CN117195139B - Chronic disease health data dynamic monitoring method based on machine learning - Google Patents
Chronic disease health data dynamic monitoring method based on machine learning Download PDFInfo
- Publication number
- CN117195139B CN117195139B CN202311475875.2A CN202311475875A CN117195139B CN 117195139 B CN117195139 B CN 117195139B CN 202311475875 A CN202311475875 A CN 202311475875A CN 117195139 B CN117195139 B CN 117195139B
- Authority
- CN
- China
- Prior art keywords
- health monitoring
- leaf node
- monitoring data
- data included
- binary tree
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000036541 health Effects 0.000 title claims abstract description 337
- 238000012544 monitoring process Methods 0.000 title claims abstract description 327
- 208000017667 Chronic Disease Diseases 0.000 title claims abstract description 64
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000010801 machine learning Methods 0.000 title claims abstract description 22
- 230000002159 abnormal effect Effects 0.000 claims abstract description 121
- 230000005856 abnormality Effects 0.000 claims abstract description 83
- 238000012937 correction Methods 0.000 claims abstract description 22
- 238000012545 processing Methods 0.000 claims abstract description 13
- 230000036772 blood pressure Effects 0.000 claims description 18
- 230000006870 function Effects 0.000 claims description 9
- 206010020772 Hypertension Diseases 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 6
- 238000001514 detection method Methods 0.000 abstract description 11
- 208000024172 Cardiovascular disease Diseases 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 201000010099 disease Diseases 0.000 description 2
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 208000014085 Chronic respiratory disease Diseases 0.000 description 1
- 206010028980 Neoplasm Diseases 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 201000011510 cancer Diseases 0.000 description 1
- 206010012601 diabetes mellitus Diseases 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention relates to the technical field of data processing, in particular to a chronic disease health data dynamic monitoring method based on machine learning, which comprises the following steps: acquiring health monitoring data corresponding to chronic diseases to be monitored of a patient to be monitored within a preset time period; performing abnormal fluctuation analysis processing on each health monitoring data included in each leaf node in each constructed target binary tree; determining an abnormal deviation index, a target abnormal index and a correction length corresponding to each piece of health monitoring data included in each leaf node; determining a target abnormality score corresponding to each health monitoring data through an isolated forest algorithm according to all correction lengths corresponding to each health monitoring data; judging whether the patient to be monitored has abnormal chronic disease conditions within a preset time period. According to the invention, through data processing on the health monitoring data, the accuracy of abnormality detection on the health data and the accuracy of monitoring on the chronic disease health data are improved.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a chronic disease health data dynamic monitoring method based on machine learning.
Background
With the development of technologies such as cloud computing, artificial intelligence, big data and the like, a machine learning algorithm is continuously and iteratively optimized, and powerful technical support is provided for dynamic monitoring of chronic disease health data. By combining machine learning and telemonitoring techniques, real-time monitoring and management of the patient may be achieved. This is particularly important for long-term chronically ill patients, who may have an improved quality of life.
In practical application, the historical health data of the chronic diseases are usually collected for analysis, whether the chronic diseases of the patient are abnormal is judged by detecting whether the historical health data are abnormal data or not, and the patient can respond to the abnormal quickly and can be treated timely in the subsequent dynamic monitoring process. Thus, anomaly detection of collected historical health data is critical. When abnormality detection is performed on data, the following methods are generally adopted: and according to the path lengths corresponding to all leaf nodes where the data are located, performing anomaly detection on the acquired data through an isolated forest algorithm.
However, when the collected historical health data is detected abnormally by the isolated forest algorithm directly according to the path lengths corresponding to all leaf nodes where the data is located, the following technical problems often exist:
Because the path lengths corresponding to the leaf nodes of the same layer in each binary tree in the isolated forest algorithm are always the same, but the historical health data contained in different leaf nodes are always different, so that some leaf nodes in the same layer of leaf nodes contain normal data, and some leaf nodes contain abnormal data.
Disclosure of Invention
The summary of the invention is provided to introduce a selection of concepts in a simplified form that are further described below in the detailed description. The summary of the invention is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.
In order to solve the technical problem of poor accuracy of monitoring chronic disease health data due to poor accuracy of abnormal detection of health data, the invention provides a dynamic monitoring method of chronic disease health data based on machine learning.
The invention provides a chronic disease health data dynamic monitoring method based on machine learning, which comprises the following steps:
acquiring health monitoring data corresponding to chronic diseases to be monitored of a patient to be monitored within a preset time period;
constructing a target binary tree set through an isolated forest algorithm according to all health monitoring data in the preset time period;
performing abnormal fluctuation analysis processing on each piece of health monitoring data included in each leaf node in each target binary tree in the target binary tree set to obtain an abnormal fluctuation index corresponding to each piece of health monitoring data included in each leaf node;
determining an abnormal deviation index corresponding to each piece of health monitoring data included in each leaf node according to all pieces of health monitoring data in the binary tree layer where each leaf node is located;
determining a target abnormality index corresponding to each health monitoring data included in each leaf node according to the abnormality fluctuation index and the abnormality deviation index corresponding to each health monitoring data included in each leaf node and all health monitoring data in all target binary trees;
correcting the path length corresponding to each leaf node according to the target abnormality index corresponding to each health monitoring data included in each leaf node to obtain the corrected length corresponding to each health monitoring data included in each leaf node;
Determining a target abnormality score corresponding to each health monitoring data in the preset time period through an isolated forest algorithm according to all correction lengths corresponding to each health monitoring data in the preset time period;
and judging whether the chronic disease abnormality to be monitored exists in the patient to be monitored within a preset time period according to all the target abnormality scores.
Optionally, the chronic disease to be monitored is hypertension, and the health monitoring data includes: high-voltage data and low-voltage data.
Optionally, the performing an abnormal fluctuation analysis processing on each piece of health monitoring data included in each leaf node in each target binary tree in the target binary tree set to obtain an abnormal fluctuation index corresponding to each piece of health monitoring data included in each leaf node includes:
performing curve fitting on high-voltage data included in all health monitoring data in the preset time period to obtain a high-voltage fluctuation curve, and performing curve fitting on low-voltage data included in all health monitoring data in the preset time period to obtain a low-voltage fluctuation curve, wherein the abscissa of the high-voltage fluctuation curve is the acquisition time, the ordinate of the high-voltage fluctuation curve is the high-voltage data, the abscissa of the low-voltage fluctuation curve is the acquisition time, and the ordinate of the low-voltage fluctuation curve is the low-voltage data;
Determining a curve segment between coordinate points corresponding to every two adjacent extremum values in the high-voltage fluctuation curve as a high-voltage sub-curve segment, and determining a curve segment between coordinate points corresponding to every two adjacent extremum values in the low-voltage fluctuation curve as a low-voltage sub-curve segment;
and determining an abnormal fluctuation index corresponding to each piece of health monitoring data included by the leaf node according to the high-pressure sub-curve segment to which the high-pressure data included by each piece of health monitoring data included by the leaf node belongs and the low-pressure sub-curve segment to which the low-pressure data included by the leaf node belongs.
