present application;

[0053] Figure 4 This is a structural block diagram of a device for predicting a respiratory signal provided in one embodiment of the present application. DETAILED DESCRIPTION

[0054] In order to more clearly understand the purpose, technical solutions and advantages of the present application, the present application is described and illustrated below in conjunction with the accompanying drawings and embodiments.

[0055] Unless otherwise defined, the technical terms or scientific terms involved in this application shall have the general meaning understood by people with general skills in the technical field to which this application belongs. The words "one", "a", "the", "these" and the like in this application do not indicate a quantitative limitation, and they may be singular or plural. The terms "include", "comprise", "have" and any variants thereof involved in this application are intended to cover non-exclusive inclusions; for example, a process, method and system, product or device comprising a series of steps or modules (units) is not limited to the listed steps or modules (units), but may include unlisted steps or modules (units), or may include other steps or modules (units) inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like involved in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The "multiple" involved in this application refers to two or more. "And / or" describes the association relationship of associated objects, indicating that there may be three relationships, for example, "A and / or B" may mean: A exists alone, A and B exist at the same time, and B exists alone. Generally, the character " / " indicates that the objects associated with each other are in an "or" relationship. The terms "first", "second", "third", etc. in this application are only used to distinguish similar objects and do not represent a specific ordering of the objects.

[0056] The method embodiment provided in this embodiment can be executed in a terminal, a computer or a similar computing device. For example, it can be executed on a radiotherapy or chemotherapy device. Figure 1 FIG. 4 is a hardware structure diagram of a terminal of the method for predicting a respiratory signal of this embodiment. Figure 1 As shown, the terminal may include one or more ( Figure 1 Only one is shown in the figure) a processor 102 and a memory 104 for storing data, wherein the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA. The above terminal may also include a transmission device 106 and an input and output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is for illustration only and does not limit the structure of the above terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations shown.

[0057] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the prediction method of the respiratory signal in the present embodiment. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, to implement the above method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0058] The transmission device 106 is used to receive or send data via a network. The above network includes a wireless network provided by the communication provider of the terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (Radio Frequency, referred to as RF) module, which is used to communicate with the Internet wirelessly.

[0059] In this embodiment, a method for predicting a respiratory signal is provided. Figure 2 is a flow chart of the method for predicting the respiratory signal of this embodiment, such as Figure 2 As shown, the process includes the following steps:

[0060] Step S210, obtaining a target breathing signal to be predicted.

[0061] The target respiratory signal to be predicted may be a respiratory signal that needs to be predicted or analyzed by an algorithm or model. The specific method of obtaining the target respiratory signal to be predicted may be to obtain the target respiratory signal to be predicted by one or more methods such as a respiratory flow meter, a chest impedance sensor, an abdominal belt sensor, and an RPM (Remote Patient Monitoring) system.

[0062] Step S220, based on the key timing element features of each breathing signal sample in the preset breathing signal sample library and the key timing element features of the target breathing signal to be predicted, all breathing signal samples whose similarity with the target breathing signal to be predicted is greater than a preset similarity threshold are selected from the preset breathing signal sample library to obtain an adaptation signal set.

[0063] In this step, the above-mentioned preset breathing signal sample library is a sample library determined by the breathing signal samples of multiple users collected in advance. It should be noted that in order to ensure that the target breathing signal to be predicted can obtain matching samples with similarity conditions in the preset breathing signal sample library, it is necessary to ensure that the number of breathing signal samples in the preset breathing signal sample library is sufficient. The above-mentioned key timing meta-features can be timing meta-features whose influence on the similarity between breathing signals meets the preset threshold conditions, that is, they can be multiple timing meta-features with the greatest (most critical) influence on the similarity between breathing signals. For example, the above-mentioned key timing meta-features can be timing meta-features of the preset type that have the greatest influence on the similarity between breathing signals, selected from the global timing meta-features of the breathing signal. The above-mentioned timing meta-features can be key features extracted from the breathing signal to reflect the characteristic properties of the breathing pattern. Specifically, the above-mentioned time-series meta-features may be one of the respiratory cycle length, average respiratory rate, ratio of inhalation time to exhalation time, respiratory depth, tidal volume, baseline drift, spectrum peak position, power spectrum density, harmonic component, fractal dimension, entropy (such as sample entropy, approximate entropy, etc.), chaotic characteristics, waveform shape (such as peak, trough, etc.), rising edge and falling edge (i.e., the slope of the signal in the inhalation and exhalation phases), asymmetry, mean, variance, skewness, and kurtosis (the sharpness of the signal distribution) of the respiratory signal. The above-mentioned global time-series meta-features may be all types of time-series meta-features of the respiratory signal.

[0064] Among them, the above-mentioned preset similarity threshold can be specifically set according to specific needs. It should be noted that the size of the above-mentioned preset similarity threshold will affect the number of samples in the adaptation signal set. The larger the preset similarity threshold, the smaller the number of samples in the obtained adaptation signal set. The smaller the preset similarity threshold, the larger the number of samples in the obtained adaptation signal set. Based on this, the number of samples in the adaptation signal set can be adjusted by adjusting the size of the preset similarity threshold. The above-mentioned adaptation sample set can be a set of all respiratory signal samples in a preset respiratory signal sample library whose similarity with the target respiratory signal to be predicted is greater than the preset similarity threshold.

[0065] The specific process of obtaining the adaptation signal set mentioned above may be to first determine the similarity between each breathing signal sample in the preset breathing signal sample library and the target breathing signal to be predicted based on the key timing element characteristics of each breathing signal sample in the preset breathing signal sample library and the key timing element characteristics of the target breathing signal to be predicted, and obtain each first similarity result between each breathing signal sample in the preset breathing signal sample library and the target breathing signal to be predicted, and then, based on each first similarity result between each breathing signal sample in the preset breathing signal sample library and the target breathing signal to be predicted, select from the preset breathing signal sample library the breathing signal samples whose similarity results with the target breathing signal to be predicted meet preset conditions, and obtain the adaptation signal set.

[0066] Step S230, according to the similarities of each trend component and each detail component between each respiratory signal sample in the adaptation signal set and the target respiratory signal to be predicted, select the respiratory signal sample with the highest similarity to the target respiratory signal to be predicted from the adaptation signal set as the adaptation sample.

[0067] The above trend component similarity can be the similarity result of the trend component between the two respiratory signals. The above trend component can be the part of the respiratory signal that changes gradually over time. In addition, the trend component usually does not consider the details of rapid changes, but only considers the overall direction or change trend of the signal. The above trend component similarity can be the similarity result of the detail component between the two respiratory signals. The above detail component can be the part of the respiratory signal that changes rapidly, which is used to reflect the subtle differences and local features in the respiratory process. Because the speed of calculating the similarity between the two respiratory signals based on the trend component similarity and the similarity of each detail component is slow, and the speed of calculating the similarity between the two respiratory signals based on the time sequence element feature is fast, therefore, a faster similarity calculation method is selected, and the signals with higher similarity to the respiratory signal to be predicted in the preset respiratory signal sample library are preliminarily screened to obtain an adaptation signal set. Through preliminary screening, the redundancy caused by too many signal samples can be filtered out, the amount of calculation can be reduced, and the speed of subsequent matching can be improved.

