CN111462911A - Algorithm for predicting blood pressure abnormity of cerebral apoplexy based on big data and artificial intelligence - Google Patents
Algorithm for predicting blood pressure abnormity of cerebral apoplexy based on big data and artificial intelligence Download PDFInfo
- Publication number
- CN111462911A CN111462911A CN202010254722.5A CN202010254722A CN111462911A CN 111462911 A CN111462911 A CN 111462911A CN 202010254722 A CN202010254722 A CN 202010254722A CN 111462911 A CN111462911 A CN 111462911A
- Authority
- CN
- China
- Prior art keywords
- blood pressure
- hypotension
- algorithm
- waveform
- artificial intelligence
- 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.)
- Pending
Links
- 208000006011 Stroke Diseases 0.000 title claims abstract description 33
- 230000036772 blood pressure Effects 0.000 title claims abstract description 30
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 14
- 206010008190 Cerebrovascular accident Diseases 0.000 title abstract description 18
- 230000002490 cerebral effect Effects 0.000 title abstract description 15
- 208000001953 Hypotension Diseases 0.000 claims abstract description 55
- 230000036543 hypotension Effects 0.000 claims abstract description 40
- 230000004872 arterial blood pressure Effects 0.000 claims abstract description 28
- 238000012549 training Methods 0.000 claims abstract description 27
- 238000010801 machine learning Methods 0.000 claims abstract description 11
- 238000000034 method Methods 0.000 claims abstract description 8
- 238000012545 processing Methods 0.000 claims abstract description 4
- 238000002372 labelling Methods 0.000 claims abstract description 3
- 230000005856 abnormality Effects 0.000 claims description 9
- 238000007477 logistic regression Methods 0.000 claims description 7
- 238000010200 validation analysis Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 abstract description 11
- 230000002159 abnormal effect Effects 0.000 abstract description 6
- 230000008859 change Effects 0.000 abstract description 5
- 238000009448 modified atmosphere packaging Methods 0.000 description 22
- 208000021822 hypotensive Diseases 0.000 description 5
- 230000001077 hypotensive effect Effects 0.000 description 5
- 239000013598 vector Substances 0.000 description 5
- 206010020772 Hypertension Diseases 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 206010008118 cerebral infarction Diseases 0.000 description 3
- 208000026106 cerebrovascular disease Diseases 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 230000000004 hemodynamic effect Effects 0.000 description 3
- 230000003993 interaction Effects 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 230000001154 acute effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000000747 cardiac effect Effects 0.000 description 2
- 230000001447 compensatory effect Effects 0.000 description 2
- 238000002790 cross-validation Methods 0.000 description 2
- 230000006837 decompression Effects 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 201000010099 disease Diseases 0.000 description 2
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 2
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 2
- 239000010931 gold Substances 0.000 description 2
- 229910052737 gold Inorganic materials 0.000 description 2
- 230000001631 hypertensive effect Effects 0.000 description 2
- 230000036316 preload Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 208000024891 symptom Diseases 0.000 description 2
- 208000009304 Acute Kidney Injury Diseases 0.000 description 1
- 201000006474 Brain Ischemia Diseases 0.000 description 1
- 206010008111 Cerebral haemorrhage Diseases 0.000 description 1
- 206010008120 Cerebral ischaemia Diseases 0.000 description 1
- 206010008132 Cerebral thrombosis Diseases 0.000 description 1
- 101000969594 Homo sapiens Modulator of apoptosis 1 Proteins 0.000 description 1
- 201000001429 Intracranial Thrombosis Diseases 0.000 description 1
- 102000004866 Microtubule-associated protein 1B Human genes 0.000 description 1
- 108090001040 Microtubule-associated protein 1B Proteins 0.000 description 1
- 102100021440 Modulator of apoptosis 1 Human genes 0.000 description 1
- 101100131116 Oryza sativa subsp. japonica MPK3 gene Proteins 0.000 description 1
- 208000033626 Renal failure acute Diseases 0.000 description 1
- 101100456045 Schizosaccharomyces pombe (strain 972 / ATCC 24843) map3 gene Proteins 0.000 description 1
- 208000032851 Subarachnoid Hemorrhage Diseases 0.000 description 1
- 201000011040 acute kidney failure Diseases 0.000 description 1
- 208000037892 acute myocardial injury Diseases 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000003542 behavioural effect Effects 0.000 description 1
- 239000008280 blood Substances 0.000 description 1
- 210000004369 blood Anatomy 0.000 description 1
