A method for processing health sign monitoring data of neurological patients

Abstract

The present invention relates to the field of medical data processing, and particularly to a method for processing health sign monitoring data of neurological patients, including: obtaining multiple sign data sets of neurological patients; dividing each sign data set into multiple sleep stage data sets; dividing each sleep stage data set into multiple sleep intervals; obtaining suspicious data of each sign data set according to the interval lengths of the respective sleep intervals; obtaining the suspicious intervals of each sign data set; obtaining the first abnormal weight of each suspicious data according to the volatility of each suspicious data and the length of the suspicious interval to which it belongs; obtaining respective comparison data of each suspicious data, and obtaining the second abnormal weight of each suspicious data according to the comparison data, thereby obtaining the abnormal degree of each suspicious data; obtaining abnormal data according to the abnormal degree of each suspicious data, thereby realizing the targeted identification of abnormal signs of neurological patients. The present invention is more efficient, convenient, and more targeted.

Description

Technical Field

[0001] The present invention relates to the field of medical data processing, and particularly to a method for processing health sign monitoring data of neurological patients. Background Art

[0002] Sleep is a basic and necessary physiological phenomenon of the body, which has a huge impact on human health, specifically manifested in the relief of the body and the relaxation of mental stress. And some special groups are more in need of high-quality sleep. For example, neurological patients, the quality of their sleep has a certain degree of influence on the improvement or deterioration of their condition. Therefore, how to improve the sleep quality of neurological patients is a very important issue.

[0003] The existing methods for improving the sleep quality of neurological patients mainly include drug treatment or non-drug treatment according to the suggestions of attending physicians. The sleep quality is determined by the abnormal signs of the body during sleep, and the abnormal signs are obtained from the data of various signs of neurological patients during sleep. The existing judgment of abnormal signs through various data during sleep is mainly through the analysis of the sign data of neurological patients during sleep by attending physicians. This method is extremely inconvenient for both patients and doctors, with slow efficiency, and it is inevitable to have errors when analyzing too much data. This method can help to a certain extent in improving the sleep quality of neurological patients, but it does not specifically solve the reason for their low sleep quality, so the improvement effect is not very significant. Summary of the Invention

[0004] The present invention provides a method for processing health sign monitoring data of neurological patients to solve the existing problems.

[0005] The method for processing health sign monitoring data of neurological patients of the present invention adopts the following technical solutions:

[0006] An embodiment of the present invention provides a method for processing health sign monitoring data of neurological patients, the method comprising the following steps:

[0007] Obtain different sign data at each moment during the sleep of neurological patients to obtain a plurality of sign data sets;

[0008] Cluster according to the magnitudes of the data values in each sign data set to obtain a plurality of sleep stage data sets for each sign data set; obtain a plurality of sleep intervals for each sleep stage data set according to the continuity of the sampling times of the data in each sleep stage data set; obtain suspicious data for each sign data set according to the interval lengths of the respective sleep intervals;

[0009] Obtain the suspicious interval of each vital sign dataset according to the continuity of the sampling times of all suspicious data in each vital sign dataset; obtain the volatility of each suspicious data based on all suspicious data in the sleep interval to which each suspicious data belongs; obtain the first abnormal weight of each suspicious data according to the volatility of each suspicious data in each vital sign dataset and the length of the suspicious interval to which each suspicious data belongs; obtain the comparison data of each suspicious data in other vital sign datasets according to the sampling times of each suspicious data in each vital sign dataset, and obtain the second abnormal weight of each suspicious data according to the volatility of the sleep interval to which the comparison data belongs; obtain the degree of abnormality of each suspicious data according to the first abnormal weight and the second abnormal weight of each suspicious data;

[0010] Obtain abnormal data according to the degree of abnormality of each suspicious data; identify abnormal signs of neurological patients based on the abnormal data.

