Acupuncture point state discrimination method and device based on midnight-noon ebb-flow time vector, and medium

Abstract

The invention discloses an acupuncture point state judgment method and device based on a midnight-noon ebb-flow time vector and a medium. The method comprises the steps that to-be-judged resistance information of a target acupuncture point skin area is collected; inputting the resistance information to be discriminated into an acupuncture point state discrimination model, and outputting an acupuncture point state; the acupoint state discrimination model is obtained by training an XGBoost model through a training sample set, and the training sample set comprises multiple pieces of sample data; each piece of sample data comprises a resistance value of a sample acupuncture point skin area within a set time period for continuous preset days, acquisition time and a label of each sample acupuncture point skin area; the label is in an acupoint state. According to the method, the resistance characteristic and the time vector analysis technology are effectively combined, and the accuracy of acupoint state judgment is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of machine learning, and particularly to a method, device, and medium for discriminating acupoint states based on the time vector of the meridian and collateral flow infusion. Background Art

[0002] With the continuous integration of traditional Chinese medicine theory and modern technology, the analysis of the electrical properties of meridians and acupoints has become an important means to study the health status of the human body. Acupoints are important concepts in traditional Chinese medicine theory. Traditionally, it is believed that they are closely related to the functions of zang-fu organs and are important treatment sites for therapies such as acupuncture and massage. Modern research has found that there is a certain correlation between the resistance characteristics of acupoints and the physiological and pathological states of the human body. For example, when an acupoint is in an active state, that is, the acupoint-opening state, the resistance value is usually low, while when the acupoint is in a static state, that is, the non-acupoint-opening state, the resistance value is high.

[0003] However, the existing judgment generally only relies on static resistance measurement, and it is difficult to comprehensively consider the dynamic change law of the resistance value over time. It is impossible to make full use of the time law in the traditional Chinese medicine meridian and collateral flow infusion theory, and the discrimination accuracy of whether an acupoint is in the "acupoint-opening" state is relatively low. At the same time, external factors such as individual differences, environmental interference, and equipment performance further affect the accuracy and consistency of the measurement results. Summary of the Invention

[0004] To solve the above technical problems, the present invention provides a method, device, and medium for discriminating acupoint states based on the time vector of the meridian and collateral flow infusion, effectively combining resistance characteristics and time vector analysis technology to improve the accuracy of acupoint state discrimination.

[0005] To solve the above technical problems, the first aspect of the present invention discloses a method for discriminating acupoint states based on the time vector of the meridian and collateral flow infusion, including:

[0006] Collecting the resistance information to be discriminated in the skin area of the target acupoint;

[0007] Inputting the resistance information to be discriminated into the acupoint state discrimination model to output the acupoint state; the acupoint state discrimination model is obtained by training the XGBoost model with a training sample set, and the training sample set includes multiple sample data; each sample data includes the resistance value, acquisition time, and label of each sample acupoint skin area within a continuous preset number of days in a set time period in the skin area of a sample acupoint; the label is the acupoint state.

[0008] In some embodiments, the resistance information to be discriminated in the skin area of the target acupoint includes the continuous preset number of days in the skin area of the target acupoint and the resistance value to be discriminated at the same time period.

[0009] In some embodiments, the acupoint state discrimination model is obtained by training the XGBoost model with a training sample set, including:

[0010] Collect a number of sample data; each said sample data includes the resistance information of the sample skin area and the collection time, and the resistance information of the sample skin area is the resistance value within the sample skin area for a continuous preset number of days and at the same time period;

[0011] Mark the resistance information with time and acupoint status to obtain the label of each sample data;

[0012] Conduct maximum differential data screening on the sample data to generate a training sample set containing a specified number of sample data;

[0013] Iteratively train the XGBoost model through the training sample set until the training termination condition is met to obtain an acupoint status discrimination model.

[0014] In some embodiments, conducting maximum differential data screening on the sample data includes:

[0015] Calculate the standard deviation, residual error, and average value of the sample data, and eliminate highly abnormal values in the sample data according to the Pauta criterion;

[0016] Eliminate the highly abnormal values, and use the average value of adjacent sample data to replace the highly abnormal values;

[0017] Arrange the sample data in ascending order, calculate the average value, standard deviation, maximum deviation value, and minimum deviation value of the sample data, and determine suspicious values according to the maximum deviation value and the minimum deviation value; calculate the Grubbs statistic for the suspicious values;

[0018] Determine the critical value of the Grubbs table corresponding to the sample data volume, compare the critical value with the Grubbs statistic, and determine and extract the maximum difference value.