Optionally, the formula corresponding to the abnormal fluctuation index corresponding to the health monitoring data included in the leaf node is:
；
；/>the method comprises the steps of carrying out a first treatment on the surface of the Wherein A is imj Is an abnormal fluctuation index corresponding to the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree; i is the sequence number of the target binary tree in the target binary tree set; m is the sequence number of the leaf node in the ith target binary tree; j is the serial number of the health monitoring data included in the mth leaf node; sigma (sigma) imj1 Is a high-voltage fluctuation factor corresponding to a high-voltage sub-curve segment where high-voltage data included in the j-th health monitoring data included in the m-th leaf node is located in the i-th target binary tree; sigma (sigma) 1 Is the accumulated value of the high-voltage fluctuation factors corresponding to all the high-voltage sub-curve segments; a, a imj Is the high-voltage data included in the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree; a is the accumulated value of all high-voltage data in the high-voltage sub-curve section where the high-voltage data included in the j-th health monitoring data included in the m-th leaf node in the ith target binary tree; sigma (sigma) imj2 In the ith target binary tree, the low-voltage fluctuation factor corresponding to the low-voltage sub-curve segment where the low-voltage data included in the jth health monitoring data included in the mth leaf node is located; sigma (sigma) 2 Is the accumulated value of the low-voltage fluctuation factors corresponding to all low-voltage sub-curve segments; b imj Is low-voltage data included in the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree; b is the accumulated value of all low-voltage data in the low-voltage sub-curve section where the low-voltage data included in the j-th health monitoring data included in the m-th leaf node is located in the i-th target binary tree; the absolute value function is taken; g imj1 And g imj2 In the ith target binary tree, the ordinate corresponding to the two endpoints of the high-voltage sub-curve segment where the high-voltage data included in the jth health monitoring data included in the mth leaf node are located; t is t imj1 And t imj2 In the ith target binary tree, the corresponding abscissa of two endpoints of a high-voltage sub-curve segment where the high-voltage data included in the jth health monitoring data included in the mth leaf node is located; g imj1 And G imj2 In the ith target binary tree, the ordinate corresponding to two endpoints of a low-voltage sub-curve segment where low-voltage data included in the jth health monitoring data included in the mth leaf node are located; t (T) imj1 And T imj2 In the ith target binary tree, the corresponding abscissa of two endpoints of the low-voltage sub-curve segment where the low-voltage data included in the jth health monitoring data included in the mth leaf node is located.
Optionally, the determining, according to all the health monitoring data in the binary tree layer where each leaf node is located, an abnormal deviation index corresponding to each health monitoring data included in each leaf node includes:
and determining an abnormal deviation index corresponding to each piece of health monitoring data included in the leaf node according to the low-voltage data and the high-voltage data included in each piece of health monitoring data included in the leaf node and the low-voltage data and the high-voltage data included in all pieces of health monitoring data in the binary tree layer where the leaf node is located.
Optionally, the formula corresponding to the abnormal deviation index corresponding to the health monitoring data included in the leaf node is:
θ imj =norm(a imj -B im +(b imj -H im ) A) is provided; wherein θ imj Is an abnormal deviation index corresponding to the j-th health monitoring data included in the mth leaf node in the ith target binary tree in the target binary tree set; i is the sequence number of the target binary tree in the target binary tree set; m is the sequence number of the leaf node in the ith target binary tree; j is the serial number of the health monitoring data included in the mth leaf node; norm () is a normalization function; a, a imj Is the high-voltage data included in the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree; b (B) im Is the average value of high-voltage data included in all health monitoring data in a binary tree layer where an mth leaf node in the ith target binary tree is located; b imj Is low-voltage data included in the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree; h im Is the average value of low-voltage data included in all health monitoring data in the binary tree layer where the mth leaf node is located in the ith target binary tree.
Optionally, the determining, according to the abnormality fluctuation index and the abnormality deviation index corresponding to each piece of health monitoring data included in each leaf node and all pieces of health monitoring data in all target binary trees, the target abnormality index corresponding to each piece of health monitoring data included in each leaf node includes:
Determining low-pressure data and high-pressure data included in the health monitoring data as blood pressure data;
and determining the target abnormal index corresponding to each health monitoring data of the leaf node according to the abnormal fluctuation index and the abnormal deviation index corresponding to each health monitoring data of the leaf node, the average value of all blood pressure data in the target binary tree where the leaf node is located and the average value of all blood pressure data in the target binary tree.
Optionally, the formula corresponding to the target abnormality index corresponding to the health monitoring data included in the leaf node is:
the method comprises the steps of carrying out a first treatment on the surface of the Wherein (1)>Is a target abnormality index corresponding to the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree in the target binary tree set; i is the sequence number of the target binary tree in the target binary tree set; m is the sequence number of the leaf node in the ith target binary tree; j is the serial number of the health monitoring data included in the mth leaf node; norm () is a normalization function; θ imj Is an abnormal deviation index corresponding to the j-th health monitoring data included in the mth leaf node in the ith target binary tree in the target binary tree set; a is that imj Is an abnormal fluctuation index corresponding to the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree in the target binary tree set; mu (mu) im The average value of all blood pressure data in the target binary tree where the mth leaf node in the ith target binary tree in the target binary tree set is located; μ is the average of all blood pressure data in all target binary trees.
Optionally, the correcting the path length corresponding to each leaf node according to the target abnormality index corresponding to each health monitoring data included in each leaf node to obtain a corrected length corresponding to each health monitoring data included in each leaf node includes:
and determining the product of the target abnormality index corresponding to each piece of health monitoring data included in the leaf node and the path length corresponding to the leaf node as the correction length corresponding to each piece of health monitoring data included in the leaf node.
Optionally, the determining whether the patient to be monitored has the chronic disease abnormality to be monitored within the preset time period according to all the target abnormality scores includes:
when the target abnormality score corresponding to the health monitoring data is larger than a preset abnormality threshold, determining the health monitoring data as target abnormality data;
and when the target abnormal data exist in the preset time period, judging that the chronic disease condition abnormality to be monitored exists in the patient to be monitored in the preset time period.