- 230000017531 blood circulation Effects 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 230000002526 effect on cardiovascular system Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 208000035474 group of disease Diseases 0.000 description 1
- 208000037907 haemorrhagic injury Diseases 0.000 description 1
- 230000002008 hemorrhagic effect Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000000302 ischemic effect Effects 0.000 description 1
- 230000002045 lasting effect Effects 0.000 description 1
- 231100000518 lethal Toxicity 0.000 description 1
- 230000001665 lethal effect Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 235000019837 monoammonium phosphate Nutrition 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 230000009022 nonlinear effect Effects 0.000 description 1
- 230000035479 physiological effects, processes and functions Effects 0.000 description 1
- 230000006461 physiological response Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002537 thrombolytic effect Effects 0.000 description 1
- 230000002861 ventricular Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/50—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for simulation or modelling of medical disorders
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Data Mining & Analysis (AREA)
- Biomedical Technology (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
Abstract
The invention discloses an algorithm for predicting abnormal blood pressure caused by cerebral apoplexy based on big data and artificial intelligence, relating to the technical field of blood pressure prediction and comprising the following steps: preparing a clinical database of available data containing arterial pressure waveforms for training; step (2): labeling the periods of hypotension and non-hypotension in the database as a training data set; and (3): processing the arterial pressure waveform to extract waveform features; and (4): the waveform features are mapped to predict hypotension events using the training data. The method is based on the high-fidelity arterial pressure waveform, predicts the blood pressure through a machine learning algorithm, applies machine learning to the arterial blood pressure waveform, creates an algorithm for predicting the blood pressure, accurately predicts the blood pressure change, provides an early warning function under the condition of abnormal blood pressure, and effectively prevents the occurrence of cerebral apoplexy.
Description
Technical Field
The invention relates to the technical field of blood pressure prediction, in particular to an algorithm for predicting abnormal blood pressure caused by cerebral apoplexy based on big data and artificial intelligence.
Background
On average, one person dies every 16 seconds and the age of stroke is advanced, and stroke is also one of the important causes of death in the young and middle aged people between 15 and 49 years old. Cerebral apoplexy is a group of diseases which take cerebral ischemia and hemorrhagic injury symptoms as main clinical manifestations, is also called cerebral apoplexy or cerebrovascular accident, has extremely high fatality rate and disability rate, is mainly divided into hemorrhagic cerebral apoplexy (cerebral hemorrhage or subarachnoid hemorrhage) and ischemic cerebral apoplexy (cerebral infarction and cerebral thrombosis), and is most common with cerebral infarction. The cerebral apoplexy is acute and the fatality rate is high, which is one of the most important lethal diseases in the world.
Cerebral stroke or stroke is a disease characterized by a loss of local nerve function caused by a disturbance in the blood circulation in the brain. Hypertension is one of the causes of cerebral stroke.
The gold rescue time of the cerebral apoplexy is 3.5 hours, and the expert introduces the time, and the rescue of the cerebral apoplexy is to race at the following time, compete for minutes and take seconds, and the earlier the rescue is better. In the treatment of the apoplexy, the clinical gold rescue time about the apoplexy is available. After the stroke happens, the patient needs to be sent to a hospital for thrombolysis emergency treatment within 3.5 hours, and the time is not more than 6 hours at the latest, otherwise, the disability rate and the fatality rate are increased by times.
Many wearable devices are available to detect blood pressure, but do not have a reliable blood pressure warning function. Because the traditional statistical model cannot capture the heterogeneity of the hypertensive patient, especially in the case of poor control of hypertension. There was no integration of genetic, behavioral and environmental factors in clinical trials. This results in the inability to capture complex biological changes and consistent physiological output of hypertensive patients.
