Hemodialysis data processing method and system, computer and storage medium

Abstract

The invention provides a hemodialysis data processing method and system, a computer and a storage medium. The method comprises the following steps that blood pressure waveform data and unbalance syndrome data in the dialysis process are collected; and calculating a metabolic state transition matrix based on the dynamic characteristic difference of the spatial-temporal characteristic tensor, determining state attribution based on the metabolic state transition matrix, and judging whether to output an early warning signal or not by calculating a blood pressure fluctuation factor and an unbalance risk factor. The overall deviation degree of the current state and the target metabolism state is quantified through a metabolism state transition matrix, state attribution is determined based on Softmax probability normalization, then whether early warning is determined or not is judged by calculating a blood pressure fluctuation factor and an unbalance risk factor, dialysis complication early warning is upgraded from'isolated parameter threshold judgment 'to'multi-dimensional process quantification', and the accuracy of dialysis complication early warning is improved. Early intervention and accurate resource allocation are realized, and the monitoring effect is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of hemodialysis, in particular to a hemodialysis data processing method and system, a computer and a storage medium. BACKGROUND

[0002] Hemodialysis is one of the renal replacement therapy methods for patients with acute and chronic renal failure. It removes metabolic waste, maintains electrolyte and acid-base balance, and removes excess water in the body by inducing blood to flow out of the body, through the dialyzer, and exchanging materials between the inside and outside of the hollow fiber through diffusion, ultrafiltration, adsorption and convection principle. And the purified blood is returned to achieve hemodialysis.

[0003] Complications are prone to occur during hemodialysis, including common hypotension and imbalance syndrome. The conventional detection method is to monitor the systolic blood pressure and determine whether there is a risk of hypotension by a specific threshold detection; or determine whether there is a risk of imbalance syndrome by brain imaging, clinical symptoms, etc. The above methods of determining complications by specific threshold detection, specific examination or clinical experience cannot timely capture the pathological signals before acute events, and the monitoring effect is not good. SUMMARY

[0004] In view of the deficiencies of the prior art, the purpose of the present application is to provide a hemodialysis data processing method and system, a computer and a storage medium, which aims to solve the technical problems that the prior art cannot timely capture the pathological signals before acute events and the monitoring effect is not good.

[0005] To achieve the above purpose, in a first aspect, the present application provides a hemodialysis data processing method, comprising the following steps: Collecting blood pressure waveform data and imbalance syndrome data during dialysis to construct a spatiotemporal feature tensor of data changes; Calculating a metabolic state transition matrix based on the dynamic feature difference of the spatiotemporal feature tensor to determine state attribution based on the metabolic state transition matrix, the metabolic state including a steady state, a hypotension early warning state and an imbalance syndrome risk state; If the metabolic state migrates to the hypotension early warning state, calculating a blood pressure fluctuation factor based on the blood pressure waveform data, and if the metabolic state migrates to the imbalance syndrome risk state, calculating an imbalance risk factor based on the imbalance syndrome data: If the blood pressure fluctuation factor is greater than a first preset value or the imbalance risk factor is greater than a second preset value, outputting a warning signal.

[0006] According to an aspect of the above technical solution, the blood pressure fluctuation factor comprises an acute factor and a chronic factor, the first preset value comprises a first threshold value and a second threshold value corresponding to the acute factor and the chronic factor respectively, and the calculation expression of the blood pressure fluctuation factor is: ; ; In the formula, is the blood pressure fluctuation factor, is the second derivative of the blood pressure waveform signal corresponding to the acute factor, is the energy probability distribution of the blood pressure signal after wavelet decomposition, H is the entropy value of the energy probability distribution corresponding to the chronic factor, and t is the signal sampling time; The calculation expression of the imbalance risk factor is: ; In the formula, is the imbalance risk factor, is the change amount of the blood sodium concentration, is the time interval, is the change amount of the cerebral oxygen saturation, is the baseline value of the cerebral oxygen saturation, is the maximum viscosity value, are weight coefficients corresponding to the blood sodium concentration and the cerebral oxygen saturation respectively.

[0007] According to an aspect of the above technical solution, the step of calculating the metabolic state transition matrix based on the dynamic feature difference of the space-time feature tensor specifically comprises: quantifying the deviation degree of the current physiological state and the target state by the Frobenius norm; based on the deviation degree and by the metabolic state transition matrix, realizing multi-state transition risk assessment according to Softmax probability normalization.