The invention has the following beneficial effects:
according to the machine learning-based dynamic monitoring method for the chronic disease health data, the technical problem of poor accuracy of monitoring the chronic disease health data caused by poor accuracy of abnormal detection of the health data is solved by carrying out data processing on the health monitoring data, and the accuracy of abnormal detection of the health data and the accuracy of monitoring the chronic disease health data are improved. Firstly, acquiring each health monitoring data corresponding to the chronic disease to be monitored of the patient to be monitored within a preset time period, and conveniently carrying out abnormality detection on the health monitoring data. And then, constructing a target binary tree set, so that the abnormal condition of each health monitoring data included in each leaf node can be conveniently analyzed, and the misjudgment of an abnormal result caused by only considering the path length can be avoided to a certain extent. Then, the abnormal fluctuation analysis processing is performed on each health monitoring data included in each leaf node, so that the abnormal fluctuation index corresponding to each health monitoring data included in each leaf node can be quantified, and the larger the value of the abnormal fluctuation index is, the more likely that the health monitoring data included in the leaf node is abnormal data is indicated. Moreover, because the path lengths corresponding to the leaf nodes of the same layer in each target binary tree in the isolated forest algorithm are always the same, but the health monitoring data contained in different leaf nodes are always different, the accuracy of determining the abnormal deviation index corresponding to each health monitoring data contained in each leaf node can be improved by comprehensively considering all the health monitoring data in the binary tree layer where each leaf node is located, and the larger the value is, the more likely that the health monitoring data contained in the leaf node is abnormal data is. Continuing, the abnormal fluctuation index and the abnormal deviation index corresponding to each health monitoring data included in each leaf node and all health monitoring data in all target binary trees are comprehensively considered, so that the accuracy of determining the target abnormal index corresponding to each health monitoring data included in each leaf node can be improved, and the larger the value is, the more likely that the health monitoring data included in the leaf node is abnormal data is indicated. And then, correcting the path length corresponding to each leaf node based on the target abnormality index corresponding to each health monitoring data included in each leaf node, wherein the abnormal degrees of the health monitoring data included in different leaf nodes of the same layer are possibly different, and the corresponding correction lengths are often different, so that the health monitoring data with different abnormal degrees in the same layer can be distinguished by the correction lengths to a certain extent, and the misjudgment of an abnormal result caused by considering only the path length can be avoided to a certain extent. Finally, based on all correction lengths corresponding to each health monitoring data, determining a target abnormality score corresponding to each health monitoring data through an isolated forest algorithm, judging whether a patient to be monitored has an abnormality of the chronic disease state in a preset time period based on all target abnormality scores, realizing the monitoring of the chronic disease health data, and compared with the method of directly detecting the abnormality of the collected health monitoring data according to the path lengths corresponding to all leaf nodes where the data are located through the isolated forest algorithm, the method quantifies the abnormality condition of each health monitoring data included in each leaf node, for example, the quantified correction lengths can distinguish health monitoring data with different abnormality degrees included in the same layer to a certain extent, and can avoid the erroneous judgment of the abnormality result caused by considering only the path length to a certain extent, thereby improving the accuracy of abnormality detection of the health data and further improving the accuracy of monitoring the chronic disease health data.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flow chart of the machine learning based chronic disease health data dynamic monitoring method of the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the present invention to achieve the preset purpose, the following detailed description is given below of the specific implementation, structure, features and effects of the technical solution according to the present invention with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, 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 belongs.
The invention provides a machine learning-based chronic disease health data dynamic monitoring method, which comprises the following steps:
acquiring health monitoring data corresponding to chronic diseases to be monitored of a patient to be monitored within a preset time period;
constructing a target binary tree set through an isolated forest algorithm according to all health monitoring data in a preset time period;
performing abnormal fluctuation analysis processing on each piece of health monitoring data included in each leaf node in each target binary tree in the target binary tree set to obtain an abnormal fluctuation index corresponding to each piece of health monitoring data included in each leaf node;
determining an abnormal deviation index corresponding to each piece of health monitoring data included in each leaf node according to all pieces of health monitoring data in the binary tree layer where each leaf node is located;
determining a target abnormality index corresponding to each health monitoring data included in each leaf node according to the abnormality fluctuation index and the abnormality deviation index corresponding to each health monitoring data included in each leaf node and all health monitoring data in all target binary trees;
correcting the path length corresponding to each leaf node according to the target abnormality index corresponding to each health monitoring data included in each leaf node to obtain the corrected length corresponding to each health monitoring data included in each leaf node;
Determining a target abnormality score corresponding to each health monitoring data in a preset time period through an isolated forest algorithm according to all correction lengths corresponding to each health monitoring data in the preset time period;
and judging whether the chronic disease abnormality to be monitored exists in the patient to be monitored within a preset time period according to all the target abnormality scores.
The following detailed development of each step is performed:
referring to fig. 1, a flow of some embodiments of a machine learning based chronic disease health data dynamic monitoring method according to the present invention is shown. The chronic disease health data dynamic monitoring method based on machine learning comprises the following steps:
step S1, acquiring each health monitoring data corresponding to the chronic disease to be monitored of the patient to be monitored within a preset time period.
In some embodiments, each health monitoring data corresponding to a chronic disease to be monitored for a preset period of time for a patient to be monitored may be obtained.
Wherein, the patient to be monitored can be a patient to be monitored for chronic disease health condition. The preset time period may be a preset time period. For example, the duration corresponding to the preset time period may be one week. The chronic disease to be monitored can be a preset chronic disease to be subjected to health monitoring. Chronic diseases with long disease course and complex etiology, which are difficult to cure, include cardiovascular diseases, diabetes, cancer, chronic respiratory diseases, etc. The prevention and control of these diseases requires long-term management and treatment, and therefore dynamic monitoring is very important for the health management of chronically ill patients. For example, the chronic disease to be monitored may be hypertension in cardiovascular diseases. The health monitoring data may be data of multiple dimensions related to the chronic disease to be monitored. For example, if the chronic disease to be monitored is hypertension, the health monitoring data may include two dimensional data, which may be high pressure data and low pressure data, respectively. If the chronic disease to be monitored is hypertension, the health monitoring data may include: high-voltage data and low-voltage data. Wherein the high voltage data is also called high voltage. Low voltage data is also known as low voltage.
It should be noted that, obtaining each health monitoring data corresponding to the chronic disease to be monitored of the patient to be monitored within the preset time period can facilitate the subsequent abnormal detection of the health monitoring data.