[0068] In this step, according to the above-mentioned trend component similarities and detail component similarities between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted, the breathing signal sample with the highest similarity to the target breathing signal to be predicted is selected from the adaptation signal set as the adaptation sample. It can be that according to the trend component similarities and detail component similarities between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted, each second similarity result between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted is determined, and then, according to each second similarity result between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted, the breathing signal sample with the highest similarity to the target breathing signal to be predicted in the adaptation signal set is determined as the adaptation sample.

[0069] Step S240, the optimal complete prediction model corresponding to the adapted sample is used as the target prediction model matched with the target respiratory signal to be predicted; the optimal complete prediction model corresponding to the adapted sample is the complete prediction model corresponding to the optimal prediction result obtained when the adapted sample is predicted by using multiple complete prediction models corresponding to the adapted sample.

[0070] In order to ensure that after determining the adapted sample, the target prediction model that matches the target respiratory signal to be predicted can be quickly determined, it is necessary to predetermine the optimal complete prediction model of each respiratory signal sample in the preset respiratory signal sample library. Among them, the optimal complete prediction model of each respiratory signal sample in the preset respiratory signal sample library can be a complete prediction model corresponding to the optimal prediction result obtained by using multiple complete prediction models corresponding to the respiratory signal sample to predict the respiratory signal sample respectively. Before using multiple complete prediction models corresponding to the respiratory signal sample to predict the respiratory signal sample respectively and obtain the complete prediction model corresponding to the optimal prediction result, it is necessary to first determine the multiple complete prediction models corresponding to the respiratory signal sample. The above-mentioned determination of multiple complete prediction models corresponding to the respiratory signal sample can be that for each respiratory signal sample in the respiratory signal sample library, each initial prediction model in the preset multiple initial prediction models is trained respectively to obtain multiple complete prediction models corresponding to each respiratory signal sample in the respiratory signal sample library.

[0071] The following example illustrates the process of determining the optimal complete prediction model for each respiratory signal sample in the preset respiratory signal sample library:

[0072] For example, there are two respiratory signal samples, namely, respiratory signal sample A and respiratory signal sample B, in the preset respiratory signal sample library, and there are two respiratory signal models, namely, respiratory signal model D and respiratory signal model E.

[0073] Wherein, the breathing signal sample A corresponds to the first complete breathing signal model D and the first complete breathing signal model E, and the breathing signal sample B corresponds to the second complete breathing signal model D and the second complete breathing signal model E. The breathing signal sample A is predicted by the first complete breathing signal model D and the first complete breathing signal model E respectively, and the breathing signal sample B is predicted by the second complete breathing signal model D and the second complete breathing signal model E, so as to obtain the first prediction result of the breathing signal sample A and the second prediction result of the breathing signal sample A, the first prediction result of the breathing signal sample B and the second prediction result of the breathing signal sample B. If the first prediction result of the breathing signal sample A is better than the second prediction result of the breathing signal sample A, then the first complete breathing signal model D is the optimal complete prediction model of the breathing signal sample A. If the first prediction result of the breathing signal sample B is better than the second prediction result of the breathing signal sample B, then the second complete breathing signal model D is the optimal complete prediction model of the breathing signal sample B. Wherein, the first prediction result of the breathing signal sample A is the prediction result obtained by predicting the breathing signal sample A using the first complete breathing signal model D. The second prediction result of the above-mentioned breathing signal sample A is the prediction result obtained by predicting the breathing signal sample A using the first complete breathing signal model E. The first prediction result of the above-mentioned breathing signal sample B is the prediction result obtained by predicting the breathing signal sample B using the second complete breathing signal model D. The second prediction result of the above-mentioned breathing signal sample B is the prediction result obtained by predicting the breathing signal sample B using the second complete breathing signal model E. The above-mentioned first complete breathing signal model D is a complete breathing signal model that meets the preset requirements obtained by training the breathing signal model D according to the breathing signal sample A. The above-mentioned first complete breathing signal model E is a complete breathing signal model that meets the preset requirements obtained by training the breathing signal model E according to the breathing signal sample A. The above-mentioned second complete breathing signal model D is a complete breathing signal model that meets the preset requirements obtained by training the breathing signal model D according to the breathing signal sample B. The above-mentioned second complete breathing signal model E is a complete breathing signal model that meets the preset requirements obtained by training the breathing signal model E according to the breathing signal sample B.

[0074] Step S250: using the target prediction model to perform respiration prediction on the target respiration signal to be predicted.

[0075] The above steps S210 to S250 pre-match the trained optimal complete breathing signal model for each breathing signal sample in the preset breathing signal sample library, and through secondary matching, that is, the first level matching is performed according to the key time sequence element features, and the second level matching is performed according to the trend component similarity and the detail component similarity, so as to quickly and accurately match the adaptation sample with the highest similarity to the target breathing signal to be predicted in the preset breathing signal sample library. Because the adaptation sample has the highest similarity with the breathing signal to be predicted, the optimal complete breathing signal model of the adaptation sample is used to predict the breathing signal to be predicted, which can ensure the accuracy of the breathing signal prediction, and the prediction model does not need to be trained in the entire model determination process, and no delay is introduced. The secondary matching process has a small amount of calculation, which solves the problem that the prior art introduces a certain delay and wastes computing resources.

[0076] In one embodiment, based on step S220, according to the key timing element features of each breathing signal sample in the preset breathing signal sample library and the key timing element features of the target breathing signal to be predicted, a breathing signal sample having a similarity with the target breathing signal to be predicted greater than a preset similarity threshold is selected from the preset breathing signal sample library to obtain an adaptation signal set, including:

[0077] Step S222, based on the key timing meta-features of each breathing signal sample in the preset breathing signal sample library, the key timing meta-features of the target breathing signal to be predicted, and the similarity characterization coefficients of each key timing meta-feature, determine each first similarity result between each breathing signal sample in the preset breathing signal sample library and the target breathing signal to be predicted.

[0078] The similarity representation coefficients of the above-mentioned key time-series element features may be representation coefficients used to measure the importance of each key time-series element feature to the similarity between respiratory signals.

[0079] Among them, the above-mentioned first similarity results between each breathing signal sample in the preset breathing signal sample library and the target breathing signal to be predicted are determined according to the key timing element features of each breathing signal sample in the preset breathing signal sample library, the key timing element features of the target breathing signal to be predicted, and the similarity characterization coefficients of each key timing element features. The specific calculation process is as follows:

[0080]

[0081] in, It refers to the first similarity result between the nth respiratory signal sample in the respiratory signal sample library and the target respiratory signal to be predicted, N is the type of key time sequence element feature, is the similarity representation coefficient of the i-th key temporal element feature, is the i-th key temporal element feature of the target respiratory signal to be predicted, is the i-th key timing element feature of the n-th breathing signal sample in the breathing signal sample library.

[0082] Step S224, based on each first similarity result between each breathing signal sample in the preset breathing signal sample library and the target breathing signal to be predicted, a breathing signal sample whose similarity result with the target breathing signal to be predicted is greater than a preset similarity threshold is selected from the preset breathing signal sample library to obtain an adaptation signal set.