[0019] In some embodiments, after conducting maximum differential data screening on the sample data, it further includes:

[0020] Normalize the sample data that has completed maximum differential data screening;

[0021] Eliminate sample data with a data length less than the preset length threshold, and truncate sample data with a data length greater than the preset length threshold, so that the data length of the sample data is all the preset length threshold.

[0022] In some embodiments, the acupoint status discrimination model includes an input layer, a task layer, and an output layer;

[0023] The input layer is used to receive the resistance information to be discriminated, preprocess the resistance information to be discriminated, and convert the resistance information to be discriminated into an input vector;

[0024] The task layer includes a set of decision trees. Each decision tree takes the input vector as input and the acupoint state as output; the acupoint states output by all decision trees are weighted and calculated to obtain a predicted value of the acupoint state;

[0025] The output layer is used to receive the predicted value of the acupoint state from the task layer and generate the acupoint state.

[0026] In some embodiments, the resistance information to be discriminated in the target acupoint skin area includes the resistance values at at least 40 time points in the same time period from the Jia day to the Kui day in the target acupoint skin area.

[0027] In some embodiments, the resistance information to be discriminated is collected by an FPC flexible sensor.

[0028] According to the second aspect of the present invention, a computer device is disclosed, including: a processor and a memory; wherein, the memory stores a computer program, and the computer program is adapted to be loaded and executed by the processor to perform the steps of the acupoint state discrimination method based on the time vector of the meridian and collateral flow as described above.

[0029] According to the third aspect of the present invention, a computer storage medium is disclosed, on which a computer program is stored. When the computer program is executed by a processor, the steps of the acupoint state discrimination method based on the time vector of the meridian and collateral flow as described above are implemented.

[0030] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0031] The present invention provides an acupoint state discrimination method, device, and medium based on the time vector of the meridian and collateral flow, which can directly and accurately judge the acupoint state through the resistance information to be discriminated in the target skin area, effectively combine the resistance characteristics and the time vector analysis technology, improve the accuracy of acupoint state discrimination, and further provide a reliable basis for judging the human health status. Description of the Drawings

[0032] Figure 1 It is a schematic flow chart of the acupoint state discrimination method based on the time vector of the meridian and collateral flow provided by the present invention;

[0033] Figure 2 It is a schematic flow chart of step S2 of the acupoint state discrimination method based on the time vector of the meridian and collateral flow provided by the present invention;

[0034] Figure 3 It is a schematic flow chart of step S23 of the acupoint state discrimination method based on the time vector of the meridian and collateral flow provided by the present invention. Detailed implementation mode

[0035] For better understanding and implementation, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0036] The terms "include" and "have" in the embodiments of the present invention and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or modules does not necessarily limit to those clearly listed steps or modules, but may include other steps or modules not clearly listed or inherent to these processes, methods, products or devices.

[0037] The embodiments of the present invention disclose a method, device, and medium for discriminating acupoint states based on the time vector of the meridian and collateral flow injection.

[0038] As Figure 1 shown, this method includes the following steps:

[0039] Step S1: Collect the resistance information to be discriminated in the skin area of the target acupoint.

[0040] The resistance information to be discriminated includes the resistance value to be discriminated and the acquisition time. Generally, the resistance value to be discriminated of the skin area of the user's target acupoint is collected through a sensor, and the sensor is attached near the skin area of the target acupoint. Before collection, to improve the accuracy of collection, the sensor needs to be calibrated, and it is necessary to ensure that the collection environment is constant in temperature and humidity to reduce the influence of environmental interference on the resistance value. During the collection process, according to traditional Chinese medicine theory, the specific acupoint position is determined. In a fixed time period within 10 consecutive days, for example, from 13:00 to 15:00 every day, the sensor automatically records the resistance value to be discriminated of the target acupoint, and stores the data collected every minute to form time series data of the resistance value to be discriminated.