As an example, if the chronic disease to be monitored is hypertension, and the duration corresponding to the preset time period is one week, the high-pressure data and the low-pressure data of the patient to be monitored may be collected once every preset time period in one week, and combined into the health monitoring data. The preset duration may be a preset duration. For example, the preset duration may be half an hour.
And S2, constructing a target binary tree set through an isolated forest algorithm according to all health monitoring data in a preset time period.
In some embodiments, the target binary tree set may be constructed according to all the health monitoring data within the preset time period through an isolated forest algorithm.
It should be noted that, the objective binary tree set is constructed, so that the abnormal condition of each health monitoring data included in each leaf node can be conveniently analyzed later, and misjudgment of an abnormal result caused by considering only path length can be avoided to a certain extent.
As an example, a plurality of binary trees can be obtained through an isolated forest algorithm according to all health monitoring data in a preset time period, each binary tree is taken as a target binary tree, and all the target binary trees are combined into a target binary tree set.
And S3, carrying out abnormal fluctuation analysis processing on each piece of health monitoring data included in each leaf node in each target binary tree in the target binary tree set to obtain an abnormal fluctuation index corresponding to each piece of health monitoring data included in each leaf node.
In some embodiments, the abnormal fluctuation analysis processing may be performed on each health monitoring data included in each leaf node in each target binary tree in the target binary tree set, so as to obtain an abnormal fluctuation index corresponding to each health monitoring data included in each leaf node.
It should be noted that, performing the abnormal fluctuation analysis processing on each health monitoring data included in each leaf node may quantify the abnormal fluctuation index corresponding to each health monitoring data included in each leaf node, and the larger the value, the more likely the health monitoring data included in the leaf node is abnormal data.
As an example, this step may include the steps of:
the first step, performing curve fitting on high-voltage data included in all the health monitoring data in the preset time period to obtain a high-voltage fluctuation curve, and performing curve fitting on low-voltage data included in all the health monitoring data in the preset time period to obtain a low-voltage fluctuation curve.
The abscissa of the high-voltage fluctuation curve may be the acquisition time. The ordinate of the high voltage fluctuation curve may be high voltage data. The abscissa of the low-voltage fluctuation curve may be the acquisition time. The ordinate of the low pressure fluctuation curve may be low pressure data.
And secondly, determining a curve segment between coordinate points corresponding to every two adjacent extreme values in the high-voltage fluctuation curve as a high-voltage sub-curve segment, and determining a curve segment between coordinate points corresponding to every two adjacent extreme values in the low-voltage fluctuation curve as a low-voltage sub-curve segment.
Thirdly, determining, according to the high-voltage sub-curve segment to which the high-voltage data included in each health monitoring data included in the leaf node belongs and the low-voltage sub-curve segment to which the low-voltage data included in the high-voltage sub-curve segment belongs, a formula corresponding to an abnormal fluctuation index corresponding to each health monitoring data included in the leaf node may be:
；
；/>the method comprises the steps of carrying out a first treatment on the surface of the Wherein A is imj Is an abnormal fluctuation index corresponding to the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree. i is the sequence number of the target binary tree in the set of target binary trees. m is the sequence number of the leaf node in the ith target binary tree. j is the sequence number of the health monitoring data included in the mth leaf node. Sigma (sigma) imj1 Is a high-voltage fluctuation factor corresponding to a high-voltage sub-curve segment where high-voltage data included in the j-th health monitoring data included in the m-th leaf node is located in the i-th target binary tree; sigma (sigma) imj1 The j-th health monitoring data includes high-voltage data which is not extreme, and when the high-voltage data falls on one high-voltage sub-curve segment, the high-voltage fluctuation factor corresponding to the high-voltage sub-curve segment where the high-voltage data is located is calculated as an example. If the high-voltage data included in the j-th health monitoring data is an extremum, the high-voltage data falls on two high-voltage sub-curve segments, and the average value of the high-voltage fluctuation factors corresponding to the two high-voltage sub-curve segments can be used as the high-voltage fluctuation factor corresponding to the whole high-voltage sub-curve segment where the high-voltage data is located. Sigma (sigma) 1 Is the accumulated value of the high voltage fluctuation factors corresponding to all the high voltage sub-curve segments. a, a imj Is the high voltage data included in the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree. a is the high voltage of the high voltage data included in the j-th health monitoring data included in the m-th leaf node in the i-th target binary treeThe accumulated value of all high voltage data in the sub-curve segment. Sigma (sigma) imj2 In the ith target binary tree, the low-voltage fluctuation factor corresponding to the low-voltage sub-curve segment where the low-voltage data included in the jth health monitoring data included in the mth leaf node is located; sigma (sigma) imj2 The j-th health monitoring data includes low-voltage data which is not extreme, and when the low-voltage data falls on one low-voltage sub-curve segment, the low-voltage fluctuation factor corresponding to the low-voltage sub-curve segment is calculated. If the low-pressure data included in the j-th health monitoring data is an extremum, the low-pressure data falls on two low-pressure sub-curve segments, and the average value of the low-pressure fluctuation factors corresponding to the two low-pressure sub-curve segments can be used as the low-pressure fluctuation factor corresponding to the whole low-pressure sub-curve segment where the low-pressure data is located. Sigma (sigma) 2 Is the accumulated value of the low-voltage fluctuation factors corresponding to all the low-voltage sub-curve segments. b imj Is the low pressure data included in the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree. b is the accumulated value of all low-voltage data in the low-voltage sub-curve segment where the low-voltage data included in the j-th health monitoring data included in the m-th leaf node is located in the i-th target binary tree. I is a function taking absolute value. g imj1 And g imj2 In the ith target binary tree, the ordinate corresponding to the two endpoints of the high-voltage sub-curve segment where the high-voltage data included in the jth health monitoring data included in the mth leaf node are located. t is t imj1 And t imj2 In the ith target binary tree, the corresponding abscissa of two endpoints of a high-voltage sub-curve segment where the high-voltage data included in the jth health monitoring data included in the mth leaf node is located. G imj1 And G imj2 In the ith target binary tree, the ordinate corresponding to the two endpoints of the low-voltage sub-curve segment where the low-voltage data included in the jth health monitoring data included in the mth leaf node is located. T (T) imj1 And T imj2 In the ith target binary tree, the corresponding abscissa of two endpoints of the low-voltage sub-curve segment where the low-voltage data included in the jth health monitoring data included in the mth leaf node is located.