Disclosure of Invention
The invention provides an algorithm for predicting abnormal blood pressure caused by cerebral apoplexy based on big data and artificial intelligence, wherein a model is built according to the blood pressure data of each person, an accurate algorithm model aiming at the individual user is obtained, the blood pressure change is accurately predicted, an early warning function is provided under the condition of abnormal blood pressure, and the cerebral apoplexy is effectively prevented.
The technical scheme of the invention is realized as follows:
an algorithm for predicting blood pressure abnormalities leading to cerebral stroke based on big data and artificial intelligence, comprising the steps of:
step (1): preparing a clinical database of available data containing arterial pressure waveforms for training;
step (2): labeling the periods of hypotension and non-hypotension in the database as a training data set;
and (3): processing the arterial pressure waveform to extract waveform features;
and (4): the waveform features are mapped to predict hypotension events using the training data.
Preferably, in step (1), the available clinical database includes: a retrospective cohort for training consisting of several patient recordings and several minutes of arterial waveform recordings and several hypotensive episodes; prospective hospital cohort for external validation, including recordings of several patients and recordings of arterial waveforms for several minutes and several hypotensive episodes.
Preferably, in step (2), mean arterial pressure <65mmHg is defined as definitive hypotension, mean arterial pressure >75mmHg is defined as definitive non-hypotension, and early recognition of hypotension is defined as 15 minutes before the actual event where mean arterial pressure is below 65mmHg for at least 1 minute.
Preferably, in the step (3), several combined features are extracted from the arterial pressure waveform of the training data set, and the data matrix of these features is respectively positive and negative for model training.
Preferably, the step of selecting a plurality of the combination features is:
step (31) for positive and negative data segments of the training data set, retaining the feature that the area under the curve is greater than 0.8;
step (32) selects successive forward features by logistic regression.
Preferably, in the step (4), the arterial pressure waveform feature is mapped to a prediction of hypotension using machine learning, and the prediction from the logistic regression model ranging from 0 to 1 is multiplied by 100 to be scaled, thereby obtaining a hypotension prediction index.
The invention has the beneficial effects that: based on the high-fidelity arterial pressure waveform, the blood pressure is predicted through a machine learning algorithm, the machine learning is applied to the arterial blood pressure waveform, an algorithm for predicting the blood pressure is established, the blood pressure change is accurately predicted, an early warning function is provided under the condition of abnormal blood pressure, and the occurrence of cerebral apoplexy is effectively prevented.
Drawings
FIG. 1 is a flow chart of an algorithm for predicting blood pressure abnormalities that cause cerebral stroke based on big data and artificial intelligence in accordance with the present invention.
Detailed Description
To further illustrate the various embodiments, the invention provides the accompanying drawings. The accompanying drawings, which are incorporated in and constitute a part of this disclosure, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the embodiments. Those skilled in the art will appreciate still other possible embodiments and advantages of the present invention with reference to these figures. Elements in the figures are not drawn to scale and like reference numerals are generally used to indicate like elements.
The invention will now be further described with reference to the accompanying drawings and detailed description.
As shown in fig. 1, an algorithm for predicting blood pressure abnormalities that cause cerebral stroke based on big data and artificial intelligence is developed using two different data sources: (1) a retrospective cohort for training consisting of 1,334 patient records and 545,959 minute arterial waveform records and 25,461 hypotension episodes; (2) a prospective local hospital cohort was used for external validation, including recordings of 204 patients and 33,236 minutes of arterial waveform recordings and 1,923 hypotensive episodes.
Mean arterial pressure (hereinafter referred to as MAP) is directly calculated from the arterial pressure waveform data. Hypotension is defined as any period of MAP <65mmHg lasting at least 1 minute, which according to studies indicates that MAP <65mmHg is the threshold for increased likelihood of acute kidney injury and myocardial injury. 1MAP >75mmHg is considered hypotension. Clearly, the definition of hypotension in the real world cannot be based on purely binary, all-or-nothing thresholds. The inflection point where the incidence of complications is considered to be increased is therefore in the MAP range between 65 and 75mmHg, which may be considered as a "grey zone" where ambiguity and some risk coexist. For accuracy, models were built based only on deterministic hypotension (MAP <65mmHg) and deterministic non-hypotension (MAP >75mmHg) data.