[0041] In this application, the flexible FPC (Flexible Printed Circuit) flexible FPC sensor can be well attached to the skin area of the target acupoint, ensuring good adhesion between the sensor and the skin, with a thin and light material, avoiding measurement errors caused by poor contact, facilitating the measurement of acupoints throughout the body, and avoiding the problems of overly troublesome data collection and complex derivation of data analysis relying on professionals. Specific acupoint electrodes can select different types of electrodes, such as gel electrodes or flexible electrodes, to adapt to different measurement requirements and cost considerations.

[0042] The resistance information to be discriminated refers to the sequence of resistance values and their corresponding time information collected from the skin area of the target acupoint for a continuous preset number of days within the same time period, reflecting the dynamic changes in the resistance characteristics of the target acupoint. Its data range is usually between several hundred ohms and several thousand ohms, depending on whether the acupoint is in the open-acupoint state or the non-open-acupoint state. When in the open-acupoint state, the resistance value is usually low, while in the "non-open-acupoint" state, the resistance value is high. In addition, the resistance information to be discriminated also includes the acquisition timestamp and the acupoint position identifier, which are used for subsequent data analysis and processing. According to the Na Jia method of the theory of the flow of qi along the meridians, the preset number of days can be set to two rounds of 10 days each. From Jia day to Gui day, at 13:00 - 15:00 every day (the hour of Wei), the acupoint data of the volunteers are collected. Continuously collecting for 10 days is one set of two rounds of data, and a total of 6 sets are collected.

[0043] In this application, the resistance information to be discriminated is collected at at least 40 time points within the same time period. By analyzing the resistance information to be discriminated at multiple time points, the fluctuation trend and its stability of the resistance value can be identified, thereby determining whether the target acupoint is in the open-acupoint state. The open-acupoint state usually shows a low resistance value and strong change regularity, while the non-open-acupoint state shows a high resistance value or large fluctuations. This refined data collection can improve the discrimination accuracy of the model and help accurately identify the physiological active state of the acupoint.

[0044] In this application, 50 time points are continuously collected within 2 hours. The sampling interval for each time point is about 2.4 minutes, reasonably balancing the data density and the acquisition efficiency of the resistance data, avoiding missing the detailed changes in the resistance due to too long a sampling interval or generating a large amount of redundant data and increasing the calculation difficulty due to too short a sampling interval. It can not only ensure the capture accuracy of the resistance value change law, but also optimize the efficiency of model training and application, while maintaining the simplicity of the acquisition process and the controllability of data processing, improving the reliability and consistency of the data, providing a sufficient basis for analyzing the correlation between acupoint resistance and the active time of the meridians.

[0045] Step S2: Input the resistance information to be discriminated into the acupoint state discrimination model to output the acupoint state; the acupoint state discrimination model is obtained by training the XGBoost model with a training sample set, and the training sample set includes multiple sample data; each sample data includes the resistance value, acquisition time, and label of each sample acupoint skin area within a continuous preset number of days in the set time period; the label is the acupoint state.

[0046] After obtaining the resistance information to be discriminated, it is input into the acupoint state discrimination model. The acupoint state discrimination model is obtained by training an XGBoost model with a training sample set, and the training sample set includes multiple sample data; each sample data includes the resistance values within a continuous preset number of days in a set time period in a sample acupoint skin area, the acquisition time, and the label of each sample acupoint skin area; the label is the acupoint state.

[0047] The acupoint state discrimination model is obtained by training an XGBoost model with a training sample set, as Figure 2 shown. The specific training steps include:

[0048] Step S21: Collect a number of sample data; each of the sample data includes the resistance information of the sample skin area and the acquisition time, and the resistance information of the sample skin area is the resistance values within a continuous preset number of days and the same time period.

[0049] Specifically, the sample data is to collect the resistance information to be discriminated of the sample acupoint skin area in the way of step S1, the resistance information values of the sample acupoint skin area in the same time period within a continuous preset number of days, the acquisition time, and the area of the sample acupoint skin, providing time series features for subsequent model analysis. The acquisition of the resistance values is carried out in a fixed time period, such as from 13:00 to 15:00 every day, and a resistance data sequence with a time pattern is formed through continuous sampling.

[0050] Step S22: Label the resistance information with respect to time and acupoint state to obtain the label of each sample data.