When sigma imj1 The larger the value, the more fluctuation degree of the high-voltage sub-curve section where the j-th health monitoring data includes is, the more likely the j-th health monitoring data includes is abnormal data. Thus, whenThe larger the data, the more abnormal the data of the high voltage included in the j-th health monitoring data is often described. When sigma is imj2 The larger the value, the more fluctuation degree of the low-voltage sub-curve section where the low-voltage data included in the jth health monitoring data is located is often indicated, and the more likely the low-voltage data included in the jth health monitoring data is abnormal data is often indicated. Thus, when- >The larger the data, the more abnormal the low-voltage data included in the jth health monitoring data tends to be. When a is imj The larger the data, the larger the high-voltage data included in the jth health monitoring data is, the more likely the high-voltage data included in the jth health monitoring data is abnormal data is. Thus->The larger the data, the more abnormal the data of the high voltage included in the j-th health monitoring data is often described. When b imj The larger the data, the larger the low-voltage data included in the jth health monitoring data is, the more likely the low-voltage data included in the jth health monitoring data is abnormal. So when->The larger the data, the more abnormal the low-voltage data included in the jth health monitoring data tends to be. Thus when A imj The larger the data, the more abnormal the j health monitoring data, the more likely the patient to be monitored is abnormal.
And S4, determining an abnormal deviation index corresponding to each piece of health monitoring data included in each leaf node according to all pieces of health monitoring data in the binary tree layer where each leaf node is located.
In some embodiments, the abnormal deviation index corresponding to each health monitoring data included in each leaf node may be determined according to all health monitoring data in the binary tree layer in which each leaf node is located.
The binary tree layer where the leaf node is located is that layer of the leaf node in the target binary tree.
It should be noted that, because the path lengths corresponding to the leaf nodes of the same layer in each target binary tree in the isolated forest algorithm are often the same, but the health monitoring data contained in different leaf nodes are often different, therefore, by comprehensively considering all the health monitoring data in the binary tree layer where each leaf node is located, the accuracy of determining the abnormal deviation index corresponding to each health monitoring data contained in each leaf node can be improved, and the larger the value is, the more likely that the health monitoring data contained in the leaf node is abnormal data is.
As an example, according to the low-voltage data and the high-voltage data included in each health monitoring data included in the leaf node, and the low-voltage data and the high-voltage data included in all the health monitoring data in the binary tree layer where the leaf node is located, the formula corresponding to the abnormal deviation index corresponding to each health monitoring data included in the leaf node may be determined as follows:
θ imj =norm(a imj -B im +(b imj -H im ) A) is provided; wherein θ imj Is an abnormal deviation index corresponding to the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree in the target binary tree set. i is the sequence number of the target binary tree in the set of target binary trees. m is the sequence number of the leaf node in the ith target binary tree. j is the sequence number of the health monitoring data included in the mth leaf node. norm () is a normalization function. a, a imj Is the high voltage data included in the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree. B (B) im Is the average value of the high-voltage data included in all the health monitoring data in the binary tree layer where the mth leaf node is located in the ith target binary tree. b imj Is the ithAnd the jth health monitoring data included in the mth leaf node in the target binary tree comprises low-voltage data. H im Is the average value of low-voltage data included in all health monitoring data in the binary tree layer where the mth leaf node is located in the ith target binary tree.
It should be noted that when a imj -B im When the data is larger, the high-voltage data included in the j-th health monitoring data included in the m-th leaf node is relatively higher than the high-voltage data included in the binary tree layer where the m-th leaf node is located, and the high-voltage data included in the j-th health monitoring data included in the m-th leaf node is relatively abnormal. When b imj -H im When the data is larger, the low-voltage data included in the j-th health monitoring data included in the m-th leaf node is relatively higher than the low-voltage data included in the binary tree layer where the m-th leaf node is located, and the low-voltage data included in the j-th health monitoring data included in the m-th leaf node is relatively abnormal. Thus when theta imj The larger the size, the more abnormal the jth health monitoring data included in the mth leaf node will be.
And S5, determining a target abnormality index corresponding to each health monitoring data included in each leaf node according to the abnormality fluctuation index and the abnormality deviation index corresponding to each health monitoring data included in each leaf node and all health monitoring data in all target binary trees.
In some embodiments, the target anomaly index corresponding to each health monitoring data included in each leaf node may be determined according to the anomaly fluctuation index and the anomaly deviation index corresponding to each health monitoring data included in each leaf node and all health monitoring data in all target binary trees.
It should be noted that, by comprehensively considering the abnormal fluctuation index and the abnormal deviation index corresponding to each health monitoring data included in each leaf node and all the health monitoring data in all the target binary trees, the accuracy of determining the target abnormal index corresponding to each health monitoring data included in each leaf node can be improved, and the larger the value is, the more likely that the health monitoring data included in the leaf node is abnormal data is indicated.
As an example, this step may include the steps of:
first, low-pressure data and high-pressure data included in the health monitoring data are determined as blood pressure data.
The second step, according to the abnormal fluctuation index and the abnormal deviation index corresponding to each health monitoring data of the leaf node, the average value of all blood pressure data in the target binary tree where the leaf node is located, and the average value of all blood pressure data in the target binary tree, determining that a formula corresponding to the target abnormal index corresponding to each health monitoring data included in the leaf node may be:
the method comprises the steps of carrying out a first treatment on the surface of the Wherein (1)>Is a target abnormality index corresponding to the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree in the target binary tree set. i is the sequence number of the target binary tree in the set of target binary trees. m is the sequence number of the leaf node in the ith target binary tree. j is the sequence number of the health monitoring data included in the mth leaf node. norm () is a normalization function. θ imj Is an abnormal deviation index corresponding to the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree in the target binary tree set. A is that imj Is an abnormal fluctuation index corresponding to the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree in the target binary tree set. Mu (mu) im Is the average value of all blood pressure data in the target binary tree where the mth leaf node in the ith target binary tree in the target binary tree set is located. μ is the average of all blood pressure data in all target binary trees.
It should be noted that when A imj The larger the data, the more abnormal the j health monitoring data, the more likely the patient to be monitored is abnormal. When theta is as imj The larger the size, the more often the mth leaf node is explainedThe more abnormal the j-th health monitoring data the point includes is relative. When mu im The greater μ, the greater the μ tends to indicate that the higher the blood pressure data in the target binary tree in which the mth leaf node is located is relative to the blood pressure data in the overall target binary tree, the more the blood pressure data in the target binary tree in which the mth leaf node is located is relative to abnormal, and the more the jth health monitoring data included in the mth leaf node is likely to be abnormal. Therefore, whenWhen the data is larger, it is often indicated that the j health monitoring data included in the mth leaf node in the ith target binary tree is relatively abnormal, and it is often indicated that the condition of the patient to be monitored is more likely to be abnormal at this time. Thus->The weight given to the path length corresponding to the mth leaf node when the j-th health monitoring data included in the mth leaf node is judged abnormally can be characterized.