[0008] According to an aspect of the above technical solution, the calculation expression of the metabolic state transition matrix is: ; In the formula, is the metabolic state transition matrix, is the feature difference tensor corresponding to the target state, is the cluster center of the target state, is the Frobenius norm, and k is the metabolic state category index, is the cluster center of the kth metabolic state.

[0009] According to an aspect of the above technical solution, the step of updating the cluster center of the metabolic state specifically comprises: Extract the feature difference tensors of several consecutive time periods in the current batch, and filter the feature difference tensors belonging to the target state based on the indicator function; The sample weight coefficient of the target state sample is calculated based on the number of target state samples in the historical state and the number of target state samples in the current batch. The state affiliation is determined based on the metabolic state transition matrix, and the cluster center of the target state is updated in combination with the sample weight coefficients. Then, the metabolic state transition matrix for the next time period is calculated based on the updated cluster center of the target state.

[0010] According to one aspect of the above technical solution, the calculation expression for the update mechanism of the cluster centers of metabolic states is as follows: ; ; In the formula, The cluster centers of the updated target state of class j, The cluster centers of the j-th target state before the update. The sample weight coefficients are... This refers to the set of samples in the current batch that belong to the target state. For the feature difference tensor, This represents the number of samples in the current batch that belong to the target state. This represents the number of target state samples in the historical states.

[0011] According to one aspect of the above technical solution, the step of determining the state affiliation based on the metabolic state transition matrix specifically includes: Based on the metabolic state transition matrix, the probability of belonging to the target state is calculated. If the probability of belonging is greater than an empirical threshold, the current metabolic state is determined to be the target state.

[0012] Secondly, the present invention provides a hemodialysis data processing system, comprising: The data acquisition module is used to collect blood pressure waveform data and imbalance syndrome data during dialysis in order to construct a spatiotemporal feature tensor of data changes; The attribution module is used to calculate the metabolic state transition matrix based on the dynamic feature difference of the spatiotemporal feature tensor, and to determine the state attribution based on the metabolic state transition matrix. The metabolic state includes a stable state, a hypotension warning state, and an imbalance syndrome risk state. The calculation module is used to calculate a blood pressure fluctuation factor based on the blood pressure waveform data if the metabolic state transitions to a hypotension warning state, and to calculate an imbalance risk factor based on the imbalance syndrome data if the metabolic state transitions to an imbalance syndrome state. The early warning module is used to output an early warning signal if the blood pressure fluctuation factor is greater than a first preset value or the imbalance risk factor is greater than a second preset value.

[0013] According to one aspect of the above technical solution, the attribution module is specifically used for: The deviation between the current physiological state and the target state is quantified using the Frobenius paradigm. Based on the deviation, multi-state transition risk assessment is achieved by normalizing the metabolic state transition matrix according to the Softmax probability.

[0014] According to one aspect of the above technical solution, the system further includes: The update module extracts the feature difference tensors of several consecutive time periods in the current batch, and filters the feature difference tensors belonging to the target state based on the indicator function; The sample weight coefficient of the target state sample is calculated based on the number of target state samples in the historical state and the number of target state samples in the current batch. The state affiliation is determined based on the metabolic state transition matrix, and the cluster center of the target state is updated in combination with the sample weight coefficients. Then, the metabolic state transition matrix for the next time period is calculated based on the updated cluster center of the target state.

[0015] According to one aspect of the above technical solution, the attribution module is specifically used for: Based on the metabolic state transition matrix, the probability of belonging to the target state is calculated. If the probability of belonging is greater than an empirical threshold, the current metabolic state is determined to be the target state.

[0016] Compared with existing technologies, the beneficial effects of this invention are as follows: by establishing a spatiotemporal feature tensor based on data such as blood pressure, serum sodium, and cerebral oxygen saturation, and then quantifying the overall deviation between the current state and the target metabolic state through a metabolic state transition matrix, the state attribution is determined based on Softmax probability normalization, and then the blood pressure fluctuation factor and imbalance risk factor are calculated to determine whether to determine the warning, thus upgrading the dialysis complication warning from "isolated parameter threshold judgment" to "multi-dimensional process quantification", realizing early intervention and precise resource allocation, and improving the monitoring effect. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the hemodialysis data processing method in the first embodiment of the present invention; Figure 2 This is a structural block diagram of the hemodialysis data processing system in the second embodiment of the present invention; Figure 3 This is a schematic diagram of the hardware structure of the computer in the third embodiment of the present invention; The following detailed description, in conjunction with the accompanying drawings, will further illustrate the present invention. Detailed Implementation

[0018] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Several embodiments of the invention are illustrated in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.