[0051] After the collection is completed, the sample data needs to be labeled with time and acupoint state. By combining the theory of traditional Chinese medicine's midnight-noon ebb-flow and the change law of resistance values, the corresponding acupoint state labels are added to the resistance information at each time point. In some embodiments, the labeling content includes the acupoint state of the sample acupoint corresponding to the resistance value, that is, being in the acupoint-opening state or the non-acupoint-opening state, and the label will be used as the output target of the training sample to guide the model to learn the association between the resistance value and the acupoint state.

[0052] Step S23: Perform maximum-differentiation data screening on the sample data to generate a training sample set containing a specified number of sample data.

[0053] To improve the accuracy of the model, maximum-differentiation data screening is performed on the sample data. In the screening process, the abnormal values caused by equipment errors or environmental interference are removed to ensure the reliability of the sample data. At the same time, the data length is standardized so that the time series lengths of all sample data are consistent, optimizing the quality of the sample data and enabling the model to better capture the dynamic characteristics and time patterns of the resistance values.

[0054] Specifically, as Figure 3 shown, screening the maximum differential data from the sample data includes:

[0055] Step S231: Calculate the standard deviation, residual error, and average value of the sample data, and eliminate the highly abnormal values in the sample data according to the Pauta criterion;

[0056] Arrange all the sample data in ascending order to form a sequence Xi (i = 1, 2, 3, n), and calculate the average value of the sequence Xi and the residual error Calculate the standard deviation S of the sequence according to Bessel's formula:

[0057]

[0058] where Xd is the resistance value of a certain sample data.

[0059] If the residual error of the resistance value Xd of a certain sample data and satisfies |ui| > 3S, then Xd is considered a highly abnormal value and needs to be eliminated. This discriminant is the Pauta criterion.

[0060] Step S232: Eliminate the highly abnormal values and use the average value of adjacent sample data to replace the highly abnormal values;

[0061] For the positions occupied by the original highly abnormal values to be eliminated, there should be new values to replace them. The replacement value needs to satisfy the following formula X d =(X d-1 +X d+1 ) / 2.

[0062] Step S233: Arrange the sample data in ascending order, calculate the average value, standard deviation, maximum deviation value, and minimum deviation value of the sample data, and determine the suspicious values according to the maximum deviation value and the minimum deviation value; Calculate the Grubbs statistic for the suspicious values.

[0063] Arrange all the sample data in ascending order. Try to use a sorting algorithm with a low time complexity and ensure stability. Among the existing sorting methods, radix sorting is relatively stable and has a time complexity of O(n). Calculate the average value, standard deviation, maximum deviation value, and minimum deviation value of the sample data, where the maximum deviation value is

[0064] The minimum deviation value is

[0065] Compare the maximum deviation value with the minimum deviation value. If the maximum deviation value α1 is greater than the minimum deviation value α2, that is, α2 < α1， Then the maximum value X in the sample data can be considered max as a suspicious value. Conversely, the minimum value X in the sample data can be considered min as a suspicious value.

[0066] Calculate the Grubbs statistic through the following formula, that is, the G i value:

[0067]

[0068] where i is the arrangement serial number of the suspicious value.

[0069] Step S234: Determine the critical value of the Grubbs table corresponding to the sample data volume, compare the critical value with the Grubbs statistic, and determine and extract the maximum difference value.

[0070] Compare the Gi value calculated in step S233 with the critical value G given in the Grubbs table pn If the calculated G i is greater than the critical value G in the table pn , it can be judged that this value is an outlier, and at the same time, the maximum difference value is extracted, and the extraction of the maximum difference value is completed.

[0071] The sample data after screening is constructed into a training sample set, which contains a specified number of high-quality samples. Each sample in the training sample set contains the normalized resistance value sequence and time points of the input features) and the corresponding status label ("acupoint opening" or "non-acupoint opening").

[0072] Furthermore, since the range of acupoint resistance values is relatively large, after completing the screening of the maximum differential data, it also includes:

[0073] Step S235: Normalize the sample data after completing the screening of the maximum differential data;

[0074] Eliminate the sample data with a data length less than the preset length threshold, and truncate the sample data with a data length greater than the preset length threshold, so that the data length of the sample data is all the preset length threshold.