[0083] The above steps S222 to S224, through the determination of the adaptation signal set, facilitate the subsequent selection of the breathing signal sample with the highest similarity to the target breathing signal to be predicted from the adaptation signal set as the adaptation sample based on the similarities of each trend component and each detail component between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted.

[0084] Specifically, in one embodiment, before step S222, the following steps are included:

[0085] Step S2212, determining the key timing element features of each respiratory signal sample in the preset respiratory signal sample library, and determining the key timing element features of the target respiratory signal to be predicted.

[0086] The above-mentioned determination of the key timing meta-features of each respiratory signal sample in the preset respiratory signal sample library may be performed by extracting the key timing meta-features of each respiratory signal sample from the global timing meta-features of each respiratory signal sample in the preset respiratory signal sample library. The above-mentioned determination of the key timing meta-features of the target respiratory signal to be predicted may be performed by obtaining the key timing meta-features of the target respiratory signal to be predicted according to the preset timing meta-feature acquisition method. For example, the key timing meta-features include the length of the respiratory cycle and the ratio of the inhalation time to the exhalation time. The length of the respiratory cycle of the target respiratory signal to be predicted may be determined by counting the peak and valley values ​​of the target respiratory signal to be predicted. The time from the start of one inhalation to the end of inhalation of the target respiratory signal to be predicted may be the inhalation time, and the time from the end of inhalation to the start of the next inhalation may be the exhalation time, and then, the inhalation time may be divided by the exhalation time to obtain the ratio of the inhalation time to the exhalation time of the target respiratory signal to be predicted. It should be noted that the key timing meta-features do not only include the length of the respiratory cycle and the ratio of the inhalation time to the exhalation time, but are obtained according to actual calculations. Different timing meta-features are acquired in different ways. This embodiment only lists one of the cases. This embodiment does not specifically limit the specific methods for acquiring each timing meta-feature. As long as the preset timing meta-feature acquisition method can be used, the key timing meta-features of the target respiratory signal to be predicted can be acquired.

[0087] Step S2214, determining the similarity characterization coefficient of each key timing element feature based on the total absolute error of the key timing element feature for all breathing signal samples in the preset breathing signal sample library.

[0088] The calculation process of determining the similarity characterization coefficient of each key timing element feature based on the total absolute error of all breathing signal samples in the preset breathing signal sample library is as follows:

[0089]

[0090] in, is the total absolute error of the i-th key timing element feature for all breathing signal samples in the preset breathing signal sample library.

[0091] The above steps S2212 to S2214, through the determination of the key timing meta-features of each breathing signal sample in the preset breathing signal sample library, the key timing meta-features of the target breathing signal to be predicted, and the determination of the similarity characterization coefficients of each key timing meta-feature, facilitate the subsequent determination of each first similarity result between each breathing signal sample in the preset breathing signal sample library and the target breathing signal to be predicted based on the key timing meta-features of each breathing signal sample in the preset breathing signal sample library, the key timing meta-features of the target breathing signal to be predicted, and the similarity characterization coefficients of each key timing meta-features.

[0092] In addition, in one embodiment, the above step S2212, before determining the key timing element features of each respiratory signal sample in the preset respiratory signal sample library and determining the key timing element features of the target respiratory signal to be predicted, includes the following steps:

[0093] Step S2202, obtaining the global time-series element feature of each respiratory signal sample in a preset respiratory signal sample library.

[0094] Among them, obtaining the global time-series meta-features of each respiratory signal sample in the preset respiratory signal sample library can be to obtain the global time-series meta-features of each respiratory signal sample in the preset respiratory signal sample library according to the preset time-series meta-feature acquisition method. It should be noted that the above-mentioned preset time-series meta-feature acquisition method can be different according to the type of time-series meta-feature, and the specific preset time-series meta-feature acquisition method is not specifically limited in this embodiment, as long as the global time-series meta-features of each respiratory signal sample in the preset respiratory signal sample library can be obtained through the preset time-series meta-feature acquisition method.

[0095] Step S2204, based on the global time-series meta-features of each respiratory signal sample in the preset respiratory signal sample library, determine the total absolute error of each time-series meta-feature in the global time-series meta-features for all respiratory signal samples in the preset respiratory signal sample library.

[0096] In this step, the calculation formula for the total absolute error of the ath time sequence element feature in the above-mentioned global time sequence element feature for all respiratory signal samples in the preset respiratory signal sample library is as follows:

[0097]

[0098] in, is the pairing result obtained by taking the breathing signal samples with the highest similarity to the breathing signal samples in the preset breathing signal sample library as the matching samples of the breathing signal samples in the preset breathing signal sample library according to the a-th time sequence meta-feature in the time sequence meta-feature. j is the j-th breathing signal sample in the preset breathing signal sample library.

[0099] Step S2206, based on the total absolute errors of each timing meta-feature in the global timing meta-feature for all respiratory signal samples in the preset respiratory signal sample library, select a preset number of timing meta-features in the global timing meta-features with the smallest total absolute errors for all respiratory signal samples in the preset respiratory signal sample library, and determine them as key timing meta-features.

[0100] The above preset number can be set according to specific needs.

[0101] The above steps S2202 to S2206, by obtaining the global time-series meta-features of each respiratory signal sample in the preset respiratory signal sample library, further determine the total absolute error of each time-series meta-feature in the global time-series meta-feature for all respiratory signal samples in the preset respiratory signal sample library, so as to select the time-series meta-features with the smallest total absolute error for all respiratory signal samples in the preset respiratory signal sample library from the global time-series meta-features, and determine them as key time-series meta-features. By selecting the key time-series meta-features that have a greater impact on the similarity of the respiratory signals, it is convenient to remove the redundant time-series meta-features that have a small impact on the similarity of the respiratory signals. It is convenient to directly calculate the similarity of the respiratory signals based on the key time-series meta-features in the subsequent process, thereby reducing the amount of calculation.

[0102] In one embodiment, the above step S2204, based on the global time sequence meta-features of each respiratory signal sample in the preset respiratory signal sample library, determines the total absolute error of each time sequence meta-feature in the global time sequence meta-feature for all respiratory signal samples in the preset respiratory signal sample library, including:

[0103] Step S22041, for each time sequence meta-feature in the global time sequence meta-feature, determine the similarity between each breathing signal sample in the preset breathing signal sample library and other breathing signal samples in the preset breathing signal sample library according to the time sequence meta-feature.

[0104] As mentioned above, according to the timing element feature, determining the similarity between each breathing signal sample in the preset breathing signal sample library and other breathing signal samples in the preset breathing signal sample library can be based on the current timing element feature, and calculating the similarity between each breathing signal sample in the preset breathing signal sample library and other breathing signal samples in the preset breathing signal sample library. For example, the global timing element feature includes the breathing cycle length and the ratio of the inhalation time to the exhalation time, then according to the breathing cycle length, each first similarity between each breathing signal sample in the preset breathing signal sample library and other breathing signal samples in the preset breathing signal sample library is determined, and according to the ratio of the inhalation time to the exhalation time, each second similarity between each breathing signal sample in the preset breathing signal sample library and other breathing signal samples in the preset breathing signal sample library is determined.