[0019] It should be noted that when a component is said to be "fixed to" another component, it can be directly on the other component or there may be an intervening component. When a component is said to be "connected to" another component, it can be directly connected to the other component or there may be an intervening component. The terms "vertical," "horizontal," "left," "right," and similar expressions used in this document are for illustrative purposes only.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0021] Example 1 Please see Figure 1 The figure shows a flowchart of the hemodialysis data processing method in the first embodiment of the present invention. As shown in the figure, the method includes the following steps: Step S100 involves collecting blood pressure waveform data and imbalance syndrome data during dialysis to construct a spatiotemporal feature tensor of data changes. Specifically, systolic / diastolic blood pressure waveform dynamics are captured using a pressure sensor to obtain blood pressure waveform data; the patient's serum sodium concentration is obtained through online monitoring of the dialysis equipment; and near-infrared spectroscopy is used to measure the difference in absorption of different wavelengths of light by oxyhemoglobin and deoxyhemoglobin in brain tissue by using near-infrared light to penetrate the skull, thereby calculating local brain oxygen saturation.

[0022] Step S200: Calculate the metabolic state transition matrix based on the dynamic feature difference of the spatiotemporal feature tensor, and determine the state affiliation based on the metabolic state transition matrix. The metabolic states include a stable state, a hypotension warning state, and a state at risk of imbalance syndrome. Specifically, the aforementioned spatiotemporal feature tensor... Where c is the parameter channel (including the blood pressure waveform data, blood sodium concentration and brain oxygen saturation mentioned above), s is the spatial node, t is the sampling time, and the dynamic feature difference is the tensor difference between two adjacent sampling points. Furthermore, in this embodiment, the calculation expression for the above-mentioned metabolic state transition matrix is ​​as follows: ; In the formula, This is the metabolic state transition matrix. The feature difference tensor corresponding to the target state. Cluster centers for the target state Let k be the Frobenius norm, and k be the metabolic state category index. The cluster center is the k-th metabolic state. Specifically, in some application scenarios of this embodiment, the feature difference tensor corresponding to the target state includes blood pressure acceleration, blood sodium change rate, and brain oxygen fluctuation. Furthermore, in this embodiment, the steps of the mechanism for updating the cluster centers of metabolic states specifically include: Extract the feature difference tensors of several consecutive time periods in the current batch, and filter the feature difference tensors belonging to the target state based on the indicator function; The sample weight coefficient of the target state sample is calculated based on the number of target state samples in the historical state and the number of target state samples in the current batch. The state affiliation is determined based on the metabolic state transition matrix, and the cluster center of the target state is updated in combination with the sample weight coefficients. Then, the metabolic state transition matrix for the next time period is calculated based on the updated cluster center of the target state.

[0023] Furthermore, in this embodiment, the calculation expression for the update mechanism of the cluster centers of the above-mentioned metabolic states is: ; ; In the formula, The cluster centers of the updated target state of class j, The cluster centers of the j-th target state before the update. The sample weight coefficients are... This refers to the set of samples in the current batch that belong to the target state. For the feature difference tensor, This represents the number of samples in the current batch that belong to the target state. This represents the number of target state samples in the historical states. Specifically, the above... Used to evaluate the probability of the j-th type of target state. Used to quantify the overall offset of the target state, if A larger deviation indicates a greater risk of imbalance syndrome.

[0024] Preferably, the step of determining the state affiliation based on the metabolic state transition matrix specifically includes: Based on the metabolic state transition matrix, the probability of belonging to the target state is calculated. If the probability of belonging is greater than an empirical threshold, the current metabolic state is determined to be the target state. The empirical threshold is preferably 0.6.