[0011] Preferably, the method for obtaining multiple sleep intervals of each sleep stage dataset is as follows:

[0012] Sort each vital sign data in each sleep stage dataset in ascending order of sampling time, and divide the vital sign data with a difference of 1 between sampling times into one interval to obtain multiple sleep intervals in each sleep stage dataset.

[0013] Preferably, the method for obtaining suspicious data of each vital sign dataset is as follows:

[0014] Obtain the maximum interval length of each sleep interval in each sleep stage dataset, and use the product of a preset adjustment parameter and the maximum interval length as the reference length; obtain all sleep intervals in each sleep stage dataset with an interval length less than the reference length, and use each data in all the sleep intervals as the suspicious data in each sleep stage dataset. The suspicious data in all sleep stage datasets constitute the suspicious data of each vital sign dataset.

[0015] Preferably, the method for obtaining the volatility of each suspicious data in each vital sign dataset is as follows:

[0016] According to the occurrence frequency of each suspicious data in the sleep interval to which each suspicious data belongs, obtain the information entropy of the sleep interval to which each suspicious data belongs; obtain the mean value of all suspicious data in the sleep interval corresponding to each suspicious data, calculate the absolute value of the difference between the data value of each suspicious data and the mean value, and use the product of the information entropy of the sleep interval to which each suspicious data belongs and the absolute value as the volatility of each suspicious data.

[0017] Preferably, the method for obtaining the second abnormal weight of each suspicious data is as follows:

[0018] According to the sampling time of each suspicious data in each vital sign dataset, obtain the vital sign data corresponding to the sampling time in the remaining each vital sign dataset as the comparison data of each suspicious data in the remaining each vital sign dataset;

[0019] Obtain the sleep intervals to which the respective comparison data corresponding to each suspicious data belong, calculate the information entropy of the sleep intervals to which the respective comparison data of each suspicious data belong and the mean value of all data in the sleep intervals; calculate the difference between each comparison data and the data mean value of the sleep interval to which the respective comparison data belong, and multiply the absolute value of the difference by the information entropy of the sleep interval to which the respective comparison data belong, and the obtained result is used as the volatility of each comparison data;

[0020] Calculate the ratio between the volatility of the respective comparison data of each suspicious data and the volatility of each suspicious data, and obtain the second abnormal weight value of each suspicious data according to the mean value of all ratios.

[0021] Preferably, the method for obtaining abnormal data is:

[0022] Set an abnormal threshold according to the empirical value. When the abnormal degree of each suspicious data is greater than the abnormal threshold, it is considered that each suspicious data is abnormal data, otherwise the suspicious data is not abnormal data.

[0023] The beneficial effects of the present invention are: by collecting various vital sign data representing the health conditions of neurological patients during the sleep process, and then using the characteristics that the various vital sign data of neurological patients during the sleep process vary greatly between different sleep stages to divide different sleep stages, and then dividing multiple sleep intervals according to the continuity of the sampling time of the single vital sign data within the same sleep stage, and initially extracting suspicious data according to the length of the sleep interval; distinguishing normal data from pseudo-abnormal data and abnormal data according to the volatility of the sleep intervals to which each suspicious data belongs and the length of the suspicious intervals to which each suspicious data belongs, and then further distinguishing pseudo-abnormal data and abnormal data according to the volatility of the vital sign data collected at the corresponding moments of each suspicious data in different vital sign datasets, and sending the obtained abnormal data to the doctor, so as to ensure that the abnormal data generated by the various vital signs of neurological patients during the sleep process can be obtained more accurately and conveniently, so as to assist the doctor in specifically identifying the abnormal signs that appear during the sleep process of neurological patients. Description of the Drawings

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0025] Figure 1 It is a step flowchart of a method for processing health sign monitoring data of neurological patients according to the present invention. Detailed implementation manners

[0026] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following will, in conjunction with the accompanying drawings and preferred embodiments, describe in detail the specific implementation manners, structures, features, and effects of a method for processing health sign monitoring data of neurological patients proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0027] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0028] The following will specifically describe the specific solution of a method for processing health sign monitoring data of neurological patients provided by the present invention with reference to the accompanying drawings.