[0075] Through normalization, the resistance value is mapped between [0-1], and the processing of the consistency of the sample data length ensures the accuracy and consistency of the model training data, and improves the generalization ability of the model. Transform the dimensional dataset into a scalar, simplify the calculation, improve the data performance, and make the sample data comparable, and further generate a training sample set containing a specified number of sample data.

[0076] Step S24: Iteratively train the XGBoost model through the training sample set until the training termination condition is met, and obtain an acupoint state discrimination model.

[0077] The training sample set is input into the XGBoost model to start training. An acupoint state discrimination model is constructed using the XGBoost machine learning model. XGBoost is an ensemble learning algorithm based on the nesting of multiple algorithms. By repeatedly training multiple weak classifiers and combining the classification results with weights, a powerful classifier for acupoint state discrimination is constructed, improving the accuracy and stability of the acupoint state discrimination model.

[0078] During the training process, the XGBoost model optimizes the objective function through multiple iterations. In each iteration, the XGBoost model calculates the gradient of the loss function and updates the splitting nodes and weights of the decision tree according to the gradient. The reconstructed decision tree fits the current error and adds the output result to the overall predicted value of the XGBoost model with weights. In this way, the model gradually reduces the error between the predicted value and the true value, continuously improving the classification accuracy. Regularization processing is also introduced to control the complexity of the model and avoid overfitting.

[0079] The training process stops when the termination condition is met. For example, when the loss function converges or reaches the preset maximum number of iterations, the model stops optimizing and the acupoint state discrimination model is obtained. The finally obtained XGBoost model can accurately discriminate the state of the target acupoint according to the input resistance information and time characteristics.

[0080] Specifically, the data of the training sample set is input into the XGBoost model. Taking the eigenvalue Rx as the classification point, iterative classification training is performed on m = 1, 2,..., M decision trees to obtain the classification result Y of predicting acupoint opening / closure. m The difference between the predicted result of the model and the label is recorded as the residual. Among them, the eigenvalue is the normalized resistance value, and M is the total number of decision trees. For each sample data Xi, i = 1, 2,..., n, the negative gradient, that is, the residual, is calculated:

[0081]

[0082] Taking the residual R mi as the target value, and then taking the sample data (xi, Rmi), i = 1, 2,..., n as the training data of the next decision tree. Among them, n is the total number of sample data, y i is the label of the i-th sample data, f m-1 (xi) is the predicted result of the (m - 1)-th tree, and L(y i , f(xi)) is the loss function. When the training iterates to the k-th decision tree, the loss function reaches the relative optimal solution and the iterative training stops. The final predicted result of the model is the sum of the predicted results of the k decision trees.

[0083] The training sample set is divided into a training set and a test set according to a certain ratio. In some embodiments, the training set accounts for 80% and the test set accounts for 20%. The data of the training sample set is introduced into the XGBoost model, and the Xgboost model is iteratively trained to continuously improve and correct the model, obtaining an acupoint discrimination model. The data of the test set is used to test the acupoint discrimination model, and the discrimination accuracy rate of the acupoint discrimination model is obtained. The acupoint discrimination model determines whether a randomly selected acupoint on the human body is in an open state according to the resistance value.

[0084] The acupoint discrimination model includes an input layer, a task layer, and an output layer; the input layer is used to receive the resistance information to be discriminated, preprocess the resistance information to be discriminated, and convert the resistance information to be discriminated into an input vector; the task layer includes a set of decision trees, each decision tree takes the input vector as input and the acupoint state as output; the acupoint states output by all decision trees are weighted and calculated to obtain a predicted acupoint state value; the output layer is used to receive the predicted acupoint state value of the task layer and generate an acupoint state.

[0085] The input layer performs steps such as maximum differential data screening, normalization, and time series extraction in step S23, converting the original resistance information to be discriminated into a standardized input vector that the model can process, and each dimension of the input vector corresponds to the resistance value at a time point. The input layer passes the input vector to the set of decision trees in the task layer.