[0025] Step S300: If the metabolic state transitions to a hypotension warning state, calculate the blood pressure fluctuation factor based on the blood pressure waveform data; if the metabolic state transitions to an imbalance risk state, calculate the imbalance risk based on the imbalance syndrome data. Specifically, in this embodiment, the blood pressure fluctuation factor includes an acute factor and a chronic factor. The first preset value includes a first threshold and a second threshold corresponding to the acute factor and the chronic factor, respectively. The calculation expression for the blood pressure fluctuation factor is as follows: ; ; In the formula, As a blood pressure fluctuation factor, This is the second derivative of the blood pressure waveform signal corresponding to the acute factor. Let H be the energy probability distribution of the blood pressure signal after wavelet decomposition, H be the entropy value of the energy probability distribution corresponding to the chronic factor, and t be the signal sampling time. The calculation expression for the imbalance risk factor is as follows: ; In the formula, As a risk factor for imbalance, This represents the change in serum sodium concentration. For time intervals, This represents the change in brain oxygen saturation. This is the baseline value for brain oxygen saturation. This is the maximum viscosity value. These are the weighting coefficients corresponding to serum sodium concentration and cerebral oxygen saturation, respectively.

[0026] Step S400: If the blood pressure fluctuation factor is greater than a first preset value or the imbalance risk factor is greater than a second preset value, an early warning signal is output. Specifically, the aforementioned blood pressure fluctuation factor includes acute factors and chronic factors, wherein... This indicates that the curvature acceleration of the blood pressure waveform exceeds a first threshold, signifying an acute hypotensive event. This indicates that if the chronic factor exceeds the second threshold, an early warning signal needs to be output. When both are triggered, the acute factor is responded to first. The entropy value H is used to quantify the progressive failure of the autonomic nervous system, and the entropy value threshold is used to achieve the "early intervention window". The rate of change in serum sodium concentration is used to quantify changes in osmotic pressure. The relative fluctuation of brain oxygen saturation is used to quantify brain injury. The preferred value is 0.6. The preferred value is 0.4. The weight value can be optimized according to individual differences. For example, the weight of cerebral oxygen saturation can be increased for patients with a history of stroke.

[0027] Understandably, in some application scenarios of this embodiment, when the target state is unstable, risk warnings are given based on the target state and the monitoring frequency is increased, and blood pressure fluctuation factors or imbalance risk factors are calculated until the patient enters a stable state or an early warning signal is issued.

[0028] In summary, the hemodialysis data processing method in the above embodiments of the present invention establishes a spatiotemporal feature tensor based on data such as blood pressure, serum sodium, and cerebral oxygen saturation. Then, it quantifies the overall deviation between the current state and the target metabolic state through a metabolic state transition matrix, determines the state affiliation based on Softmax probability normalization, and then determines whether to issue an early warning by calculating blood pressure fluctuation factors and imbalance risk factors. This upgrades the early warning of dialysis complications from "isolated parameter threshold judgment" to "multi-dimensional process quantification," enabling early intervention and precise resource allocation, and improving monitoring effectiveness.

[0029] Example 2 A second embodiment of this application also provides a hemodialysis data processing system for implementing the embodiments and preferred embodiments described herein, which will not be repeated hereafter. As used below, the terms "module," "unit," "subunit," etc., can refer to a combination of software and / or hardware that performs a predetermined function. Although the system described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0030] like Figure 2 As shown, the system includes: a data acquisition module 100, a data attribution module 200, a calculation module 300, and an early warning module 400.

[0031] The acquisition module 100 is used to acquire blood pressure waveform data and imbalance syndrome data during dialysis to construct a spatiotemporal feature tensor of data changes; The attribution module 200 is used to calculate the metabolic state transition matrix based on the dynamic feature difference of the spatiotemporal feature tensor, and to determine the state attribution based on the metabolic state transition matrix. The metabolic state includes a stable state, a hypotension warning state, and an imbalance syndrome risk state. The calculation module 300 is used to calculate a blood pressure fluctuation factor based on the blood pressure waveform data if the metabolic state transitions to a hypotension warning state, and to calculate an imbalance risk factor based on the imbalance syndrome data if the metabolic state transitions to an imbalance syndrome state. The early warning module 400 is used to output an early warning signal if the blood pressure fluctuation factor is greater than a first preset value or the imbalance risk factor is greater than a second preset value.

[0032] Preferably, in this embodiment, the attribution module 200 is specifically used for: The deviation between the current physiological state and the target state is quantified using the Frobenius paradigm. Based on the deviation, multi-state transition risk assessment is achieved by normalizing the metabolic state transition matrix according to the Softmax probability.