To eliminate the effect of sudden pressure drops due to artifacts or external events, rather than due to the patient's own physiological response, the segment of depressurization data where the MAP drop rate exceeds 0.5mmHg/s is excluded from the analysis. A rate of decline greater than 0.5mmHg/s equals a MAP decline greater than 30mmHg within 1 minute, which is considered to be beyond the predictive range of the algorithm as it is more likely to be associated with an acute event (e.g., sudden blood loss or transducer altitude change) rather than an ongoing hypotensive episode.
To ensure algorithm-related clarity, the early identification period of hypotension is defined as 15 minutes before the actual event where the MAP is below 65mmHg for at least 1 minute. For comparison, it was also evaluated whether hypotension events could be predicted by the percent change in MAP (Δ MAP). Four different Δ MAPs were calculated and evaluated: Δ MAP20s, Δ MAP1 min, Δ MAP3 min and Δ MAP5 min (the difference between the two MAP values is 20s, 1,3 and 5 min apart).
The hypotensive events were calculated by determining sections of at least 1 minute in duration such that all data points in the section showed MAP <65 mmHg. One event or positive data point is selected as the sample recorded 5, 10 or 15 minutes before the onset of the hypotensive event. The non-hypotensive events were calculated by determining a continuous portion of data points for 30 minutes such that the portion was at least 20 minutes from any hypotensive event, and all data points in the portion showed a MAP >75 mmHg. The non-event or negative data point is the center point of the non-buck event.
2,603,125 combined features are extracted from the arterial pressure waveforms of the training data set for model training, the data matrices for these features being positive and negative, respectively. The training is repeated multiple times using different patient subsets in the training dataset and using different definitions for positive and negative data points. The performance of each model after feature selection and training was evaluated using a cross-validation dataset. The final model is selected according to performance based on the prediction error of the cross-validation dataset and the general behavior of each patient. In order to retain only the most useful features, these features undergo a two-step feature selection process: (1) for positive and negative data segments of the training data set, retaining the feature that the area under the curve is greater than 0.8; (2) successive forward features are selected by logistic regression.
A L logistic regression is a classification method that predicts binary responses based on one or more model input features.
Where "x" is the independent variable (feature) vector, "ω" is the corresponding vector of coefficients, and h ω (x) is the logical model of the dependent variable. Solving the logistic function will yield:
the logical function is a continuous function in the range of 0, 1. During model training, an optimal coefficient vector "ω" (0 — no event, 1 — event) is calculated from a set of model input feature vectors "xi" and corresponding observation classes of the training set "yi", i — 1, 2, …, N, and the log-likelihood cost function is as follows:
the solution is obtained by minimizing the above constraints.
Vex cost function with respect to coefficient vector "ω". Once "ω" is determined using available training data, a logical model is applied to the new data (xt) to compute a prediction of the event "p (xt)".
The predictions produced by the logistic regression model ranging from 0 to 1 are then scaled by multiplying by 100. We named the prediction index for hypotension.
Receiver operating characteristic analysis captures false negatives, or occurs when the algorithm erroneously drops before hypotension, including a clear hypotension region with MAP <65mmHg and a critical hypotension gray region of 65 to 75 mmHg. Receiver operational characteristic analysis also captures false positives, or those that occur when the algorithm erroneously improves algorithm efficiency, hypotension does not occur, and the MAP is above 75 mmHg. In receiver operating characteristic analysis where negative plate is selected as the data point for MAP >75mmHg, the only limitation is that the receiver operating characteristic analysis may not show the effect of false positives within the 65 to 75mmHg boundary range. Clinically, MAP remains an important intermediate region between 65 and 75mmHg, which may still present a risk of complications, and false positives may be beneficial if they suggest more concern to the hemodynamic characteristics of the patient. Hypotension predictive analysis (algorithm output and hypotension analysis frequency) compared to the actual occurrence of hypotension events. In this analysis, the frequency of the occurrence of a voltage drop event in data samples in different ranges of the algorithm output is plotted. The analysis was as follows:
(1) blood pressure onset is defined as MAP <65mmHg for at least 1 minute;
(2) the event sample was taken for exactly "t" minutes (t ═ 5, 10, 15 minutes before the onset of hypotensive episodes;
(3) a non-hypotensive episode with MAP >75mmHg is at least 20 minutes from any hypotensive episode;
(4) non-event samples are taken as the midpoint of each 30 minute non-hypotensive episode;
(5) for a given data set, accumulating and segmenting all algorithm output values of the event and non-event samples into algorithm output bins;
(6) for each bin, the percentage of event samples in that bin is the event incidence, since one event in the event samples occurs within "t" minutes.