And S6, correcting the path length corresponding to each leaf node according to the target abnormality index corresponding to each health monitoring data included in each leaf node, and obtaining the corrected length corresponding to each health monitoring data included in each leaf node.
In some embodiments, the path length corresponding to each leaf node may be corrected according to the target abnormality index corresponding to each health monitoring data included in each leaf node, so as to obtain the corrected length corresponding to each health monitoring data included in each leaf node.
It should be noted that, based on the target abnormality index corresponding to each health monitoring data included in each leaf node, the path length corresponding to each leaf node is corrected, and the abnormality degrees of the health monitoring data included in different leaf nodes of the same layer may be different, and the corresponding correction lengths are often different, so that the correction lengths can distinguish health monitoring data with different abnormality degrees in the same layer to a certain extent, and misjudgment of an abnormality result caused by considering only the path length can be avoided to a certain extent.
As an example, a product of the target abnormality index corresponding to each health monitoring data included in the leaf node and the path length corresponding to the leaf node may be determined as a correction length corresponding to each health monitoring data included in the leaf node.
And S7, determining a target abnormality score corresponding to each health monitoring data in a preset time period through an isolated forest algorithm according to all correction lengths corresponding to each health monitoring data in the preset time period.
In some embodiments, the target anomaly score corresponding to each health monitoring data in the preset time period may be determined according to all the correction lengths corresponding to each health monitoring data in the preset time period through an isolated forest algorithm.
It should be noted that, when the target binary tree set is constructed according to all the health monitoring data in the preset time period through the isolated forest algorithm, the same health monitoring data is often allocated to a plurality of target binary trees, and then the same health monitoring data may be located at a plurality of leaf nodes, and one health monitoring data included in one leaf node corresponds to one correction length, so that one health monitoring data in the preset time period corresponds to a plurality of correction lengths. And based on all correction lengths corresponding to each health monitoring data, the accuracy of determining the target abnormality score corresponding to each health monitoring data is improved.
As an example, this step may include the steps of:
And in the first step, determining the average value of all the correction lengths corresponding to each health monitoring data as the average path length corresponding to the health monitoring data.
And secondly, calculating the abnormal score of the health monitoring data through an isolated forest algorithm according to the average path length corresponding to each health monitoring data, and normalizing the abnormal score of the health monitoring data to obtain a target abnormal score corresponding to the health monitoring data.
When the data anomaly is detected by the isolated forest algorithm, the average value of the path lengths corresponding to all the leaf nodes where the data are located is often directly used as the average path length corresponding to the data.
And S8, judging whether the chronic disease condition abnormality to be monitored exists in the patient to be monitored within a preset time period according to all the target abnormality scores.
In some embodiments, whether the patient to be monitored has the chronic disease abnormality to be monitored within the preset time period can be judged according to all the target abnormality scores, so that the monitoring of the chronic disease health data is realized.
As an example, this step may include the steps of:
and determining the health monitoring data as target abnormal data when the target abnormal score corresponding to the health monitoring data is larger than a preset abnormal threshold value.
The preset abnormal threshold may be a preset threshold. For example, the preset anomaly threshold value may be 0.7.
And secondly, when target abnormal data exist in the preset time period, judging that the patient to be monitored has abnormal chronic disease condition to be monitored in the preset time period.
In summary, compared with the method for detecting the abnormality of the collected health monitoring data directly according to the path lengths corresponding to all the leaf nodes where the data are located through an isolated forest algorithm, the method provided by the invention quantifies the abnormality of each health monitoring data included in each leaf node, for example, the quantified correction length can relatively and accurately represent the abnormality of the corresponding health monitoring data included in the corresponding leaf node, the quantified correction length can distinguish health monitoring data with different abnormality degrees included in the same layer to a certain extent, and the erroneous judgment of the abnormality result caused by considering only the path length can be avoided to a certain extent, so that the accuracy of abnormality detection of the health data is improved, and the accuracy of monitoring chronic disease health data is further improved.
The above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention and are intended to be included within the scope of the invention.
Claims (9)
1. The chronic disease health data dynamic monitoring method based on machine learning is characterized by comprising the following steps of:
acquiring health monitoring data corresponding to chronic diseases to be monitored of a patient to be monitored within a preset time period;
constructing a target binary tree set through an isolated forest algorithm according to all health monitoring data in the preset time period;
performing abnormal fluctuation analysis processing on each piece of health monitoring data included in each leaf node in each target binary tree in the target binary tree set to obtain an abnormal fluctuation index corresponding to each piece of health monitoring data included in each leaf node;
Determining an abnormal deviation index corresponding to each piece of health monitoring data included in each leaf node according to all pieces of health monitoring data in the binary tree layer where each leaf node is located;
determining a target abnormality index corresponding to each health monitoring data included in each leaf node according to the abnormality fluctuation index and the abnormality deviation index corresponding to each health monitoring data included in each leaf node and all health monitoring data in all target binary trees;
correcting the path length corresponding to each leaf node according to the target abnormality index corresponding to each health monitoring data included in each leaf node to obtain the corrected length corresponding to each health monitoring data included in each leaf node;
determining a target abnormality score corresponding to each health monitoring data in the preset time period through an isolated forest algorithm according to all correction lengths corresponding to each health monitoring data in the preset time period;
judging whether the chronic disease condition abnormality to be monitored exists in the patient to be monitored within a preset time period according to all the target abnormality scores;
correcting the path length corresponding to each leaf node according to the target abnormality index corresponding to each health monitoring data included in each leaf node to obtain the corrected length corresponding to each health monitoring data included in each leaf node, including:
And determining the product of the target abnormality index corresponding to each piece of health monitoring data included in the leaf node and the path length corresponding to the leaf node as the correction length corresponding to each piece of health monitoring data included in the leaf node.
2. The machine learning based chronic disease health data dynamic monitoring method of claim 1, wherein the chronic disease to be monitored is hypertension, and the health monitoring data comprises: high-voltage data and low-voltage data.