[0033] Preferably, in this embodiment, the system further includes: The update module extracts the feature difference tensors of several consecutive time periods in the current batch, and filters the feature difference tensors belonging to the target state based on the indicator function; The sample weight coefficient of the target state sample is calculated based on the number of target state samples in the historical state and the number of target state samples in the current batch. The state affiliation is determined based on the metabolic state transition matrix, and the cluster center of the target state is updated in combination with the sample weight coefficients. Then, the metabolic state transition matrix for the next time period is calculated based on the updated cluster center of the target state.

[0034] Preferably, in this embodiment, the attribution module 200 is specifically used for: Based on the metabolic state transition matrix, the probability of belonging to the target state is calculated. If the probability of belonging is greater than an empirical threshold, the current metabolic state is determined to be the target state.

[0035] It should be noted that the modules can be functional modules or program modules, and can be implemented in software or hardware. For modules implemented in hardware, the modules can reside in the same processor; or the modules can be located in different processors in any combination.

[0036] Example 3 A third embodiment of this application provides a computer that may include a processor 81 and a memory 82 storing computer program commands.

[0037] Specifically, the processor 81 may include a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0038] The memory 82 may include a large-capacity storage device for data or commands. For example, and not limitingly, the memory 82 may include a hard disk drive (HDD), a floppy disk drive, a solid-state drive (SSD), flash memory, an optical disk drive, a magneto-optical disk drive, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 82 may include removable or non-removable (or fixed) media. Where appropriate, the memory 82 may be internal or external to a data processing device. In a particular embodiment, the memory 82 is non-volatile memory. In a particular embodiment, the memory 82 includes read-only memory (ROM) and random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), an electrically alterable read-only memory (EAROM), or flash memory, or a combination of two or more of these. Where appropriate, the RAM can be Static Random-Access Memory (SRAM) or Dynamic Random-Access Memory (DRAM). DRAM can be Fast Page Mode Dynamic Random-Access Memory (FPMDRAM), Extended Data Out Dynamic Random-Access Memory (EDODRAM), Synchronous Dynamic Random-Access Memory (SDRAM), etc.

[0039] The memory 82 can be used to store or cache various data files that need to be processed and / or communicated, as well as possible computer program commands executed by the processor 81.

[0040] The processor 81 reads and executes computer program commands stored in the memory 82 to implement any of the hemodialysis data processing methods in the above embodiments.

[0041] In some embodiments, the computer may further include a communication interface 83 and a bus 80. For example, Figure 3 As shown, the processor 81, memory 82, and communication interface 83 are connected through bus 80 and complete communication with each other.

[0042] The communication interface 83 is used to enable communication between the various modules, devices, units, and / or equipment in the embodiments of this application. The communication interface 83 can also enable data communication with other components such as external devices, image / data acquisition devices, databases, external storage, and image / data processing workstations.

[0043] Bus 80 includes hardware, software, or both, that couples computer components together. Bus 80 includes, but is not limited to, at least one of the following: data bus, address bus, control bus, expansion bus, and local bus. For example, and not as a limitation, bus 80 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses, or a combination of two or more of these. Where appropriate, bus 80 may include one or more buses. Although specific buses are described and illustrated in the embodiments of this application, this application considers any suitable bus or interconnection.

[0044] Example 4 The fourth embodiment of this application provides a readable storage medium. This readable storage medium stores computer program commands; when executed by a processor, these computer program commands implement any of the hemodialysis data processing methods described in the above embodiments.

[0045] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the 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.

[0046] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for processing hemodialysis data, characterized in that, Includes the following steps: Blood pressure waveform data and imbalance syndrome data were collected during dialysis to construct a spatiotemporal feature tensor of data changes; The metabolic state transition matrix is ​​calculated based on the dynamic feature difference of the spatiotemporal feature tensor, and the state assignment is determined based on the metabolic state transition matrix. The metabolic state includes a stable state, a hypotension warning state, and an imbalance syndrome risk state. If the metabolic state transitions to a hypotension warning state, a blood pressure variability factor is calculated based on the blood pressure waveform data. If the metabolic state transitions to a disequilibrium syndrome risk state, an imbalance risk factor is calculated based on the disequilibrium syndrome data. If the blood pressure fluctuation factor is greater than the first preset value or the imbalance risk factor is greater than the second preset value, an early warning signal is output.