The algorithm of the embodiment is developed on the basis of a prediction algorithm of a machine learning model, and the algorithm can predict the hypotension in real time. The algorithm output indicates the likelihood that the patient's condition will be prone to a depressurization event. When the algorithm output is low, the likelihood of a voltage drop event is also low and the event interval tends to be long. Conversely, when the algorithm output is higher, the likelihood of a voltage drop event is higher, and the event interval time tends to be shorter.
The algorithm is based on detecting physiological signals in high resolution arterial pressure waveforms that are affected by attenuation of cardiovascular compensatory mechanisms that typically occur prior to hypotension.
Cardiac preload, afterload and contractile force. The early stages of instability appear to manifest as subtle and complex changes in the correlation between different physiological variables. The variability, complexity and physiologically relevant dynamics of the arterial pressure waveform characteristics occur before significant clinical symptoms occur. Decompression event the basic focus of the algorithm is to detect the earliest occurrence of these dynamic changes in the arterial pressure waveform and use them to predict an impending decompression event. Specifically, the algorithm detects dynamic changes corresponding to physiological interactions between left ventricular contractility, preload, and afterload. The main challenge in detecting these complex changes is that they are highly diverse. They are not only unrecognizable to the human eye, but also undetectable by simple signal processing algorithms. To detect the multivariate variability and interactions prior to a depressurization event, the algorithm uses complex machine learning techniques.
The machine learning method is a powerful mathematical tool and can accurately quantify dynamic multivariate interconnection. It is emphasized that in an algorithm, machine learning techniques can mathematically quantify the complex course of the cardiac compensation mechanism; they do not capture statistical relationships as most machine learning based algorithms used in the past. The evaluation of physiological associations is crucial to this algorithm as it represents the effect of dynamic links between thousands of automatically derived hemodynamics. Function, all from arterial waveforms. The assessment of the physiological link involves computing linear and nonlinear combinations of all 3,022 variability/complexity features in the initially calculated arterial waveform. These combined features then provide key information about nonlinear effects and dynamic physiology.
All 3,022 of the respective linear features interact. As arterial blood pressure drops to MAP levels of 65 to 70mmHg, a 1: 1 is linearly related. The algorithm output then rises sharply as the hypotensive event is approached, apparently due to the increase in interconnection characteristics detected by the model in the arterial pressure waveform when hypotension is imminent. The algorithm uses this large comprehensive analysis of interaction effects to evaluate compensatory mechanisms and capture cross-correlation changes between thousands of automatically derived hemodynamic characteristics, all from arterial waveforms, that are predictive of the onset of hypotension.
In addition to the potential clinical value of this development, analysis of physiological waveforms can be enhanced by using computer science techniques and reverse engineering to open up new areas of research for basic physiological studies. While this may be an important step forward, there are still many unsolved problems with using real-time prediction algorithms in a surgical environment relative to an intensive care unit or medical facility. These problems are not specific to the development of hypotension prediction algorithms, and can be applied to any prediction algorithm in the context of rapid development of invasive processes.
The present invention is not limited to the above preferred embodiments, and any modifications, equivalent substitutions, improvements, etc. within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (6)
1. An algorithm for predicting blood pressure abnormalities that cause cerebral stroke based on big data and artificial intelligence, characterized by: the method comprises the following steps:
step (1): preparing a clinical database of available data containing arterial pressure waveforms for training;
step (2): labeling the periods of hypotension and non-hypotension in the database as a training data set;
and (3): processing the arterial pressure waveform to extract waveform features;
and (4): the waveform features are mapped to predict hypotension events using the training data.