[0086] The task layer is the core part of the model, which is composed of a set of multiple decision trees. These decision trees work in parallel or serial ways, learning the mapping relationship between the input vector and the acupoint state during the training process, and learning the correlation between the resistance value at the time point and the acupoint state. Each decision tree in the task layer extracts key features from the input vector, splits nodes layer by layer, and divides the input data into finer-grained regions. For example, when the resistance value at a certain time point is lower than a specific threshold, the decision tree may determine it as "open acupoint". The input vector is passed to the task layer, and each decision tree takes the input vector as input. Through multiple iterations, each decision tree gradually corrects the error of the previous round and improves the prediction accuracy of the model. The task layer performs weighted calculation on the output results of all decision trees to obtain a predicted value of the acupoint state.

[0087] The output layer receives the predicted acupoint state value calculated by the task layer and determines the acupoint state according to the set threshold. For example, when the predicted value is greater than 0.5, the output acupoint state is open acupoint, otherwise the output acupoint state is non-open acupoint.

[0088] The present invention uses the acupoint state discrimination model to identify the specific acupoint state based on the resistance information of the target acupoint skin area to achieve the goal of preventing disease. Furthermore, the sensor's electroacupuncture function can be used to perform electroacupuncture treatment on the acupoint based on the acupoint state, avoiding side effects such as skin breakage, needle dizziness, infection, and skin pain. This allows patients to perform the treatment at home, regardless of the treatment location. The treatment can significantly reduce local pain, promote blood circulation and lymphatic return in local tissues, induce skeletal muscle contraction, prevent muscle atrophy, increase smooth muscle tone, act on ganglia and nerve segments, and regulate autonomic nervous system function. The combination of these two methods has a significant therapeutic effect.

[0089] The acupoint status discrimination model judges the acupoint status based on the resistance information to be judged at 50 time points within a fixed time period of a preset number of consecutive days. It has wide applicability and fills the gap in related fields. It is not only suitable for traditional acupoint intelligent positioning, but also can be used for electro-sensitized acupoint intelligent positioning, avoiding the problem of no fixed acupoint resistance value division standard. The basic method of self-comparison avoids the influence of individual differences and skin resistance differences caused by factors such as equipment movement and environmental changes, and can achieve more accurate real-time acupoint discrimination.

[0090] The wearable flexible acupoint sensor used in the present invention can be attached to the surface of human skin for a long time without feeling, which simplifies the collection operation process and has the characteristics of portability, comfort and long-term collection. The acupoint electrodes in the patent can use different types of electrodes, such as gel electrodes or flexible electrodes, to adapt to different measurement needs and cost considerations. By introducing the relationship between acupoints and time as an observation indicator, exploring the theory of meridians in clinical trials has important guiding significance, which helps to promote the modernization of traditional Chinese medicine. By verifying and applying traditional Chinese medicine theories through scientific methods, it enhances the influence and recognition of traditional Chinese medicine in the field of modern medicine. The present invention not only improves the standardization and accuracy of acupoint resistance measurement, but also provides new tools and methods for the clinical application and scientific research of traditional Chinese medicine.

[0091] The present invention also provides a computer device, which may include: a memory storing executable program code;

[0092] a processor coupled to a memory;

[0093] A transceiver used to communicate with other devices or communication networks, and to receive or send network messages;

[0094] A bus used to connect memory, processors, and transceivers for internal communication.

[0095] The transceiver receives the messages transmitted over the network, passes them to the processor via the bus. The processor calls the executable program code stored in the memory via the bus for processing, and passes the processing results to the transceiver for transmission via the bus, thereby implementing the method provided by the embodiments of the present application.

[0096] The embodiments of the present application also provide a non-transitory machine-readable storage medium, on which an executable program is stored. When the executable program is run by a processor, the processor is caused to execute the method provided by the above embodiments.

[0097] The embodiments of the present invention disclose a computer-readable storage medium that stores a computer program for electronic data exchange, wherein the computer program causes a computer to execute the described method.

[0098] The embodiments of the present invention disclose a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute the described method.

[0099] The above embodiments can be referred to the description of the method and will not be elaborated here.

[0100] The embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separated. The components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed to multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of the present embodiment. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0101] Through the specific descriptions of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the above technical solutions, in essence, or the parts that contribute to the prior art can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, and the storage medium includes read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc memories, magnetic disk memories, tape memories, or any other computer-readable medium that can be used to carry or store data.