[0105] Step S22042, based on the similarity between each breathing signal sample in the preset breathing signal sample library and other breathing signal samples in the preset breathing signal sample library, the breathing signal sample with the highest similarity to each breathing signal sample in the preset breathing signal sample library is used as each matching sample of each breathing signal sample in the preset breathing signal sample library.

[0106] If, for a time-series element feature such as the breathing cycle length, each first similarity between each breathing signal sample in the preset breathing signal sample library and other breathing signal samples in the preset breathing signal sample library is obtained, the breathing signal sample with the highest first similarity is used as each matching sample of each breathing signal sample in the preset breathing signal sample library. For example, the preset breathing signal library includes breathing signal sample A, breathing signal sample B, breathing signal sample C, and breathing signal sample D. If the calculation is based on the time sequence element feature of the breathing cycle length, the similarity between breathing signal sample A and breathing signal sample B is 10%, the similarity between breathing signal sample A and breathing signal sample C is 70%, the similarity between breathing signal sample A and breathing signal sample D is 60%, the similarity between breathing signal sample B and breathing signal sample C is 89%, the similarity between breathing signal sample B and breathing signal sample D is 77%, and the similarity between breathing signal sample C and breathing signal sample D is 99%, then breathing signal sample C is the matching sample of breathing signal sample A, breathing signal sample C is the matching sample of breathing signal sample B, breathing signal sample D is the matching sample of breathing signal sample C, and breathing signal sample C is the matching sample of breathing signal sample D. According to the above method, each matching sample of each breathing signal sample in the preset breathing signal sample library based on each time sequence element feature is determined.

[0107] Step S22043, using the optimal complete prediction model corresponding to each matching sample of each respiratory signal sample in the preset respiratory signal sample library, predict each respiratory signal sample in the respiratory signal sample library to obtain the prediction error of each respiratory signal sample.

[0108] If the preset breathing signal library includes breathing signal sample A, breathing signal sample B, breathing signal sample C and breathing signal sample D, and based on the time series element feature of the breathing cycle length, breathing signal sample C is a matching sample of breathing signal sample A, breathing signal sample C is a matching sample of breathing signal sample B, breathing signal sample D is a matching sample of breathing signal sample C, and breathing signal sample C is a matching sample of breathing signal sample D, then the optimal complete prediction model corresponding to each matching sample of each breathing signal sample in the preset breathing signal sample library is used to predict each breathing signal sample in the breathing signal sample library. The process of measuring and obtaining the prediction error of each breathing signal sample is as follows: using the optimal complete prediction model corresponding to breathing signal sample C to predict breathing signal sample A to obtain the prediction error of breathing signal sample A, using the optimal complete prediction model corresponding to breathing signal sample C to predict breathing signal sample B to obtain the prediction error of breathing signal sample B, using the optimal complete prediction model corresponding to breathing signal sample D to predict breathing signal sample C to obtain the prediction error of breathing signal sample C, and using the optimal complete prediction model corresponding to breathing signal sample C to predict breathing signal sample D to obtain the prediction error of breathing signal sample D. According to the above method, based on each time-series element feature, the optimal complete prediction model corresponding to each matching sample of each breathing signal sample in the preset breathing signal sample library is determined, and each breathing signal sample in the breathing signal sample library is predicted to obtain the prediction error of each breathing signal sample.

[0109] Step S22044: The sum of the prediction errors of all the respiratory signal samples is taken as the total absolute error of the time series meta-feature for all the respiratory signal samples in the preset respiratory signal sample library.

[0110] If the preset breathing signal library includes breathing signal sample A, breathing signal sample B, breathing signal sample C and breathing signal sample D, and the calculation is performed based on the time series meta-feature of the breathing cycle length, then the optimal complete prediction model corresponding to each matching sample of each breathing signal sample in the preset breathing signal sample library is used to predict each breathing signal sample in the breathing signal sample library, and the prediction error of each breathing signal sample is obtained as the prediction error of breathing signal sample A, the prediction error of breathing signal sample B, the prediction error of breathing signal sample C and the prediction error of breathing signal sample D. Then, the specific process of taking the sum of the prediction errors of all breathing signal samples as the total absolute error of all breathing signal samples in the preset breathing signal sample library by the time series meta-feature is as follows:

[0111] Calculate the sum of the prediction error of breathing signal sample A, the prediction error of breathing signal sample B, the prediction error of breathing signal sample C and the prediction error of breathing signal sample D, and use the calculation result as the total absolute error of the time series element feature of the breathing cycle length for all breathing signal samples in the preset breathing signal sample library.

[0112] Step S22045, traverse all the time sequence meta-features in the global time sequence meta-features to obtain the total absolute error of each time sequence meta-feature for all the respiratory signal samples in the preset respiratory signal sample library.

[0113] Specifically, by looping the above steps S22041 to S22044 and traversing all the time-series meta-features in the global time-series meta-features, the total absolute error of each time-series meta-feature for all the breathing signal samples in the preset breathing signal sample library can be obtained.

[0114] The above steps S22041 to S22045 calculate the total absolute errors of all respiratory signal samples in a preset respiratory signal sample library for each timing meta-feature, so as to facilitate the subsequent calculation of the total absolute errors of all respiratory signal samples in a preset respiratory signal sample library for each timing meta-feature in the global timing meta-feature, and select a preset number of timing meta-features with the smallest total absolute errors for all respiratory signal samples in a preset respiratory signal sample library from the global timing meta-feature, and determine them as key timing meta-features.

[0115] In addition, in one embodiment, before step S230, the following steps are included:

[0116] Step S225 , decomposing each respiratory signal sample in the adaptation signal set and each intrinsic mode function of the target respiratory signal to be predicted, and obtaining the intrinsic mode function results of each respiratory signal sample in the adaptation signal set and the intrinsic mode function results of the target respiratory signal to be predicted.

[0117] The above decomposition of each breathing signal sample in the adaptation signal set and each intrinsic mode function of the target breathing signal to be predicted may be performed by using an EMD (Empirical Mode Decomposition) algorithm to decompose each breathing signal sample in the adaptation signal set and each intrinsic mode function of the target breathing signal to be predicted.

[0118] Step S226, based on the intrinsic mode function results of each respiratory signal sample in the adaptation signal set and the intrinsic mode function results of the target respiratory signal to be predicted, obtain the multi-scale permutation entropy of each respiratory signal sample in the adaptation signal set and the multi-scale permutation entropy of the target respiratory signal to be predicted.

[0119] In this step, the multi-scale permutation entropy of each breathing signal sample in the adaptation signal set can be calculated by the intrinsic mode function results of each breathing signal sample in the adaptation signal set, and the multi-scale permutation entropy of the target breathing signal to be predicted can be calculated by the intrinsic mode function results of the target breathing signal to be predicted. The above-mentioned calculation of the multi-scale permutation entropy of each breathing signal sample in the adaptation signal set by the intrinsic mode function results of each breathing signal sample in the adaptation signal set can be based on the intrinsic mode function results of each breathing signal sample in the adaptation signal set, mapping the time series of the intrinsic mode function results of each breathing signal sample to the phase space, and then calculating the probability distribution of each breathing signal sample in the phase space, and then calculating the multi-scale permutation entropy of each breathing signal sample according to the probability distribution in the phase space of each breathing signal sample.