2. The algorithm for predicting blood pressure abnormalities leading to cerebral stroke based on big data and artificial intelligence as set forth in claim 1, wherein: in the step (1), the available clinical database includes: a retrospective cohort for training consisting of several patient recordings and several minutes of arterial waveform recordings and several hypotensive episodes; prospective hospital cohort for external validation, including recordings of several patients and recordings of arterial waveforms for several minutes and several hypotensive episodes.
3. The algorithm for predicting blood pressure abnormalities leading to cerebral stroke based on big data and artificial intelligence as set forth in claim 1, wherein: in said step (2), mean arterial pressure <65mmHg is defined as definitive hypotension, mean arterial pressure >75mmHg is defined as definitive non-hypotension, and early identification period of hypotension is defined as 15 minutes before the actual event that mean arterial pressure is below 65mmHg for at least 1 minute.
4. The algorithm for predicting blood pressure abnormalities leading to cerebral stroke based on big data and artificial intelligence as set forth in claim 1, wherein: in the step (3), a plurality of combined features are extracted from the arterial pressure waveform of the training data set, and for model training, the data matrixes of the features are respectively positive values and negative values.
5. The algorithm for predicting blood pressure abnormalities leading to cerebral stroke based on big data and artificial intelligence as set forth in claim 4, wherein: the selection step of the multiple combination features comprises the following steps:
step (31) for positive and negative data segments of the training data set, retaining the feature that the area under the curve is greater than 0.8;
step (32) selects successive forward features by logistic regression.
6. The algorithm for predicting blood pressure abnormalities leading to cerebral stroke based on big data and artificial intelligence as set forth in claim 4, wherein: in the step (4), the arterial pressure waveform feature is mapped to the prediction of hypotension using machine learning, and the prediction in the range from 0 to 1 generated by the logistic regression model is multiplied by 100 to be scaled, thereby obtaining the hypotension prediction index.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010254722.5A CN111462911A (en) | 2020-04-02 | 2020-04-02 | Algorithm for predicting blood pressure abnormity of cerebral apoplexy based on big data and artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010254722.5A CN111462911A (en) | 2020-04-02 | 2020-04-02 | Algorithm for predicting blood pressure abnormity of cerebral apoplexy based on big data and artificial intelligence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111462911A true CN111462911A (en) | 2020-07-28 |
Family
ID=71682326
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010254722.5A Pending CN111462911A (en) | 2020-04-02 | 2020-04-02 | Algorithm for predicting blood pressure abnormity of cerebral apoplexy based on big data and artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111462911A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112382394A (en) * | 2020-11-05 | 2021-02-19 | 苏州麦迪斯顿医疗科技股份有限公司 | Event processing method and device, electronic equipment and storage medium |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110245631A1 (en) * | 2010-03-31 | 2011-10-06 | Sahika Genc | System, Method, and Computer Software Code for Predicting an Acute Hypotensive Episode |
| CN104765846A (en) * | 2015-04-17 | 2015-07-08 | 西安电子科技大学 | Data feature classifying method based on feature extraction algorithm |
| US20150297143A1 (en) * | 2014-04-16 | 2015-10-22 | Xerox Corporation | Assessing patient risk for an acute hypotensive episode |
| CN108378835A (en) * | 2018-02-11 | 2018-08-10 | 中国联合网络通信集团有限公司 | Blood pressure prediction technique and device |
| CN109195508A (en) * | 2016-07-22 | 2019-01-11 | 爱德华兹生命科学公司 | The derived prediction to the following low blood pressure of mean arterial pressure (MAP) |
| CN109195507A (en) * | 2016-07-08 | 2019-01-11 | 爱德华兹生命科学公司 | Low blood pressure analyzes the prediction weighting of parameter |
| CN109313732A (en) * | 2016-07-22 | 2019-02-05 | 爱德华兹生命科学公司 | Forecasting risk model optimization |