[0029] Please refer to Figure 1 , which shows a step flowchart of a method for processing health sign monitoring data of neurological patients provided by an embodiment of the present invention. The method includes the following steps:

[0030] Step S001: Obtain different sign data at each moment during the sleep of neurological patients to obtain a plurality of sign data sets.

[0031] First, collect various sign data during the entire sleep process of neurological patients. The sign data obtained in the present invention includes blood pressure SA, muscle activity SB, and blood oxygen level SC. The specific collection method can be sampled according to the specific equipment of each hospital. The sampling interval is 1 second, that is, all types of the above-mentioned sign data are sampled once every second. Collect various sign data during the entire sleep process of neurological patients. The data values of each type of sign data during the entire sleep process constitute the corresponding sign data set. Taking blood pressure SA as an example, the corresponding blood pressure sign data set , where is the th sampling time, It is the sampling time corresponding to when the body of a neurology patient just enters sleep. It is the sampling time corresponding to when the sleep of the body of a neurology patient ends. is the blood pressure data of the body of a neurology patient collected at the

[0032] Similarly, the muscle activity data and blood oxygen level data obtained at different sampling times during the entire sleep process respectively constitute a muscle activity sign data set and a blood oxygen level sign data set, which are not elaborated in this invention.

[0033] Thus, multiple sign data sets of the body of a neurology patient are obtained.

[0034] Step S002: Obtain the respective sleep stage data sets corresponding to each sign data set, divide each sleep stage data set into multiple sleep intervals; obtain the suspicious data in each sign data set according to the lengths of the respective sleep intervals.

[0035] In the sleep sign data of the body, it is generally divided into three types: normal data, pseudo-abnormal data, and abnormal data. Among them, normal data is the data generated when all signs of the body are normal during sleep; abnormal data is that a certain sign data of the body is different from the normal situation in this sleep stage. In this case, the abnormal data often has a certain duration, and due to the certain correlation relationship between different signs, when abnormal data is generated due to the abnormality of a certain sign, the other sign data will also show relatively similar abnormal phenomena; while pseudo-abnormal data refers to that when an individual is sleeping, only the data of a certain sign is abnormal at a certain moment. The generation of this abnormal data lasts for a very short time and only a certain sign data is abnormal, which will not affect the other sign data. The generation of pseudo-abnormal data is probably caused by the external environment or the accidental reaction of the body. This kind of pseudo-abnormal data cannot truly reflect the abnormal sleep signs of the body. Therefore, in this invention, the first abnormal weight and the second abnormal weight are used to judge the abnormal data of the sign data generated by an individual during sleep, exclude the pseudo-abnormal data, and identify the abnormal data.

[0036] There is a biological rhythm in the sleep process of the body, that is, approximately at Within a period of minutes, there is a cycle with 5 different stages. The International Classification of Sleep Disorders divides sleep stages into five stages: the stage of falling asleep, light sleep stage, deep sleep stage, slow-wave sleep stage, and rapid eye movement (REM) stage. That is, the entire sleep process contains multiple sleep cycles, and each sleep cycle contains five sleep stages. Normally, due to the differences in the body's activities during different sleep stages, the physical sign data in different sleep stages vary greatly, while the physical sign data values within the same sleep stage are relatively similar. Therefore, the present invention uses the K-means algorithm to cluster the data values in each physical sign dataset separately, sets the number of clustering results of K-means to 5, and each physical sign dataset can obtain five clustering results. Each clustering result corresponds to a sleep stage dataset, that is, each physical sign dataset can obtain five sleep stage datasets. The data in each sleep stage dataset are the data of this stage during different sleep cycles.