[0120] Step S227, based on the intrinsic mode function results of each respiratory signal sample in the adaptation signal set, the intrinsic mode function results of the target respiratory signal to be predicted, the multi-scale permutation entropy of each respiratory signal sample in the adaptation signal set and the multi-scale permutation entropy of the target respiratory signal to be predicted, determine the trend component and detail component of each respiratory signal sample in the adaptation signal set, and the trend component and detail component of the target respiratory signal to be predicted.

[0121] The above-mentioned intrinsic mode function results of each respiratory signal sample in the adaptation signal set, the intrinsic mode function results of the target respiratory signal to be predicted, the multi-scale permutation entropy of each respiratory signal sample in the adaptation signal set and the multi-scale permutation entropy of the target respiratory signal to be predicted, the trend component and detail component of each respiratory signal sample in the adaptation signal set, and the calculation process of the trend component and detail component of the target respiratory signal to be predicted are determined as follows:

[0122] Calculate the trend component TRC of the respiratory signal sample P in the adaptation signal set P The process is as follows:

[0123]

[0124] in, is the i-th intrinsic mode function result of the respiratory signal sample P in the adaptation signal set. is the i-th multi-scale permutation entropy of the respiratory signal sample P in the adaptation signal set.

[0125] Calculate the target breathing signal to be predicted The trend component The process is as follows:

[0126]

[0127] in, The target breathing signal to be predicted The i-th eigenmode function result of . The target breathing signal to be predicted The i-th multi-scale permutation entropy of .

[0128] Calculate the detail component DEC of the respiratory signal sample P in the adaptation signal set P The process is as follows:

[0129]

[0130] Calculate the target breathing signal to be predicted The amount of detail The process is as follows:

[0131]

[0132] Step S228, based on the trend components and detail components of each breathing signal sample in the adaptation signal set, and the trend components and detail components of the target breathing signal to be predicted, determine the similarities of each trend component and each detail component between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted.

[0133] The above-mentioned determination of the similarities of the trend components and the detail components between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted may be performed by respectively calculating the similarities of the trend components under morphological pattern coding and the detail components under symbol aggregation approximate coding between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted through the longest common subsequence. Specifically, the symbol sequence of the trend components of each breathing signal sample in the adaptation signal set and the symbol sequence of the trend components of the target breathing signal to be predicted may be calculated by an algorithm based on the morphological pattern, and the symbol sequence of the detail components of each breathing signal sample in the adaptation signal set and the symbol sequence of the detail components of the target breathing signal to be predicted may be calculated by a symbol aggregation approximate algorithm, and then, the trend component similarities under morphological pattern coding and the detail component similarities under symbol aggregation approximate coding may be calculated between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted through the longest common subsequence.

[0134] Specifically, the longest common subsequence is used to calculate the respiratory signal sample P and the target respiratory signal to be predicted The trend component similarity between the morphological patterns The calculation process is as follows:

[0135] Among them, L is the longest common subsequence length determined based on the trend component symbol sequence, and the calculation formula of L is as follows:

[0136]

[0137] Among them, L(i, j) is the first i elements in the respiratory signal sample P and the target respiratory signal to be predicted based on the trend component symbol sequence judgment The length of the longest common subsequence of the first j elements of , L(i-1, j) is the first i-1 elements in the respiratory signal sample P and the target respiratory signal to be predicted The length of the longest common subsequence of the first j elements of , L(i, j-1) is the first i elements in the respiratory signal sample P and the target respiratory signal to be predicted The length of the longest common subsequence of the first j-1 elements of , L(i-1, j-1) is the first i-1 elements in the respiratory signal sample P and the target respiratory signal to be predicted The length of the longest common subsequence of the first j-1 elements of . F P and The shortest time sequence length, F P is the symbol sequence of the trend component of the respiratory signal sample P The target breathing signal to be predicted The symbol sequence of the trend component of is the time length of the symbol sequence of the trend component of the respiratory signal sample P, The target breathing signal to be predicted The time series length of the symbol sequence of the trend component.

[0138] Specifically, the longest common subsequence is used to calculate the respiratory signal sample P and the target respiratory signal to be predicted Similarity of detail components under symbol aggregation approximate coding between The calculation process is as follows:

[0139]

[0140] Wherein, L' is the longest common subsequence length determined based on the detail component symbol sequence, and the calculation formula of L' is as follows:

[0141]

[0142] Among them, L´(i, j) is the first i elements in the respiratory signal sample P and the target respiratory signal to be predicted based on the detail component symbol sequence judgment The length of the longest common subsequence of the first j elements of . For E P and The shortest time sequence length, E P is the symbol sequence of the detail component of the respiratory signal sample P, The target breathing signal to be predicted The symbol sequence of the detail component, is the temporal length of the symbol sequence of the detail component of the respiratory signal sample P, The target breathing signal to be predicted The temporal length of the symbol sequence of the detail component.

[0143] In the above steps S225 to S228, by obtaining the intrinsic mode function results of each breathing signal sample in the adaptation signal set and the intrinsic mode function results of the target breathing signal to be predicted, and then, based on the intrinsic mode function results, obtaining the multi-scale permutation entropy of each breathing signal sample in the adaptation signal set and the multi-scale permutation entropy of the target breathing signal to be predicted, and then determining the trend component and the detail component, and then determining the similarities of each trend component and each detail component between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted based on the trend component and the detail component, which facilitates the subsequent calculation of the similarity between the breathing signals by determining the similarities of each trend component and each detail component.

[0144] In one embodiment, based on step S230, according to the similarities of each trend component and each detail component between each respiratory signal sample in the adaptation signal set and the target respiratory signal to be predicted, a respiratory signal sample having the highest similarity with the target respiratory signal to be predicted is selected from the adaptation signal set as the adaptation sample, including:

[0145] Step S232, based on the number of trend components and the number of detail components of each respiratory signal sample in the adaptation signal set and the target respiratory signal to be predicted, determine each trend similarity weight and each detail similarity weight between each respiratory signal sample in the adaptation signal set and the target respiratory signal to be predicted.

[0146] In this step, the trend similarity weight between the two respiratory signals is determined The calculation process is as follows:

[0147]

[0148] Among them, N TRC is the total number of trend components of the two respiratory signals whose similarity is to be determined, N EDM is the total number of trend components and detail components of the two respiratory signals whose similarity is to be determined.

[0149] Determine the trend similarity weight between two respiratory signals The calculation process is as follows:

[0150]

[0151] Among them, N DEC is the total number of detail components of the two respiratory signals whose similarity is to be determined.

[0152] Step S234, based on the trend similarity weights and detail similarity weights between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted, the trend component similarity and detail component similarity between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted are weightedly summed to obtain each second similarity result between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted.

[0153] The above calculation of the second similarity result between the two breathing signals The process is as follows:

[0154] Among them, TRCS is the detail component similarity between two respiratory signals, and DECS is the trend component similarity between two respiratory signals.

[0155] Step S236, based on each second similarity result between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted, select the breathing signal sample with the highest similarity to the target breathing signal to be predicted from the adaptation signal set as the adaptation sample.