| WO2019215742A1 (en) * | 2018-05-09 | 2019-11-14 | B.G. Negev Technologies And Applications Ltd., At Ben-Gurion University | System and method for hypotensive episode prediction |
-
2020
- 2020-04-02 CN CN202010254722.5A patent/CN111462911A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110245631A1 (en) * | 2010-03-31 | 2011-10-06 | Sahika Genc | System, Method, and Computer Software Code for Predicting an Acute Hypotensive Episode |
| US20150297143A1 (en) * | 2014-04-16 | 2015-10-22 | Xerox Corporation | Assessing patient risk for an acute hypotensive episode |
| CN104765846A (en) * | 2015-04-17 | 2015-07-08 | 西安电子科技大学 | Data feature classifying method based on feature extraction algorithm |
| CN109195507A (en) * | 2016-07-08 | 2019-01-11 | 爱德华兹生命科学公司 | Low blood pressure analyzes the prediction weighting of parameter |
| CN109195508A (en) * | 2016-07-22 | 2019-01-11 | 爱德华兹生命科学公司 | The derived prediction to the following low blood pressure of mean arterial pressure (MAP) |
| CN109313732A (en) * | 2016-07-22 | 2019-02-05 | 爱德华兹生命科学公司 | Forecasting risk model optimization |
| CN108378835A (en) * | 2018-02-11 | 2018-08-10 | 中国联合网络通信集团有限公司 | Blood pressure prediction technique and device |
| WO2019215742A1 (en) * | 2018-05-09 | 2019-11-14 | B.G. Negev Technologies And Applications Ltd., At Ben-Gurion University | System and method for hypotensive episode prediction |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112382394A (en) * | 2020-11-05 | 2021-02-19 | 苏州麦迪斯顿医疗科技股份有限公司 | Event processing method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Fergus et al. | Classification of caesarean section and normal vaginal deliveries using foetal heart rate signals and advanced machine learning algorithms | |
| EP3777674A1 (en) | Time series data learning and analyzing method using artificial intelligence | |
| Cömert et al. | Evaluation of fetal distress diagnosis during delivery stages based on linear and nonlinear features of fetal heart rate for neural network community | |
| Fiterau et al. | Shortfuse: Biomedical time series representations in the presence of structured information | |
| Moridani et al. | New algorithm of mortality risk prediction for cardiovascular patients admitted in intensive care unit | |
| Kumar et al. | Hybrid Bijective soft set-Neural network for ECG arrhythmia classification | |
| Banerjee et al. | A hybrid CNN-LSTM architecture for detection of coronary artery disease from ECG | |
| Arunachalam | Cardiovascular disease prediction model using machine learning algorithms | |
| Prakash et al. | A system for automatic cardiac arrhythmia recognition using electrocardiogram signal | |
| Côté-Allard et al. | Long–short ensemble network for bipolar manic-euthymic state recognition based on wrist-worn sensors | |
| Akdağ et al. | A new approach for congestive heart failure and arrhythmia classification using downsampling local binary patterns with LSTM | |
| Marathe et al. | Prediction of heart disease and diabetes using naive Bayes algorithm | |
| Manikandababu et al. | Prediction of diabetes using machine learning | |
| CN111462911A (en) | Algorithm for predicting blood pressure abnormity of cerebral apoplexy based on big data and artificial intelligence | |
| Karthikeyan et al. | An Efficient Method for Heart Disease Prediction Using Hybrid Classifier Model in Machine Learning | |
| Viswapriya et al. | A systematic method of stroke prediction model based on big data and machine learning | |
| Hong et al. | Gated temporal convolutional neural network and expert features for diagnosing and explaining physiological time series: a case study on heart rates | |
| Patil et al. | Prediction and analysis of heart disease using SVM algorithm | |
| Bahuguna et al. | Statistical Analysis and Prediction of Heart Disease Using Machine Learning | |
| Thilagavathi et al. | Heart disease prediction using machine learning algorithms | |
| Emuoyibofarhe et al. | Predictive system for heart disease using a machine learning trained model | |
| Das et al. | Accurate recognition of coronary artery disease by applying machine learning classifiers | |
| Khalil et al. | Classification of ECG signals using Legendre moments | |
| Kavya et al. | Heart Disease Prediction Using Logistic Regression | |
| Ngong et al. | Feature Extraction Methods for Predicting the Prevalence of Heart Disease |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200728 |