[0037] Taking the blood pressure physical sign dataset as an example, after clustering this physical sign dataset by K-means, five sleep stage datasets are obtained, which are respectively denoted as , and each sleep stage dataset may contain abnormal data, that is, data that should not belong to this sleep stage. However, due to the relatively close data values, this data is wrongly classified into this sleep stage dataset. And the sampling time of the blood pressure physical sign data in the same stage of the same cycle is continuous. Therefore, the present invention divides the data of different cycles in each sleep stage dataset according to the continuity of the data sampling time, that is, sorts the data in each sleep stage dataset in ascending order of sampling time, and divides the data with continuous sampling time into one interval. For example, assuming the blood pressure physical sign data contained therein are respectively , , , , , , , , , , , then will be divided into 4 sleep intervals, which are respectively: , , , , and , , .

[0038] Then, set the screening threshold based on the number of data in the longest interval, so as to conduct a check for normal data and determine suspicious data, and obtain The maximum interval length of all sleep intervals , that is, the maximum value of the number of blood pressure sign data included in each sleep interval. Set the adjustment parameter to according to the screening threshold, then the screening threshold is . All the sign data included in the sleep intervals with an interval length greater than the screening threshold are normal data, and all the sign data included in the sleep intervals with an interval length less than the screening threshold are suspicious data.

[0039] Similarly, for the sleep stage data set , perform the division of sleep intervals and the determination of suspicious data to obtain all the suspicious data in the blood pressure sign data set.

[0040] Use the above method to process other sign data sets, and finally obtain all the suspicious data in all the sign data sets.

[0041] Step S003: Obtain the first abnormal weight and the second abnormal weight of each suspicious data, and then obtain the degree of abnormality of each suspicious data.

[0042] First, sort all the suspicious data in each sign data set in ascending order of sampling time, and then divide the suspicious data with continuous sampling time into one interval to obtain each suspicious interval corresponding to each suspicious data. So far, each suspicious data corresponds to a sleep interval and a suspicious interval;

[0043] Since the suspicious data in the present invention is judged according to the length of each sleep interval, and the sleep interval refers to the sign data of a certain stage in a cycle, it cannot be excluded that when a certain sleep stage appears pseudo-abnormal data or abnormal data at the end part of a cycle, which may lead to the normal data at the end part of this sleep stage being divided into a separate sleep interval. At this time, judging according to the interval length will judge it as suspicious data, that is to say, the suspicious data obtained in step S002 may be normal data, pseudo-abnormal data and abnormal data.

[0044] Among them, the normal data within the same sleep stage changes periodically, and the sizes of the normal data within the same sleep stage are similar, that is, the data volatility of the sleep interval formed by the normal data is small, while neither the pseudo-abnormal data nor the abnormal data can meet this characteristic. Therefore, the present invention characterizes whether there is a periodic fluctuation of the data in the sleep interval according to the degree of chaos of the data in the sleep interval to which each suspicious data belongs, and judges whether the sizes of the data in the sleep interval are relatively similar according to the difference between each suspicious data and the mean value of all the suspicious data in the sleep interval to which it belongs. Distinguish normal data from abnormal data and pseudo-abnormal data according to these two characteristics;

[0045] The difference between abnormal data and pseudo-abnormal data is that pseudo-abnormal data appears accidentally and does not have continuity in time. Therefore, the present invention further determines whether the suspicious data has continuity in time according to the length of the suspicious interval corresponding to each suspicious data. Taking the i-th suspicious data in the blood pressure sign data set as an example, the sleep interval to which the suspicious data belongs is denoted as , and the suspicious interval to which the suspicious data belongs is . The first abnormal weight of the suspicious data can be expressed as:

[0046]

[0047] Wherein, is the interval length of the suspicious interval to which the i-th suspicious data in the blood pressure sign data set belongs; is the information entropy of the sleep interval to which the i-th suspicious data in the blood pressure sign data set belongs; is the data value of the suspicious data; is the average value of all suspicious data in the sleep interval to which the suspicious data belongs.

[0048] Since entropy is generally used to characterize the degree of chaos of data, the present invention uses the information entropy of the data values in the sleep interval to which the i-th suspicious data belongs to characterize the degree of chaos of the data in this sleep interval. The calculation of entropy is a well-known technology and will not be elaborated here. The smaller the entropy value, the stronger the periodic change of the data in this sleep interval. The larger this value, the more chaotic the data in this sleep interval, and the greater the possibility that the suspicious data is abnormal data. Further, according to the difference between the data value of the suspicious data and the average value of all data in the sleep interval corresponding to the suspicious data, it is judged whether the data in this sleep interval is similar in size. The greater the difference, the more abnormal the suspicious data. That is, the present invention takes as the volatility of the sleep interval of the i-th suspicious data, denoted as , to distinguish normal data, pseudo-abnormal data, and abnormal data;

[0049] Then, according to the interval length of the suspicious interval to which the i-th suspicious data belongs, it is used as another index to evaluate whether a suspicious data is abnormal data. The longer the length of the suspicious interval, the longer the duration of the occurrence of the suspicious data, and the greater the probability that the corresponding suspicious data is abnormal data, thereby excluding pseudo-abnormal data.

[0050] The above method analyzes suspicious data in the same set of physical sign data, and it cannot be excluded that some suspicious data is accidental. For example, the interval length of the suspicious interval to which the suspicious data belongs is short, and at this time, it is still impossible to distinguish pseudo-abnormal data from abnormal data. Therefore, there is a certain deviation in judging the suspicious data as abnormal data only based on the first abnormal weight. Considering that when a certain physical sign data is abnormal data, due to a certain correlation between different physical signs, when this abnormal physical sign data occurs, abnormal data will also appear in other physical sign data, while there is no correlation between different physical signs when pseudo-abnormal data occurs. Therefore, according to the acquisition time of each suspicious data in each set of physical sign data, the physical sign data with the same sampling time in the remaining sets of physical sign data is obtained as the comparison data of each suspicious data in the remaining sets of physical sign data. Then, the sleep intervals corresponding to the comparison data in the remaining sets of physical sign data are obtained, and further the volatility of each comparison data is obtained. The calculation method of the volatility of the comparison data is the same as that of the suspicious data, that is, first calculate the information entropy and the data mean of the sleep interval corresponding to the comparison data, and then the absolute value of the difference between the value of each comparison data and the corresponding data mean, and the result of multiplying the information entropy and the absolute value corresponding to each comparison data is used as the volatility of the comparison data.

[0051] Taking the difference between the volatility of each suspicious data and the volatility of its corresponding comparison data as the second abnormal weight of each suspicious data. Taking the blood pressure physical sign data set as an example, the second abnormal weight of the i-th suspicious data in this physical sign data set can be expressed as:

[0052]

[0053] where n is the number of remaining physical sign data sets, and in the present invention, n = 2; is the volatility of the i-th suspicious data, that is , and this value is obtained when calculating the first abnormal weight; is the volatility of the comparison data of the i-th suspicious data in the remaining j-th physical sign data set.

[0054] Since compared with normal data, the volatility of the sleep interval data corresponding to pseudo-abnormal data and abnormal data is relatively large, and for abnormal data, similar volatility will occur in other physical sign data sets. Therefore, when the volatility of the comparison data is compared with the volatility of the corresponding suspicious data, the obtained result is less than 1 or close to 1. Therefore, when the value is closer to 1, it indicates that the correlation between different physical sign data is stronger, and the probability that the corresponding suspicious data is abnormal data is greater; when The farther the value is from 1, the weaker the correlation between different vital sign data. At this time, the probability that the corresponding suspicious data is pseudo-abnormal data is relatively high. That is, the present invention further distinguishes pseudo-abnormal data from abnormal data according to the change of the vital sign data corresponding to the suspicious data at the same moment in different vital sign data sets, so as to ensure the accuracy of the judgment result.