[0156] In the above steps S232 to S236, by calculating each second similarity result between each respiratory signal sample in the adaptation signal set and the target respiratory signal to be predicted, the adaptation sample with the highest similarity to the target respiratory signal to be predicted is selected from the adaptation signal set. The determination of the adaptation sample facilitates the subsequent determination of the target prediction model based on the adaptation sample.

[0157] Furthermore, in one embodiment, the above step S250, using the target prediction model to perform respiration prediction on the target respiration signal to be predicted, includes:

[0158] Step S252, using the target breathing signal to be predicted, the target prediction model is trained to obtain a complete target prediction model.

[0159] Step S254, using the complete target prediction model, predicting the target breathing signal to be predicted.

[0160] In the above steps S252 to S254, the target prediction model is trained by using the target breathing signal to be predicted to obtain a complete target prediction model, and then the target breathing signal to be predicted is predicted using the complete target prediction model. The target prediction model is trained again to ensure the accuracy of the prediction of the target breathing signal to be predicted.

[0161] The present embodiment is described and illustrated below through preferred embodiments.

[0162] Figure 3 FIG. 1 is a flow chart of a method for predicting a respiratory signal provided by a preferred embodiment of the present application. Figure 3 As shown, the respiratory signal prediction method includes the following steps:

[0163] Step S310, obtaining a target breathing signal to be predicted;

[0164] Step S320, determining each first similarity result between each respiratory signal sample in the preset respiratory signal sample library and the target respiratory signal to be predicted according to the key timing element features of each respiratory signal sample in the preset respiratory signal sample library, the key timing element features of the target respiratory signal to be predicted, and the similarity characterization coefficient of each key timing element feature;

[0165] Step S330, based on each first similarity result between each breathing signal sample in the preset breathing signal sample library and the target breathing signal to be predicted, selecting a breathing signal sample whose similarity result with the target breathing signal to be predicted is greater than a preset similarity threshold from the preset breathing signal sample library, to obtain an adaptation signal set;

[0166] Step S340, determining each trend similarity weight and each detail similarity weight between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted based on the number of trend components and the number of detail components between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted;

[0167] Step S350, based on each trend similarity weight and each detail similarity weight between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted, weighted summation is performed on the trend component similarity and detail component similarity between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted, to obtain each second similarity result between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted;

[0168] Step S360, based on each second similarity result between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted, selecting a breathing signal sample with the highest similarity to the target breathing signal to be predicted from the adaptation signal set as the adaptation sample;

[0169] Step S370, using the optimal complete prediction model corresponding to the adapted sample as the target prediction model matched with the target respiratory signal to be predicted;

[0170] Step S380, using the target breathing signal to be predicted, training the target prediction model to obtain a complete target prediction model;

[0171] Step S390, using the complete target prediction model, predicting the target breathing signal to be predicted.

[0172] In the above steps S310 to S390, each breathing signal sample in the preset breathing signal sample library is pre-matched with the trained optimal complete breathing signal model, and through secondary matching, that is, the first level matching is performed according to the key timing element features, and the second level matching is performed according to the trend component similarity and the detail component similarity, so as to quickly and accurately match the adaptation sample with the highest similarity to the target breathing signal to be predicted in the preset breathing signal sample library. Because the adaptation sample has the highest similarity with the breathing signal to be predicted, the optimal complete breathing signal model of the adaptation sample is used to predict the breathing signal to be predicted, so as to ensure the accuracy of the breathing signal prediction, and the prediction model does not need to be trained in the entire model determination process, and no delay is introduced. The secondary matching process has a small amount of calculation, which solves the problem that the prior art introduces a certain delay and wastes computing resources.

[0173] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0174] Based on the same inventive concept, a breathing signal prediction device is also provided in this embodiment, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. The terms "module", "unit", "subunit", etc. used below can implement a combination of software and / or hardware for a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0175] In one embodiment, Figure 4 is a structural block diagram of a respiratory signal prediction device provided in an embodiment of the present application, such as Figure 4 As shown, the prediction device of the respiratory signal comprises:

[0176] A signal acquisition module 41 is used to acquire a target respiratory signal to be predicted;

[0177] The signal set determination module 42 is used to select, from the preset respiratory signal sample library, all respiratory signal samples whose similarity with the target respiratory signal to be predicted is greater than a preset similarity threshold, according to the key timing element features of each respiratory signal sample in the preset respiratory signal sample library and the key timing element features of the target respiratory signal to be predicted, to obtain an adaptation signal set;

[0178] A sample determination module 43, configured to select, from the adaptation signal set, a breathing signal sample having the highest similarity to the target breathing signal to be predicted, as an adaptation sample, according to the similarities of each trend component and each detail component between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted;

[0179] The model determination module 44 is used to use the optimal complete prediction model corresponding to the adaptation sample as the target prediction model matched with the target respiratory signal to be predicted; the optimal complete prediction model corresponding to the adaptation sample is the complete prediction model corresponding to the optimal prediction result obtained when the adaptation sample is predicted by using multiple complete prediction models corresponding to the adaptation sample respectively;

[0180] And a prediction module 45 is used to use the target prediction model to perform respiration prediction on the target respiration signal to be predicted.

[0181] The above-mentioned prediction device of the respiratory signal pre-matches each respiratory signal sample in the preset respiratory signal sample library with the trained optimal complete respiratory signal model, and through secondary matching, that is, the first level matching is performed according to the key time sequence element characteristics, and the second level matching is performed according to the trend component similarity and the detail component similarity, and the adaptation sample with the highest similarity to the target respiratory signal to be predicted in the preset respiratory signal sample library is quickly and accurately matched. Because the adaptation sample has the highest similarity with the respiratory signal to be predicted, the optimal complete respiratory signal model of the adaptation sample is used to predict the respiratory signal to be predicted, which can ensure the accuracy of the respiratory signal prediction, and the prediction model does not need to be trained in the entire model determination process, and no delay is introduced. The secondary matching process has a small amount of calculation, which solves the problem that the prior art introduces a certain delay and wastes computing resources.

[0182] In one of the embodiments, the signal set determination module includes a similarity calculation unit and a signal set determination unit;

[0183] A similarity calculation unit, used to determine each first similarity result between each breathing signal sample in the preset breathing signal sample library and the target breathing signal to be predicted according to the key timing element features of each breathing signal sample in the preset breathing signal sample library, the key timing element features of the target breathing signal to be predicted, and the similarity characterization coefficient of each key timing element feature;

[0184] The signal set determination unit is used to select, from the preset breathing signal sample library, breathing signal samples whose similarity results with the target breathing signal to be predicted are greater than a preset similarity threshold, based on each first similarity result between each breathing signal sample in the preset breathing signal sample library and the target breathing signal to be predicted, to obtain an adaptation signal set.

[0185] In one of the embodiments, the prediction device for the respiratory signal further includes a key time sequence element feature determination module and a characterization coefficient determination module;

[0186] A key timing element feature determination module is used to determine the key timing element features of each respiratory signal sample in a preset respiratory signal sample library, and determine the key timing element features of the target respiratory signal to be predicted;

[0187] The characterization coefficient determination module is used to determine the similarity characterization coefficient of each key timing element feature based on the total absolute error of the key timing element feature for all respiratory signal samples in a preset respiratory signal sample library.