[0102] Finally, it should be noted that: The disclosed embodiments of the present invention are only the preferred embodiments of the present invention, and are only used to illustrate the technical solutions of the present invention, rather than limiting them; Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: They can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; And these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for discriminating acupoint states based on the time vector of midnight-noon ebb-flow, characterized in that, Including: Collecting the resistance information to be discriminated in the skin area of the target acupoint; Inputting the resistance information to be discriminated into the acupoint state discrimination model and outputting the acupoint state; The acupoint state discrimination model is obtained by training the XGBoost model with a training sample set. The training sample set includes multiple sample data. Each sample data includes the resistance values, collection time, and label of each sample acupoint skin area within a continuous preset number of days in a set time period; The label is the acupoint state.

2. The acupoint state discrimination method based on the time vector of midnight-noon ebb-flow according to claim 1, wherein The resistance information to be discriminated in the skin area of the target acupoint includes the continuous preset number of days in the skin area of the target acupoint and the resistance value to be discriminated at the same time period.

3. The acupoint state discrimination method based on the time vector of the meridian and collateral flow as claimed in claim 1, wherein The acupoint state discrimination model is obtained by training the XGBoost model with a training sample set, including: Collecting a number of sample data. Each sample data includes the resistance information of the sample skin area and the collection time. The resistance information of the sample skin area is the resistance value within a continuous preset number of days and the same time period in the sample skin area; Labeling the resistance information with time and acupoint state to obtain the label of each sample data; Performing maximum differential data screening on the sample data to generate a training sample set containing a specified number of sample data; Performing iterative training on the XGBoost model through the training sample set until the training termination condition is met to obtain the acupoint state discrimination model.

4. The acupoint state discrimination method based on the time vector of the meridian and collateral flow as claimed in claim 3, wherein Performing maximum differential data screening on the sample data, including: Calculating the standard deviation, residual error, and average value of the sample data, and removing the highly abnormal values in the sample data according to the Pauta criterion; Removing the highly abnormal values and replacing the highly abnormal values with the average values of adjacent sample data; Sorting the sample data in ascending order, calculating the average value, standard deviation, maximum deviation value, and minimum deviation value of the sample data, and determining the suspicious values according to the maximum deviation value and the minimum deviation value; calculating the Grubbs statistic for the suspicious values; Determining the critical value of the Grubbs table corresponding to the sample data volume, comparing the critical value with the Grubbs statistic, and determining and extracting the maximum difference value.

5. The acupoint state discrimination method based on the time vector of the meridian and collateral flow as claimed in claim 4, wherein, After performing maximum differential data screening on the sample data, it further includes: Normalizing the sample data after completing the maximum differential data screening; Removing the sample data with a data length less than the preset length threshold and truncating the sample data with a data length greater than the preset length threshold so that the data lengths of the sample data are all the preset length thresholds.

6. The acupoint state discrimination method based on the time vector of the meridian flow injection according to claim 5, characterized in that, The acupoint state discrimination model includes an input layer, a task layer, and an output layer; The input layer is used to receive the resistance information to be discriminated, preprocess the resistance information to be discriminated, and convert the resistance information to be discriminated into an input vector; The task layer includes a set of decision trees. Each decision tree takes the input vector as input and the acupoint state as output; weighted calculation of the acupoint states output by all decision trees to obtain the acupoint state prediction value; The output layer is used to receive the acupoint state prediction value of the task layer and generate the acupoint state.

7. The acupoint state discrimination method based on the time vector of the meridian and collateral flow as claimed in claim 6, wherein, The resistance information to be discriminated for the target acupoint skin area includes the resistance values at at least 40 time points in the same time period from the Jia day to the Kui day for the target acupoint skin area.

8. The acupoint state discrimination method based on the time vector of midnight-noon ebb-flow according to claim 1, wherein The resistance information to be discriminated is collected by an FPC flexible sensor.

9. A computer device, characterized in that, It includes: A processor and a memory; wherein, the memory stores a computer program, and the computer program is adapted to be loaded and executed by the processor to perform the steps of the acupoint state discrimination method based on the time vector of the meridian flow as claimed in any one of claims 1-8.

10. A computer storage medium, characterized in that, A computer program is stored thereon, and when the computer program is executed by the processor, the steps of the acupoint state discrimination method based on the time vector of the meridian flow as claimed in any one of claims 1-8 are implemented.

Images