2. The hemodialysis data processing method according to claim 1, characterized in that, The blood pressure fluctuation factor includes an acute factor and a chronic factor. The first preset value includes a first threshold and a second threshold corresponding to the acute factor and the chronic factor, respectively. The calculation expression for the blood pressure fluctuation factor is as follows: ； ； In the formula, As a blood pressure fluctuation factor, This is the second derivative of the blood pressure waveform signal corresponding to the acute factor. Let H be the energy probability distribution of the blood pressure signal after wavelet decomposition, H be the entropy value of the energy probability distribution corresponding to the chronic factor, and t be the signal sampling time. The calculation expression for the imbalance risk factor is as follows: ； In the formula, As a risk factor for imbalance, This represents the change in serum sodium concentration. For time intervals, This represents the change in brain oxygen saturation. This is the baseline value for brain oxygen saturation. This is the maximum viscosity value. These are the weighting coefficients corresponding to serum sodium concentration and cerebral oxygen saturation, respectively.

3. The hemodialysis data processing method according to claim 1, characterized in that, The specific steps for calculating the metabolic state transition matrix based on dynamic feature difference of spatiotemporal feature tensors include: The deviation between the current physiological state and the target state is quantified using the Frobenius paradigm. Based on the deviation, multi-state transition risk assessment is achieved by normalizing the metabolic state transition matrix according to the Softmax probability.

4. The hemodialysis data processing method according to claim 3, characterized in that, The expression for calculating the metabolic state transition matrix is ​​as follows: ； In the formula, This is the metabolic state transition matrix. The feature difference tensor corresponding to the target state. Cluster centers for the target state Let k be the Frobenius norm, and k be the metabolic state category index. It serves as the cluster center for the k-th metabolic state.

5. The hemodialysis data processing method according to claim 4, characterized in that, The specific steps of the update mechanism for the cluster centers of metabolic states include: Extract the feature difference tensors of several consecutive time periods in the current batch, and filter the feature difference tensors belonging to the target state based on the indicator function; The sample weight coefficient of the target state sample is calculated based on the number of target state samples in the historical state and the number of target state samples in the current batch. The state affiliation is determined based on the metabolic state transition matrix, and the cluster center of the target state is updated in combination with the sample weight coefficients. Then, the metabolic state transition matrix for the next time period is calculated based on the updated cluster center of the target state.

6. The hemodialysis data processing method according to claim 5, characterized in that, The calculation expression for the update mechanism of cluster centers of metabolic states is as follows: ； ； In the formula, The cluster centers of the updated target state of class j, The cluster centers of the j-th target state before the update. The sample weight coefficients, This refers to the set of samples in the current batch that belong to the target state. For the feature difference tensor, This represents the number of samples in the current batch that belong to the target state. This represents the number of target state samples in the historical states.

7. The hemodialysis data processing method according to claim 6, characterized in that, The steps for determining state affiliation based on the metabolic state transition matrix specifically include: Based on the metabolic state transition matrix, the probability of belonging to the target state is calculated. If the probability of belonging is greater than an empirical threshold, the current metabolic state is determined to be the target state.

8. A hemodialysis data processing system, characterized in that, include: The data acquisition module is used to collect blood pressure waveform data and imbalance syndrome data during dialysis in order to construct a spatiotemporal feature tensor of data changes; The attribution module is used to calculate the metabolic state transition matrix based on the dynamic feature difference of the spatiotemporal feature tensor, and to determine the state attribution based on the metabolic state transition matrix. The metabolic state includes a stable state, a hypotension warning state, and an imbalance syndrome risk state. The calculation module is used to calculate a blood pressure fluctuation factor based on the blood pressure waveform data if the metabolic state transitions to a hypotension warning state, and to calculate an imbalance risk factor based on the imbalance syndrome data if the metabolic state transitions to an imbalance syndrome state. The early warning module is used to output an early warning signal if the blood pressure fluctuation factor is greater than a first preset value or the imbalance risk factor is greater than a second preset value.

9. A computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the hemodialysis data processing method as described in any one of claims 1-7.

10. A storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the hemodialysis data processing method as described in any one of claims 1-7 above.

Images