[0188] In one of the embodiments, the prediction device of the respiratory signal further includes a global time sequence element feature acquisition module, a total absolute error determination module and a key time sequence element feature selection module;

[0189] A global time sequence element feature acquisition module, used to acquire the global time sequence element feature of each respiratory signal sample in a preset respiratory signal sample library;

[0190] A total absolute error determination module is used to determine the total absolute error of each timing element feature in the global timing element feature for all breathing signal samples in the preset breathing signal sample library based on the global timing element feature of each breathing signal sample in the preset breathing signal sample library;

[0191] The key timing meta-feature selection module is used to select a preset number of timing meta-features with the smallest total absolute errors for all respiratory signal samples in a preset respiratory signal sample library from the global timing meta-features according to the total absolute errors of each timing meta-feature in the global timing meta-features for all respiratory signal samples in a preset respiratory signal sample library, and determine them as key timing meta-features.

[0192] In one of the embodiments, the total absolute error determination module includes a similarity calculation unit, a matching sample determination unit, a prediction error determination unit, a total absolute error determination unit and a traversal unit;

[0193] A similarity calculation unit, for determining, for each time sequence meta-feature in the global time sequence meta-feature, the similarity between each breathing signal sample in the preset breathing signal sample library and other breathing signal samples in the preset breathing signal sample library according to the time sequence meta-feature;

[0194] a matching sample determination unit, configured to use, based on the similarity between each breathing signal sample in the preset breathing signal sample library and other breathing signal samples in the preset breathing signal sample library, a breathing signal sample with the highest similarity to each breathing signal sample in the preset breathing signal sample library as each matching sample of each breathing signal sample in the preset breathing signal sample library;

[0195] A prediction error determination unit, used to predict each respiratory signal sample in the respiratory signal sample library using the optimal complete prediction model corresponding to each matching sample of each respiratory signal sample in the preset respiratory signal sample library, and obtain the prediction error of each respiratory signal sample;

[0196] A total absolute error determination unit, used for taking the sum of the prediction errors of all the breathing signal samples as the total absolute error of the time series meta-feature for all the breathing signal samples of the preset breathing signal sample library;

[0197] The traversal unit is used to traverse all the time sequence meta-features in the global time sequence meta-features to obtain the total absolute error of each time sequence meta-feature for all the breathing signal samples in the preset breathing signal sample library.

[0198] In one of the embodiments, the breathing signal is predicted by the device further comprising an intrinsic mode function decomposition unit, a multi-scale permutation entropy determination unit, a component determination unit and a component similarity determination unit;

[0199] An intrinsic mode function decomposition unit is used to decompose each respiratory signal sample in the adaptation signal set and each intrinsic mode function of the target respiratory signal to be predicted, and obtain the intrinsic mode function results of each respiratory signal sample in the adaptation signal set and the intrinsic mode function results of the target respiratory signal to be predicted;

[0200] A multi-scale permutation entropy determination unit, for obtaining the multi-scale permutation entropy of each respiratory signal sample in the adaptation signal set and the multi-scale permutation entropy of the target respiratory signal to be predicted based on the intrinsic mode function results of each respiratory signal sample in the adaptation signal set and the intrinsic mode function results of the target respiratory signal to be predicted;

[0201] a component determination unit, for determining the trend component and detail component of each breathing signal sample in the adaptation signal set, and the trend component and detail component of the target breathing signal to be predicted based on the intrinsic mode function results of each breathing signal sample in the adaptation signal set, the intrinsic mode function results of the target breathing signal to be predicted, the multi-scale permutation entropy of each breathing signal sample in the adaptation signal set, and the multi-scale permutation entropy of the target breathing signal to be predicted;

[0202] A component similarity determination unit is used to determine the trend component similarities and detail component similarities between each respiratory signal sample in the adaptation signal set and the target respiratory signal to be predicted based on the trend components and detail components of each respiratory signal sample in the adaptation signal set and the trend components and detail components of the target respiratory signal to be predicted.

[0203] In one of the embodiments, the sample determination module includes: a similarity weight determination unit, a second similarity calculation unit, and an adaptation sample determination unit;

[0204] A similarity weight determination unit, for determining each trend similarity weight and each detail similarity weight between each respiratory signal sample in the adaptation signal set and the target respiratory signal to be predicted based on the number of trend components and the number of detail components between each respiratory signal sample in the adaptation signal set and the target respiratory signal to be predicted;

[0205] a second similarity calculation unit, configured to perform weighted summation of trend component similarities and detail component similarities between each respiratory signal sample in the adapted signal set and the target respiratory signal to be predicted based on each trend similarity weight and each detail similarity weight between each respiratory signal sample in the adapted signal set and the target respiratory signal to be predicted, to obtain each second similarity result between each respiratory signal sample in the adapted signal set and the target respiratory signal to be predicted;

[0206] The adaptation sample determination unit is used to select, from the adaptation signal set, a breathing signal sample with the highest similarity to the target breathing signal to be predicted as the adaptation sample based on each second similarity result between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted.

[0207] In one of the embodiments, the prediction module is used to train a target prediction model using a target breathing signal to be predicted to obtain a complete target prediction model; and to predict the target breathing signal to be predicted using the complete target prediction model.

[0208] It should be noted that the above modules can be functional modules or program modules, and can be implemented by software or hardware. For modules implemented by hardware, the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.

[0209] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, any one of the respiratory signal prediction methods in the above embodiments is implemented.

[0210] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0211] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0212] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for predicting a respiratory signal, characterized in that: The method comprises: Acquire a target breathing signal to be predicted; According to the key timing element features of each breathing signal sample in the preset breathing signal sample library and the key timing element features of the target breathing signal to be predicted, all breathing signal samples whose similarity with the target breathing signal to be predicted is greater than a preset similarity threshold are selected from the preset breathing signal sample library to obtain an adaptation signal set; According to the similarities of each trend component and each detail component between each respiratory signal sample in the adaptation signal set and the target respiratory signal to be predicted, selecting a respiratory signal sample having the highest similarity with the target respiratory signal to be predicted from the adaptation signal set as an adaptation sample; The optimal complete prediction model corresponding to the adapted sample is used as the target prediction model matched with the target respiratory signal to be predicted; the optimal complete prediction model corresponding to the adapted sample is the complete prediction model corresponding to the optimal prediction result obtained when the adapted sample is predicted by using multiple complete prediction models corresponding to the adapted sample respectively; The target prediction model is used to perform respiration prediction on the target respiration signal to be predicted.

2. The method for predicting a respiratory signal according to claim 1, characterized in that: According to the key timing element features of each breathing signal sample in the preset breathing signal sample library and the key timing element features of the target breathing signal to be predicted, a breathing signal sample having a similarity with the target breathing signal to be predicted greater than a preset similarity threshold is selected from the preset breathing signal sample library to obtain an adaptation signal set, including: Determine each first similarity result between each of the breathing signal samples in the preset breathing signal sample library and the target breathing signal to be predicted according to the key timing element feature of each of the breathing signal samples in the preset breathing signal sample library, the key timing element feature of the target breathing signal to be predicted, and the similarity characterization coefficient of each of the key timing element features; Based on each of the first similarity results between each of the breathing signal samples in the preset breathing signal sample library and the target breathing signal to be predicted, a breathing signal sample whose similarity result with the target breathing signal to be predicted is greater than a preset similarity threshold is selected from the preset breathing signal sample library to obtain the adaptation signal set.