[0055] Take the product of the first abnormal weight and the second abnormal weight of each suspicious data in each vital sign data set as the degree of abnormality of the suspicious data. Then the degree of abnormality of the i-th suspicious data in the blood pressure vital sign data set is:

[0056]

[0057] Step S004: Obtain abnormal data according to the degree of abnormality of each suspicious data; identify the abnormal vital signs of neurological patients according to the abnormal data.

[0058] Perform normalization processing on the degree of abnormality of each suspicious data in each vital sign data set, that is, obtain the maximum degree of abnormality and the minimum degree of abnormality of all suspicious data in each vital sign data set, subtract the degree of abnormality of each suspicious data from the minimum degree of abnormality, and take the ratio of the obtained difference to the difference between the maximum degree of abnormality and the minimum degree of abnormality as the normalized degree of abnormality of the suspicious data. For example, the normalization result of the degree of abnormality of the i-th suspicious data in the blood pressure vital sign data set can be expressed as:

[0059]

[0060] Among them, are the maximum degree of abnormality and the minimum degree of abnormality of all suspicious data in the blood pressure vital sign data set respectively, is the degree of abnormality of the i-th suspicious data.

[0061] Set the abnormality threshold according to experience , when the normalized degree of abnormality of the i-th suspicious data is greater than , it is considered that the suspicious data is abnormal data, otherwise it is normal data.

[0062] Similarly, obtain all the abnormal data in each vital sign data set.

[0063] Send the abnormal data and the corresponding sampling time in each vital sign data set to the doctor, and the doctor further identifies these abnormal data, so as to assist the doctor in identifying the abnormal vital signs of neurological patients during sleep.

[0064] Through the above steps, the recognition of abnormal sleep signs of neurological patients is completed, so as to realize the processing of health sign monitoring data of neurological patients.

[0065] In the present invention, various sign data during the sleep process of neurological patients are collected, and then, by using the feature that the various sign data during the sleep process of neurological patients vary greatly between different sleep stages, different sleep stages are divided. Then, multiple sleep intervals are divided according to the continuity of the sampling time of single sign data within the same sleep stage, and suspicious data are initially extracted according to the length of the sleep intervals; normal data are distinguished from pseudo-abnormal data and abnormal data according to the volatility of the sleep intervals to which the respective suspicious data belong and the length of the suspicious intervals to which the respective suspicious data belong. Then, the pseudo-abnormal data and abnormal data are further distinguished according to the volatility of the sign data collected at the corresponding moments in different sign data sets for the respective suspicious data, and the obtained abnormal data are sent to the doctor, so as to ensure that abnormal data generated by various signs during the sleep process of neurological patients can be obtained more accurately and conveniently, so as to assist the doctor in specifically identifying abnormal signs that occur during the sleep process of neurological patients.

[0066] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for processing health sign monitoring data of neurological patients, characterized in that, The method includes the following steps: Obtain different physical sign data at each moment during the sleep of neurological patients to obtain multiple physical sign data sets; Cluster according to the magnitudes of the data values in each physical sign data set to obtain multiple sleep stage data sets for each physical sign data set; obtain multiple sleep intervals for each sleep stage data set according to the continuity of the sampling times of the data in each sleep stage data set; obtain suspicious data for each physical sign data set according to the interval lengths of the respective sleep intervals; Obtain the suspicious intervals for each physical sign data set according to the continuity of the sampling times of all the suspicious data in each physical sign data set; obtain the volatility of each of the suspicious data according to all the suspicious data in the sleep interval to which the respective suspicious data belongs; obtain the first anomaly weight for each of the suspicious data in each physical sign data set according to the volatility of each of the suspicious data in each physical sign data set and the length of the suspicious interval to which the respective suspicious data belongs, the first anomaly weight being determined according to the length of the suspicious interval to which the suspicious data belongs, the information entropy of the sleep interval to which the suspicious data belongs, and the difference between the data value of the suspicious data and the mean of all the data in the sleep interval to which the suspicious data belongs, the information entropy being used to characterize the degree of chaos of the sleep interval to which the suspicious data belongs, and the difference being used to characterize the similarity between the magnitudes of all the data in the sleep interval to which the suspicious data belongs; obtain the comparison data of each of the suspicious data in other physical sign data sets according to the sampling times of each of the suspicious data in each physical sign data set, and obtain the second anomaly weight for each of the suspicious data according to the volatility of the sleep interval to which the comparison data belongs; obtain the anomaly degree of each of the suspicious data according to the first anomaly weight and the second anomaly weight of each of the suspicious data; Obtain abnormal data according to the anomaly degree of each of the suspicious data; identify the abnormal physical signs of neurological patients based on the abnormal data.

2. A method for processing health sign monitoring data of neurological patients according to claim 1, characterized in that, The method for obtaining the multiple sleep intervals of each sleep stage data set is as follows: Sort the physical sign data in each sleep stage data set in ascending order of sampling time, and divide the physical sign data with a difference of 1 between the sampling times into one interval to obtain multiple sleep intervals in each sleep stage data set.

3. A method for processing health sign monitoring data of neurological patients according to claim 1, characterized in that, The method for obtaining the suspicious data of each physical sign data set is as follows: Obtain the maximum interval length of each sleep interval in each sleep stage data set, and use the product of a preset adjustment parameter and the maximum interval length as the reference length; obtain all the sleep intervals in each sleep stage data set with an interval length less than the reference length, and use the data in all the sleep intervals as the suspicious data in each sleep stage data set, and the suspicious data in all the sleep stage data sets constitute the suspicious data of each physical sign data set.

4. A method for processing health sign monitoring data of neurological patients according to claim 1, characterized in that, The method for obtaining the volatility of each of the suspicious data in each physical sign data set is as follows: Obtain the information entropy of the sleep intervals to which the respective suspicious data belong according to the occurrence frequencies of the respective suspicious data in the sleep intervals to which they belong; obtain the mean value of all the suspicious data in the sleep intervals corresponding to the respective suspicious data, calculate the absolute value of the difference between the data value of each suspicious data and the mean value, and use the product of the information entropy of the sleep interval to which each suspicious data belongs and the absolute value as the volatility of each suspicious data.

5. A method for processing health sign monitoring data of neurological patients according to claim 1, characterized in that, The method for obtaining the second anomaly weight value of each suspicious data is as follows: According to the sampling time of each suspicious data in each vital sign dataset, obtain the vital sign data corresponding to the sampling time in the remaining vital sign datasets as the comparison data of each suspicious data in the remaining vital sign datasets; Obtain the sleep intervals to which the respective comparison data corresponding to the respective suspicious data belong, calculate the information entropy of the sleep intervals to which the respective comparison data of each suspicious data belong and the mean value of all the data in the sleep intervals; Calculate the difference between each comparison data and the data mean value of the sleep interval to which the respective comparison data belong, multiply the absolute value of the difference by the information entropy of the sleep interval to which the respective comparison data belong, and use the obtained result as the volatility of each comparison data; Calculate the ratio between the volatility of each comparison data of each suspicious data and the volatility of each suspicious data, and obtain the second anomaly weight value of each suspicious data according to the mean value of all the ratios.

6. A method for processing health sign monitoring data of neurological patients according to claim 1, characterized in that, The method for obtaining the abnormal data is as follows: Set an anomaly threshold according to the empirical value. When the anomaly degree of each suspicious data is greater than the anomaly threshold, consider each suspicious data as abnormal data; otherwise, the suspicious data is not abnormal data.

Images