3. The method for dynamically monitoring chronic disease health data based on machine learning according to claim 2, wherein the performing an abnormal fluctuation analysis process on each health monitoring data included in each leaf node in each target binary tree in the target binary tree set to obtain an abnormal fluctuation index corresponding to each health monitoring data included in each leaf node comprises:
performing curve fitting on high-voltage data included in all health monitoring data in the preset time period to obtain a high-voltage fluctuation curve, and performing curve fitting on low-voltage data included in all health monitoring data in the preset time period to obtain a low-voltage fluctuation curve, wherein the abscissa of the high-voltage fluctuation curve is the acquisition time, the ordinate of the high-voltage fluctuation curve is the high-voltage data, the abscissa of the low-voltage fluctuation curve is the acquisition time, and the ordinate of the low-voltage fluctuation curve is the low-voltage data;
Determining a curve segment between coordinate points corresponding to every two adjacent extremum values in the high-voltage fluctuation curve as a high-voltage sub-curve segment, and determining a curve segment between coordinate points corresponding to every two adjacent extremum values in the low-voltage fluctuation curve as a low-voltage sub-curve segment;
and determining an abnormal fluctuation index corresponding to each piece of health monitoring data included by the leaf node according to the high-pressure sub-curve segment to which the high-pressure data included by each piece of health monitoring data included by the leaf node belongs and the low-pressure sub-curve segment to which the low-pressure data included by the leaf node belongs.
4. The method for dynamically monitoring chronic disease health data based on machine learning according to claim 3, wherein the leaf nodes comprise the following formula corresponding to abnormal fluctuation indexes corresponding to health monitoring data:
；
；/>the method comprises the steps of carrying out a first treatment on the surface of the Wherein A is imj Is an abnormal fluctuation index corresponding to the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree; i is the sequence number of the target binary tree in the target binary tree set; m is the sequence number of the leaf node in the ith target binary tree; j is the serial number of the health monitoring data included in the mth leaf node; sigma (sigma) imj1 Is a high-voltage fluctuation factor corresponding to a high-voltage sub-curve segment where high-voltage data included in the j-th health monitoring data included in the m-th leaf node is located in the i-th target binary tree; sigma (sigma) 1 Is the accumulated value of the high-voltage fluctuation factors corresponding to all the high-voltage sub-curve segments; a, a imj Is the high-voltage data included in the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree; a is the accumulated value of all high-voltage data in the high-voltage sub-curve section where the high-voltage data included in the j-th health monitoring data included in the m-th leaf node in the ith target binary tree; sigma (sigma) imj2 In the ith target binary tree, the low-voltage fluctuation factor corresponding to the low-voltage sub-curve segment where the low-voltage data included in the jth health monitoring data included in the mth leaf node is located; sigma (sigma) 2 Is the accumulated value of the low-voltage fluctuation factors corresponding to all low-voltage sub-curve segments; b imj Is low-voltage data included in the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree; b is the accumulated value of all low-voltage data in the low-voltage sub-curve section where the low-voltage data included in the j-th health monitoring data included in the m-th leaf node is located in the i-th target binary tree; the absolute value function is taken; g imj1 And g imj2 In the ith target binary tree, the ordinate corresponding to the two endpoints of the high-voltage sub-curve segment where the high-voltage data included in the jth health monitoring data included in the mth leaf node are located; t is t imj1 And t imj2 In the ith target binary tree, the corresponding abscissa of two endpoints of a high-voltage sub-curve segment where the high-voltage data included in the jth health monitoring data included in the mth leaf node is located; g imj1 And G imj2 In the ith target binary tree, the ordinate corresponding to two endpoints of a low-voltage sub-curve segment where low-voltage data included in the jth health monitoring data included in the mth leaf node are located; t (T) imj1 And T imj2 In the ith target binary tree, the corresponding abscissa of two endpoints of the low-voltage sub-curve segment where the low-voltage data included in the jth health monitoring data included in the mth leaf node is located.
5. The method for dynamically monitoring chronic disease health data based on machine learning according to claim 2, wherein the determining, according to all health monitoring data in a binary tree layer where each leaf node is located, an abnormal deviation index corresponding to each health monitoring data included in each leaf node comprises:
and determining an abnormal deviation index corresponding to each piece of health monitoring data included in the leaf node according to the low-voltage data and the high-voltage data included in each piece of health monitoring data included in the leaf node and the low-voltage data and the high-voltage data included in all pieces of health monitoring data in the binary tree layer where the leaf node is located.
6. The method for dynamically monitoring chronic disease health data based on machine learning according to claim 5, wherein the leaf nodes comprise the following formula corresponding to an abnormal deviation index corresponding to health monitoring data:
θ imj =norm(a imj -B im +(b imj -H im ) A) is provided; wherein θ imj Is an abnormal deviation index corresponding to the j-th health monitoring data included in the mth leaf node in the ith target binary tree in the target binary tree set; i is the sequence number of the target binary tree in the target binary tree set; m is the sequence number of the leaf node in the ith target binary tree; j is the serial number of the health monitoring data included in the mth leaf node; norm () is a normalization function; a, a imj Is the high-voltage data included in the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree; b (B) im Is the average value of high-voltage data included in all health monitoring data in a binary tree layer where an mth leaf node in the ith target binary tree is located; b imj Is low-voltage data included in the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree; h im Is the average value of low-voltage data included in all health monitoring data in the binary tree layer where the mth leaf node is located in the ith target binary tree.
7. The method for dynamically monitoring chronic disease health data based on machine learning according to claim 1, wherein the determining the target abnormality index corresponding to each health monitoring data included in each leaf node according to the abnormality fluctuation index and the abnormality deviation index corresponding to each health monitoring data included in each leaf node and all health monitoring data in all target binary trees comprises:
determining low-pressure data and high-pressure data included in the health monitoring data as blood pressure data;
and determining the target abnormal index corresponding to each health monitoring data of the leaf node according to the abnormal fluctuation index and the abnormal deviation index corresponding to each health monitoring data of the leaf node, the average value of all blood pressure data in the target binary tree where the leaf node is located and the average value of all blood pressure data in the target binary tree.
8. The method for dynamically monitoring chronic disease health data based on machine learning according to claim 7, wherein the leaf node comprises a formula corresponding to a target abnormality index corresponding to health monitoring data:
the method comprises the steps of carrying out a first treatment on the surface of the Wherein (1)>Is a target abnormality index corresponding to the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree in the target binary tree set; i is the sequence number of the target binary tree in the target binary tree set; m is the sequence number of the leaf node in the ith target binary tree; j is the serial number of the health monitoring data included in the mth leaf node; norm () is a normalization function; θ imj Is an abnormal deviation index corresponding to the j-th health monitoring data included in the mth leaf node in the ith target binary tree in the target binary tree set; a is that imj Is the abnormal fluctuation finger corresponding to the j-th health monitoring data included in the m-th leaf node in the i-th target binary tree in the target binary tree setMarking; mu (mu) im The average value of all blood pressure data in the target binary tree where the mth leaf node in the ith target binary tree in the target binary tree set is located; μ is the average of all blood pressure data in all target binary trees.