3. The method for predicting a respiratory signal according to claim 2, characterized in that: Before determining each first similarity result between each of the breathing signal samples in the preset breathing signal sample library and the target breathing signal according to the key timing element feature of each of the breathing signal samples in the preset breathing signal sample library, the key timing element feature of the target breathing signal, and the similarity characterization coefficient of each of the key timing element features, the method includes: Determine the key timing element feature of each of the respiratory signal samples in the preset respiratory signal sample library, and determine the key timing element feature of the target respiratory signal to be predicted; Based on the total absolute error of the key timing element feature for all the breathing signal samples in the preset breathing signal sample library, the similarity representation coefficient of each of the key timing element features is determined.

4. The method for predicting a respiratory signal according to claim 3, characterized in that: Before determining the key timing element feature of each of the respiratory signal samples in the preset respiratory signal sample library and determining the key timing element feature of the target respiratory signal to be predicted, the method includes: Acquire the global time sequence element feature of each of the respiratory signal samples in the preset respiratory signal sample library; Based on the global time-series meta-features of the respective breathing signal samples in the preset breathing signal sample library, determining the total absolute error of each time-series meta-feature in the global time-series meta-features for all breathing signal samples in the preset breathing signal sample library; According to the total absolute error of each timing meta-feature in the global timing meta-feature for all the breathing signal samples in the preset breathing signal sample library, a preset number of timing meta-features with the smallest total absolute error for all the breathing signal samples in the preset breathing signal sample library are selected from the global timing meta-features and determined as key timing meta-features.

5. The method for predicting a respiratory signal according to claim 4, characterized in that: The determining, based on the global time-series meta-features of the respective breathing signal samples in the preset breathing signal sample library, the total absolute error of each time-series meta-feature in the global time-series meta-features for all breathing signal samples in the preset breathing signal sample library comprises: For each of the timing element features in the global timing element features, determining the similarity between each of the breathing signal samples in the preset breathing signal sample library and other breathing signal samples in the preset breathing signal sample library according to the timing element features; Based on the similarity between each of the breathing signal samples in the preset breathing signal sample library and other breathing signal samples in the preset breathing signal sample library, taking the breathing signal sample with the highest similarity to each of the breathing signal samples in the preset breathing signal sample library as each matching sample of each of the breathing signal samples in the preset breathing signal sample library; Using the optimal complete prediction model corresponding to each matching sample of each breathing signal sample in the preset breathing signal sample library, predict each breathing signal sample in the breathing signal sample library to obtain the prediction error of each breathing signal sample; The sum of the prediction errors of all the breathing signal samples is used as the total absolute error of the time sequence element feature for all the breathing signal samples in the preset breathing signal sample library; All the time-series meta-features in the global time-series meta-features are traversed to obtain the total absolute error of each of the time-series meta-features for all the breathing signal samples in the preset breathing signal sample library.

6. The method for predicting a respiratory signal according to claim 1, characterized in that: Before selecting, from the adaptation signal set, a breathing signal sample having the highest similarity to the target breathing signal to be predicted as an adaptation sample according to the similarities of each trend component and each detail component between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted, the method comprises: Decomposing each respiratory signal sample in the adapted signal set and each intrinsic mode function of the target respiratory signal to be predicted, and obtaining an intrinsic mode function result of each respiratory signal sample in the adapted signal set and an intrinsic mode function result of the target respiratory signal to be predicted; Based on the intrinsic mode function results of each of the respiratory signal samples in the adapted signal set and the intrinsic mode function results of the target respiratory signal to be predicted, obtaining the multi-scale permutation entropy of each of the respiratory signal samples in the adapted signal set and the multi-scale permutation entropy of the target respiratory signal to be predicted; Determine the trend component and detail component of each of the breathing signal samples in the adapted signal set, and the trend component and detail component of the target breathing signal to be predicted based on the intrinsic mode function results of each of the breathing signal samples in the adapted signal set, the intrinsic mode function results of the target breathing signal to be predicted, the multiscale permutation entropy of each of the breathing signal samples in the adapted signal set, and the multiscale permutation entropy of the target breathing signal to be predicted; Based on the trend component and the detail component of each of the breathing signal samples in the adaptation signal set, and the trend component and the detail component of the target breathing signal to be predicted, the similarities of the trend components and the similarities of the detail components between each of the breathing signal samples in the adaptation signal set and the target breathing signal to be predicted are determined.

7. The method for predicting a respiratory signal according to claim 1, characterized in that: The step of selecting, according to the similarities of each trend component and each detail component between each respiratory signal sample in the adaptation signal set and the target respiratory signal to be predicted, a respiratory signal sample having the highest similarity to the target respiratory signal to be predicted from the adaptation signal set as an adaptation sample comprises: Determining each trend similarity weight and each detail similarity weight between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted based on the number of trend components and the number of detail components between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted; Based on each trend similarity weight and each detail similarity weight between each breathing signal sample in the adapted signal set and the target breathing signal to be predicted, weighted summation is performed on the trend component similarity and the detail component similarity between each breathing signal sample in the adapted signal set and the target breathing signal to be predicted, to obtain each second similarity result between each breathing signal sample in the adapted signal set and the target breathing signal to be predicted; Based on each second similarity result between each of the breathing signal samples in the adaptation signal set and the target breathing signal to be predicted, a breathing signal sample having the highest similarity to the target breathing signal to be predicted is selected from the adaptation signal set as the adaptation sample.

8. The method for predicting a respiratory signal according to claim 1, characterized in that: The step of using the target prediction model to perform respiration prediction on the target respiration signal to be predicted includes: Using the target breathing signal to be predicted, the target prediction model is trained to obtain a complete target prediction model; The target respiratory signal to be predicted is predicted using the complete target prediction model.

9. A respiratory signal prediction device, characterized in that: The device comprises: A signal acquisition module, used for acquiring a target respiratory signal to be predicted; A signal set determination module, configured to select, from the preset respiratory signal sample library, all respiratory signal samples whose similarity to the target respiratory signal to be predicted is greater than a preset similarity threshold, according to key timing element features of each respiratory signal sample in the preset respiratory signal sample library and key timing element features of the target respiratory signal to be predicted, to obtain an adaptation signal set; A sample determination module, configured to select, from the adaptation signal set, a breathing signal sample having the highest similarity to the target breathing signal to be predicted, as an adaptation sample, according to the similarities of each trend component and each detail component between each breathing signal sample in the adaptation signal set and the target breathing signal to be predicted; A model determination module is used to use the optimal complete prediction model corresponding to the adapted sample as the target prediction model matched with the target respiratory signal to be predicted; the optimal complete prediction model corresponding to the adapted sample is a complete prediction model corresponding to the optimal prediction result obtained by respectively predicting the adapted sample using multiple complete prediction models corresponding to the adapted sample; and a prediction module, which is used to perform respiration prediction on the target respiration signal to be predicted by using the target prediction model.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the processor implements the steps of the method for predicting a respiratory signal according to any one of claims 1 to 8.

Images