9. The method for dynamically monitoring chronic disease health data based on machine learning according to claim 1, wherein the step of determining whether the patient to be monitored has a chronic disease abnormality to be monitored within a preset time period according to all target abnormality scores comprises:
when the target abnormality score corresponding to the health monitoring data is larger than a preset abnormality threshold, determining the health monitoring data as target abnormality data;
and when the target abnormal data exist in the preset time period, judging that the chronic disease condition abnormality to be monitored exists in the patient to be monitored in the preset time period.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311475875.2A CN117195139B (en) | 2023-11-08 | 2023-11-08 | Chronic disease health data dynamic monitoring method based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311475875.2A CN117195139B (en) | 2023-11-08 | 2023-11-08 | Chronic disease health data dynamic monitoring method based on machine learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117195139A CN117195139A (en) | 2023-12-08 |
| CN117195139B true CN117195139B (en) | 2024-02-09 |
Family
ID=88989155
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311475875.2A Active CN117195139B (en) | 2023-11-08 | 2023-11-08 | Chronic disease health data dynamic monitoring method based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117195139B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117556714B (en) * | 2024-01-12 | 2024-03-22 | 济南海德热工有限公司 | Preheating pipeline temperature data anomaly analysis method for aluminum metal smelting |
| CN117747094A (en) * | 2024-02-21 | 2024-03-22 | 天津市第五中心医院 | Abnormal detection method for gynecological disease clinical sign data |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112288025A (en) * | 2020-11-03 | 2021-01-29 | 中国平安财产保险股份有限公司 | Abnormal case identification method, device and equipment based on tree structure and storage medium |
| CN113749630A (en) * | 2021-09-16 | 2021-12-07 | 华南理工大学 | Blood pressure monitoring system and method based on ECG signal and PPG signal |
| CN115482932A (en) * | 2022-09-22 | 2022-12-16 | 深圳可孚生物科技有限公司 | Multivariate blood glucose prediction algorithm based on transfer learning and glycosylated hemoglobin |
| CN115547500A (en) * | 2022-11-03 | 2022-12-30 | 深圳市龙岗区第三人民医院 | Health monitoring equipment and system for hemodialysis patient |
| CN115577275A (en) * | 2022-11-11 | 2023-01-06 | 山东产业技术研究院智能计算研究院 | Time sequence data anomaly monitoring system and method based on LOF and isolated forest |
| CN116911806A (en) * | 2023-09-11 | 2023-10-20 | 湖北华中电力科技开发有限责任公司 | Internet + based power enterprise energy information management system |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20210174954A1 (en) * | 2019-12-04 | 2021-06-10 | Koninklijke Philips N.V. | Monitoring a patient with chronic obstructive pulmonary disease |
| US20230071467A1 (en) * | 2021-08-23 | 2023-03-09 | Analytics For Life Inc. | Method and System to Non-Invasively Assess Elevated Left Ventricular End-Diastolic Pressure |
-
2023
- 2023-11-08 CN CN202311475875.2A patent/CN117195139B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112288025A (en) * | 2020-11-03 | 2021-01-29 | 中国平安财产保险股份有限公司 | Abnormal case identification method, device and equipment based on tree structure and storage medium |
| CN113749630A (en) * | 2021-09-16 | 2021-12-07 | 华南理工大学 | Blood pressure monitoring system and method based on ECG signal and PPG signal |
| CN115482932A (en) * | 2022-09-22 | 2022-12-16 | 深圳可孚生物科技有限公司 | Multivariate blood glucose prediction algorithm based on transfer learning and glycosylated hemoglobin |
| CN115547500A (en) * | 2022-11-03 | 2022-12-30 | 深圳市龙岗区第三人民医院 | Health monitoring equipment and system for hemodialysis patient |
| CN115577275A (en) * | 2022-11-11 | 2023-01-06 | 山东产业技术研究院智能计算研究院 | Time sequence data anomaly monitoring system and method based on LOF and isolated forest |
| CN116911806A (en) * | 2023-09-11 | 2023-10-20 | 湖北华中电力科技开发有限责任公司 | Internet + based power enterprise energy information management system |
Non-Patent Citations (1)
| Title |
|---|
| 基于时间序列的风电机组异常数据检测;张添良;万方数据知识服务平台;19-20 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117195139A (en) | 2023-12-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117195139B (en) | Chronic disease health data dynamic monitoring method based on machine learning | |
| Wang et al. | A high-precision arrhythmia classification method based on dual fully connected neural network | |
| Wu et al. | Predict pneumonia with chest X-ray images based on convolutional deep neural learning networks | |
| CN110018670B (en) | Industrial process abnormal working condition prediction method based on dynamic association rule mining | |
| CN107516075B (en) | Electrocardiosignal detection method and device and electronic equipment | |
| CN109117730B (en) | Real-time electrocardiogram atrial fibrillation judgment method, device and system and storage medium | |
| CN109785976A (en) | A kind of goat based on Soft-Voting forecasting system by stages | |
| CN112294341B (en) | Sleep electroencephalogram spindle wave identification method and system based on light convolutional neural network | |
| CN117312997B (en) | Intelligent diagnosis method and system for power management system | |
| CN110634566A (en) | Traditional Chinese medicine clinical diagnosis data processing system and method and information data processing terminal | |
| CN116092680B (en) | Abdominal aortic aneurysm early prediction method and system based on random forest algorithm | |
| CN115120248B (en) | Histogram-based adaptive threshold R peak detection and heart rhythm classification method and device | |
| US20220022765A1 (en) | Heart Condition Treatment and Analysis | |
| CN113995419A (en) | Atrial fibrillation risk prediction system based on heartbeat rhythm signal and application thereof | |
| CN111128327A (en) | Hypoglycemia early warning method and device | |
| US11471113B2 (en) | Determination of health status of systems equipped with sensors | |
| Guanglong et al. | ECG signal classification algorithm based on fusion features | |
| Admass et al. | Arrhythmia classification using ECG signal: A meta-heuristic improvement of optimal weighted feature integration and attention-based hybrid deep learning model | |
| CN111317458A (en) | Blood pressure detection system based on deep learning | |
| Ammour | Atrial fibrillation detection with a domain adaptation neural network approach | |
| CN114842978B (en) | Intelligent blood gas analysis detection system and method based on medical big data | |
| CN114881941A (en) | Spine health change prediction method and system based on big data | |
| Pardasani et al. | Quantitative assessment of respiratory distress using convolutional neural network for multivariate time series segmentation | |
| CN116595338B (en) | Engineering information acquisition and processing system based on Internet of things | |
| CN116010137A (en) | Abnormality detection method based on causal relationship |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |