CN114548259B - PISA fault identification method based on Semi-supervised Semi-KNN model - Google Patents
PISA fault identification method based on Semi-supervised Semi-KNN model Download PDFInfo
- Publication number
- CN114548259B CN114548259B CN202210152424.4A CN202210152424A CN114548259B CN 114548259 B CN114548259 B CN 114548259B CN 202210152424 A CN202210152424 A CN 202210152424A CN 114548259 B CN114548259 B CN 114548259B
- Authority
- CN
- China
- Prior art keywords
- blood glucose
- semi
- pisa
- constraint
- sequence
- 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
- 238000000034 method Methods 0.000 title claims abstract description 51
- 241000404883 Pisa Species 0.000 title claims abstract 32
- 239000008280 blood Substances 0.000 claims abstract description 205
- 210000004369 blood Anatomy 0.000 claims abstract description 205
- WQZGKKKJIJFFOK-GASJEMHNSA-N Glucose Natural products OC[C@H]1OC(O)[C@H](O)[C@@H](O)[C@@H]1O WQZGKKKJIJFFOK-GASJEMHNSA-N 0.000 claims abstract description 137
- 239000008103 glucose Substances 0.000 claims abstract description 137
- 238000012549 training Methods 0.000 claims abstract description 54
- 230000002159 abnormal effect Effects 0.000 claims abstract description 40
- 238000001514 detection method Methods 0.000 claims abstract description 22
- 238000007781 pre-processing Methods 0.000 claims abstract description 17
- 238000005259 measurement Methods 0.000 claims abstract description 16
- 238000012545 processing Methods 0.000 claims abstract description 14
- 125000002791 glucosyl group Chemical group C1([C@H](O)[C@@H](O)[C@H](O)[C@H](O1)CO)* 0.000 claims description 57
- 102100037651 AP-2 complex subunit sigma Human genes 0.000 claims description 16
- 101000806914 Homo sapiens AP-2 complex subunit sigma Proteins 0.000 claims description 16
- 230000005856 abnormality Effects 0.000 claims description 14
- 238000012544 monitoring process Methods 0.000 claims description 12
- 230000006870 function Effects 0.000 claims description 11
- 238000004364 calculation method Methods 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 8
- 230000008859 change Effects 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 6
- 125000004122 cyclic group Chemical group 0.000 claims description 3
- 238000003860 storage Methods 0.000 claims description 3
- 238000003745 diagnosis Methods 0.000 abstract description 6
- 238000004422 calculation algorithm Methods 0.000 description 13
- 230000008569 process Effects 0.000 description 12
- 230000000694 effects Effects 0.000 description 6
- NOESYZHRGYRDHS-UHFFFAOYSA-N insulin Chemical compound N1C(=O)C(NC(=O)C(CCC(N)=O)NC(=O)C(CCC(O)=O)NC(=O)C(C(C)C)NC(=O)C(NC(=O)CN)C(C)CC)CSSCC(C(NC(CO)C(=O)NC(CC(C)C)C(=O)NC(CC=2C=CC(O)=CC=2)C(=O)NC(CCC(N)=O)C(=O)NC(CC(C)C)C(=O)NC(CCC(O)=O)C(=O)NC(CC(N)=O)C(=O)NC(CC=2C=CC(O)=CC=2)C(=O)NC(CSSCC(NC(=O)C(C(C)C)NC(=O)C(CC(C)C)NC(=O)C(CC=2C=CC(O)=CC=2)NC(=O)C(CC(C)C)NC(=O)C(C)NC(=O)C(CCC(O)=O)NC(=O)C(C(C)C)NC(=O)C(CC(C)C)NC(=O)C(CC=2NC=NC=2)NC(=O)C(CO)NC(=O)CNC2=O)C(=O)NCC(=O)NC(CCC(O)=O)C(=O)NC(CCCNC(N)=N)C(=O)NCC(=O)NC(CC=3C=CC=CC=3)C(=O)NC(CC=3C=CC=CC=3)C(=O)NC(CC=3C=CC(O)=CC=3)C(=O)NC(C(C)O)C(=O)N3C(CCC3)C(=O)NC(CCCCN)C(=O)NC(C)C(O)=O)C(=O)NC(CC(N)=O)C(O)=O)=O)NC(=O)C(C(C)CC)NC(=O)C(CO)NC(=O)C(C(C)O)NC(=O)C1CSSCC2NC(=O)C(CC(C)C)NC(=O)C(NC(=O)C(CCC(N)=O)NC(=O)C(CC(N)=O)NC(=O)C(NC(=O)C(N)CC=1C=CC=CC=1)C(C)C)CC1=CN=CN1 NOESYZHRGYRDHS-UHFFFAOYSA-N 0.000 description 4
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 238000003825 pressing Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000000691 measurement method Methods 0.000 description 3
- 230000035790 physiological processes and functions Effects 0.000 description 3
- 238000002360 preparation method Methods 0.000 description 3
- 208000013016 Hypoglycemia Diseases 0.000 description 2
- 102000004877 Insulin Human genes 0.000 description 2
- 108090001061 Insulin Proteins 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 230000006835 compression Effects 0.000 description 2
- 238000007906 compression Methods 0.000 description 2
- 206010012601 diabetes mellitus Diseases 0.000 description 2
- 230000002218 hypoglycaemic effect Effects 0.000 description 2
- 229940125396 insulin Drugs 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 210000003722 extracellular fluid Anatomy 0.000 description 1
- 230000005764 inhibitory process Effects 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 235000012054 meals Nutrition 0.000 description 1
- 210000000496 pancreas Anatomy 0.000 description 1
- 230000001766 physiological effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000011524 similarity measure Methods 0.000 description 1
- 238000007920 subcutaneous administration Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24133—Distances to prototypes
- G06F18/24143—Distances to neighbourhood prototypes, e.g. restricted Coulomb energy networks [RCEN]
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/145—Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissue
- A61B5/14532—Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissue for measuring glucose, e.g. by tissue impedance measurement
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Emergency Medicine (AREA)
- Optics & Photonics (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The invention relates to a PISA fault identification method based on a Semi-supervised Semi-KNN model, which comprises the following steps: s10, obtaining blood glucose information to be measured in a preset time period, and preprocessing to obtain preprocessed blood glucose information to be measured; s20, based on a pre-established PISA constraint set and preprocessed blood glucose information to be measured, obtaining a constraint relation by adopting a similarity measurement processing mode; s30, inputting the preprocessed blood glucose information to be measured and the constraint relation into a pre-trained Semi-supervised Semi-KNN model, and outputting a classification result of the blood glucose information to be measured by the Semi-supervised Semi-KNN model; the Semi-supervised Semi-KNN model is a model which is obtained by training the KNN model by adopting a training data set and a PISA constraint set and is used for identifying abnormal blood sugar information. The method improves the reliability of blood glucose information detection, improves the accuracy of fault diagnosis results, and improves the processing efficiency.
Description
Technical Field
The invention relates to a PISA fault identification technology, in particular to a PISA fault identification method based on a Semi-supervised Semi-KNN model.
Background
In recent years, continuous blood glucose monitoring systems CGM have gained increasing attention. Continuous blood glucose monitoring signals are used as an adjunct to diagnosing and guiding various types of diabetes. The continuous blood glucose monitoring signal is usually analyzed by adopting a data driving method, and the problems that the blood glucose signal is easily influenced by noise, the blood glucose monitor is easily failed, the blood glucose prediction alarm is influenced by data errors, the accuracy is not high and the like exist. Fault identification methods for continuous blood glucose monitoring systems are mostly plagued by poor performance and high false positive rates, which limit the clinical utility of assistance.
In recent years, the development of digital signal processing is rapid, and the noise problem of CGM signals is solved through a finite impulse response filter and an infinite impulse response filter. CGM fault detection remains a challenge of concern and is a very active area of research application.
When the skin around the sensor used by the continuous blood glucose monitoring system CGM is exposed to a large pressure, the CGM reading will drop rapidly and algorithms based on CGM readings such as predictive pump shut down (predictive pump shut-off) rely on an estimate of the rate of glucose sensor change to shut down the insulin pump to avoid hypoglycemia. However, the PISA event cannot be timely focused when occurring at night, which can cause improper pump shut-down; in addition, the prediction algorithm also has lower prediction data caused by PISA faults and has more serious influence on prediction and early warning. Therefore, how to distinguish the PISA fault from other events with low signal values such as insulin events, hypoglycemia events, exercise events, etc. is a technical problem that needs to be solved currently. Thus, there is a need for a semi-supervised PISA fault identification method with fast enough execution time to operate in real time.
Disclosure of Invention
First, the technical problem to be solved
In view of the above-mentioned drawbacks and shortcomings of the prior art, the present invention provides a PISA fault identification method based on a Semi-supervised Semi-KNN model.
(II) technical scheme
In order to achieve the above purpose, the main technical scheme adopted by the invention comprises the following steps:
in a first aspect, an embodiment of the present invention provides a PISA fault identification method based on a Semi-supervised Semi-KNN model, including:
s10, obtaining blood glucose information to be measured in a preset time period, and preprocessing the blood glucose information to be measured to obtain preprocessed blood glucose information to be measured;
s20, based on a pre-established PISA constraint set and the preprocessed blood glucose information to be detected, adopting a similarity measurement processing mode to acquire a constraint relation of the blood glucose information to be detected;
the PISA constraint set is a set which is constructed based on priori knowledge and has ML constraint and CL constraint in the Semi-supervised Semi-KNN model training stage, and each element in the set is information of first-order differential characteristics of a blood glucose subsequence;
s30, inputting the preprocessed blood glucose information to be measured and the constraint relation into a pre-trained Semi-supervised Semi-KNN model, and outputting a classification result of the blood glucose information to be measured by the Semi-supervised Semi-KNN model;
the Semi-supervised Semi-KNN model is a model of a Semi-supervised mode for identifying abnormal blood glucose information, which is obtained by training the KNN model by adopting a training data set and the PISA constraint set, and the training data set comprises blood glucose data processed through first-order difference.
Optionally, before S10, the method further includes:
s01, acquiring a plurality of historical blood glucose data by means of CGM equipment, preprocessing each historical blood glucose data, and obtaining a blood glucose sequence; each blood glucose sequence comprises blood glucose data with PISA timestamp labels and blood glucose data without PISA timestamp labels;
s02, dividing each blood glucose sequence into a plurality of subsequences, and performing first-order difference calculation on each subsequence to obtain a training data set;
s03, forming rules according to semi-supervision constraint conditions based on priori knowledge and training data with PISA timestamp labels in a training data set, and generating a PISA constraint set;
s04, training the Semi-supervised Semi-KNN model by using the training data set and the PISA constraint set to obtain a trained Semi-supervised Semi-KNN model;
the Semi-supervised Semi-KNN model is an improved KNN model and is constructed in a Semi-supervised mode.
Optionally, the S04 includes:
s04-1, traversing all subsequences of a training data set, and constructing an offline K-dimensional search binary tree to obtain a K-D tree;
s04-2, traversing a PISA constraint set based on the K-D tree to obtain an abnormal threshold sigma of the Semi-supervised Semi-KNN model, wherein boundary thresholds of the abnormal threshold sigma 1 and sigma 2 are expressed as sigma= [ sigma 1, sigma 2];
calculating the average distance between each PISA event and other events in the PISA constraint set by adopting a DTW similarity measurement function to obtain a distance set;
the abnormality threshold σ= [ σ1, σ2] is obtained according to the following equation (1);
σ1=q3+1.5 (Q3-Q1), equation (1)
σ2＝Q1-1.5(Q3-Q1)，
Determining that the blood glucose data to be measured has an abnormal sample when the sample distance dist < sigma 2 of the ML relation in the PISA constraint; determining that a sample distance dist > sigma 1 of the CL relation in the PISA constraint of the blood glucose data to be measured is an abnormal sample;
q3 is the upper quartile in the distance set and Q1 is the lower quartile in the distance set.
Optionally, the S02 includes:
sliding window processing is carried out on each blood sugar sequence, and the blood sugar sequence is processed in the bloodIn the sugar sequence X= { X1, X2, …, xn } a sliding window of size w forms a plurality of subsequences qi= { X i ,x i+1 ,…,x i+k }，
One subset of sequences is d= { q1, q2, …, qm }, a first order difference calculation is performed for each sub-sequence qi according to formula (2),
h is the change amount of the first-order difference formula, and the value of h is 0.8-1.2;
after the first order difference is calculated for all the sub-sequences, the first order difference value of each sub-sequence is used as a training data set.
Optionally, the S10 includes:
acquiring blood glucose information to be measured for 30-45 minutes or more by means of CGM equipment;
and performing filtering treatment, and preprocessing the blood glucose information to be detected in a sliding window mode to remove isolated noise points in the blood glucose information to be detected and realize filling of missing values, so as to obtain a blood glucose sequence to be detected of the blood glucose information to be detected. That is, the blood glucose information to be measured can be filtered, traversed by a sliding window with proper size, and processed averagely when the sliding window has such micro-area, thereby removing quasi-isolated noise points of the blood glucose information to be measured. When the missing value exists in the blood glucose information to be measured due to the sensor problem, the number of the missing values which exist continuously can be judged, and then the missing value filling is carried out on the blood glucose information to be measured through a general linear interpolation method, so that a pretreated blood glucose sequence to be measured of the blood glucose information to be measured is obtained.
Optionally, the S20 includes:
when the SBD distance between the blood glucose data A of each sequence in the blood glucose sequence to be measured and one PISA event B in the PISA constraint set is smaller than the threshold lambda, namely f SBD (A,B)<Lambda, determining a constraint relation ML (A, B), and updating the PISA constraint set; lambda is a preset value greater than 0; f represents a function of the SBD distance calculated by the two sequences;
when the SBD distance of the CL constraint relation between the blood glucose data A of each sequence in the blood glucose sequence to be measured and one PISA event B in the PISA constraint set is smaller than the threshold lambda, namely f SBD (A,B)<Lambda, determining a constraint relation CL (A, B), and updating the PISA constraint set;
traversing each sequence in the blood glucose sequences to be measured, and taking the updated PISA constraint set as the constraint relation of the blood glucose information to be measured.
Optionally, the S30 includes:
based on the K-D tree and the blood glucose sequence to be measured, acquiring the nearest K data points from each data in the blood glucose sequence to be measured in a cyclic iteration mode, obtaining a K-D tree in a using stage,
traversing the constraint relation based on the K-D tree in the using stage to obtain a classification result of PISA abnormal information of the blood glucose sequence to be detected;
and calculating the actual distance between each PISA event and other events in the constraint relation by adopting a DTW similarity measurement function, and comparing the actual distance with an abnormal threshold value sigma= [ sigma 1, sigma 2] to obtain a classification result belonging to the PISA event and the non-PISA event.
Optionally, after comparing the actual distance with the anomaly threshold, determining the data quantity belonging to the ML constraint in the constraint relation, and determining the anomaly class value in the PISA event according to the data quantity.
In a second aspect, an embodiment of the present invention further provides an electronic device, including: the system comprises a memory and a processor, wherein the memory is used for storing a computer program, and the processor is used for executing the computer program stored in the memory and executing the steps of the PISA fault identification method based on the Semi-supervised Semi-KNN model in any one of the first aspect.
In a third aspect, an embodiment of the present invention further provides a computer readable storage medium, where a computer program is stored, where the computer program, when executed by a processor, implements the steps of the PISA fault identification method based on the Semi-supervised Semi-KNN model according to any one of the first aspect above.
(III) beneficial effects
The method of the embodiment of the invention carries out abnormal classification based on the Semi-KNN model, and solves the problem of uncertainty of the KNN model when abnormal detection is carried out; the prior knowledge is introduced in a constraint form for the first time to form a semi-supervised anomaly detection method, so that the effective prior knowledge is utilized to the maximum extent, and the reliability of a detection result is improved; by grading the result, the reliability of the result is improved, and the effect of helping doctors to realize clinical judgment is achieved.
In the embodiment of the invention, the fault diagnosis of the CGM sensor is firstly carried out by a semi-supervision method, and the accuracy of the fault diagnosis result is improved by introducing priori knowledge (such as expert experience); compared with the effect of the traditional unsupervised fault recognition method applied to the field of fault diagnosis of the CGM sensor, the detection result accuracy is higher, and the PISA constraint set introduced with the semi-supervised model can ensure higher recognition rate for PISA faults and confidence level for detecting uncalibrated anomalies (such as PISA abnormal events occurring at night).
In addition, the semi-supervised model provided by the invention can update the distance measurement mode from the original Euclidean distance to the DTW and SBD similarity measurement method aiming at the time sequence data type, thereby improving the measurement accuracy of the time sequence data type and accelerating the running speed of the whole calculation program.
The method can be applied to a continuous blood glucose monitoring system CGM to enhance CGM data. The fault detection not only enhances the safety of CGM, but also can avoid the reliability reduction of tasks such as treatment scheme change or forecast and early warning caused by faults. The CGM applying the method of the invention is used for detecting the pressure-sensitive sensor attenuation (PISA) false signal, and improving the confidence of detection.
Drawings
FIG. 1 is a flowchart of a PISA fault identification method based on a Semi-supervised Semi-KNN model according to an embodiment of the invention;
FIG. 2 (a) is a schematic diagram of a process for constructing a sample of a K-dimensional search binary tree K-d tree;
FIG. 2 (b) is a schematic diagram of a K-d tree;
FIG. 3 is a representation of a new sample;
FIG. 4 is a schematic diagram of the guiding effect of the constraint relationship on the anomaly detection iterative process of KNN;
fig. 5 is a flowchart of a PISA fault identification method based on a Semi-supervised Semi-KNN model according to another embodiment of the present invention.
Detailed Description
The invention will be better explained by the following detailed description of the embodiments with reference to the drawings.
Example 1
As shown in fig. 1, fig. 1 shows a flowchart of a PISA fault identification method based on a Semi-supervised Semi-KNN model, and an execution subject of the method may be any computer/electronic device/CGM, and the method may include the following steps:
s10, obtaining blood glucose information to be measured in a preset time period, and preprocessing the blood glucose information to be measured to obtain preprocessed blood glucose information to be measured.
In the embodiment, the blood glucose information to be measured with the help of CGM can be obtained for 30-45 minutes or more; and filtering the blood glucose information to be measured, preprocessing the blood glucose information to be measured in a sliding window mode to remove isolated noise points in the blood glucose information to be measured and fill up missing values, and obtaining a blood glucose sequence to be measured of the blood glucose information to be measured.
It should be noted that, the time period of the blood glucose information to be measured can be adjusted, and the adjustment is performed according to the parameter value of the sliding window in the pretreatment.
S20, based on a pre-established PISA constraint set and the preprocessed blood glucose information to be detected, adopting a similarity measurement processing mode to acquire a constraint relation of the blood glucose information to be detected;
the PISA constraint set is a set with ML constraint and CL constraint constructed based on priori knowledge in a Semi-supervised Semi-KNN model training stage, and each element in the set is information of first-order differential characteristics of a blood glucose subsequence. The PISA constraint set must include PISA information, that is, preprocessing the blood glucose data during the training phase according to the PISA timestamp label, and then creating the PISA constraint set.
S30, inputting the preprocessed blood glucose information to be measured and the constraint relation into a pre-trained Semi-supervised Semi-KNN model, and outputting a classification result of the blood glucose information to be measured by the Semi-supervised Semi-KNN model;
the Semi-supervised Semi-KNN model is a model of a Semi-supervised mode for identifying abnormal blood glucose information, which is obtained by training the KNN model by adopting a training data set and the PISA constraint set, and the training data set comprises blood glucose data processed through first-order difference. It is understood that the training data set is a data set obtained by performing a first-order difference process on the blood glucose data obtained during the training phase through the sub-sequence set obtained after the hua process.
The method of the embodiment carries out abnormal classification based on the Semi-KNN model, and solves the problem of uncertainty of the KNN model in the process of carrying out abnormal detection; the prior knowledge is introduced in a constraint form for the first time to form a semi-supervised anomaly detection method, so that the effective prior knowledge is utilized to the maximum extent, and the reliability of a detection result is improved; by grading the result, the reliability of the result is improved, and the effect of helping doctors to realize clinical judgment is achieved.
In practical applications, before the step S10, the method shown in fig. 1 may further include the following steps not shown in the drawings:
s01, acquiring a plurality of historical blood glucose data by means of a continuous blood glucose monitoring system CGM (i.e. CGM equipment), preprocessing each historical blood glucose data, and obtaining a blood glucose sequence; each blood glucose sequence comprises blood glucose data with PISA timestamp labels and blood glucose data without PISA timestamp labels;
s02, dividing each blood glucose sequence into a plurality of subsequences, and performing first-order difference calculation on each subsequence to obtain a training data set;
for example, the S02 may include:
sliding window processing is carried out on each blood sugar sequence, and in the blood sugar sequences X= { X1, X2, … and xn }, a plurality of subsequences qi= { X are formed after a sliding window with the size of w i ,x i+1 ,…,x i+k }，
One subset of sequences is d= { q1, q2, …, qm }, a first order difference calculation is performed for each sub-sequence qi according to formula (2),
h is the change amount of the first-order difference formula, and the value of h is 0.8-1.2, preferably 1;
after the first order difference is calculated for all the sub-sequences, the first order difference value of each sub-sequence is used as a training data set.
S03, forming rules according to semi-supervision constraint conditions based on priori knowledge and training data with PISA timestamp labels in a training data set, and generating a PISA constraint set;
s04, training the Semi-supervised Semi-KNN model by using the training data set and the PISA constraint set to obtain a trained Semi-supervised Semi-KNN model;
the Semi-supervised Semi-KNN model is an improved KNN model and is constructed in a Semi-supervised mode.
For example, S04 may include:
s04-1, traversing all subsequences of a training data set, and constructing an offline K-dimensional search binary tree to obtain a K-D tree;
s04-2, traversing a PISA constraint set based on the K-D tree to obtain an abnormal threshold sigma of the Semi-supervised Semi-KNN model, wherein boundary thresholds of the abnormal threshold sigma 1 and sigma 2 are expressed as sigma= [ sigma 1, sigma 2];
calculating the average distance between each PISA event and other events in the PISA constraint set by adopting a DTW similarity measurement function to obtain a distance set;
the abnormality threshold σ= [ σ1, σ2] is obtained according to the following equation (1);
σ1=q3+1.5 (Q3-Q1), equation (1)
σ2＝Q1-1.5(Q3-Q1)，
Determining that the blood glucose data to be measured has an abnormal sample when the sample distance dist < sigma 2 of the ML relation in the PISA constraint; determining that a sample distance dist > sigma 1 of the CL relation in the PISA constraint of the blood glucose data to be measured is an abnormal sample;
q3 is the upper quartile in the distance set and Q1 is the lower quartile in the distance set.
For better understanding of the above step S20, the step S20 may be specifically described as follows:
when the SBD distance between the blood glucose data A of each sequence in the blood glucose sequence to be measured and one PISA event B in the PISA constraint set is smaller than the threshold lambda, namely f SBD (A,B)<Lambda, determining a constraint relation ML (A, B), and updating the PISA constraint set; lambda is a preset value greater than 0; f (f) SBD A function representing the calculated SBD distance of the two sequences;
when the SBD distance of the CL constraint relation between the blood glucose data A of each sequence in the blood glucose sequence to be measured and one PISA event B in the PISA constraint set is smaller than the threshold lambda, namely f SBD (A,B)<Lambda, determining a constraint relation CL (A, B), and updating the PISA constraint set;
traversing each sequence in the blood glucose sequences to be measured, and taking the updated PISA constraint set as the constraint relation of the blood glucose information to be measured.
Accordingly, the step S30 may include:
based on the K-D tree and the blood glucose sequence to be measured, acquiring the nearest K data points from each data in the blood glucose sequence to be measured in a cyclic iteration mode, obtaining a K-D tree in a using stage,
traversing the constraint relation based on the K-D tree in the using stage to obtain a classification result of PISA abnormal information of the blood glucose sequence to be detected;
and calculating the actual distance between each PISA event and other events in the constraint relation by adopting a DTW similarity measurement function, and comparing the actual distance with an abnormal threshold value sigma= [ sigma 1, sigma 2] to obtain a classification result belonging to the PISA event and the non-PISA event. Specifically, after comparing the actual distance with the anomaly threshold value, determining the data quantity belonging to the ML constraint in the constraint relation, and determining the anomaly class value belonging to the PISA event according to the data quantity.
The method of the embodiment can be integrated in electronic equipment such as an abnormality detector, and the abnormality detector can identify the abnormality problem of the PISA, so that the abnormality detector is used in clinical emergency patient care, the abnormality state of a patient is effectively monitored and reliably quantified, the hysteresis in the prior art is solved, and the real-time monitoring and analysis are realized.
In the embodiment, the fault diagnosis of the CGM sensor is firstly carried out by a semi-supervision method, and the accuracy of the fault diagnosis result is improved by introducing priori knowledge (such as expert experience). Meanwhile, the uncertainty problem of the KNN algorithm in the process of executing anomaly detection is solved.
Example two
The method of the present embodiment will be described in detail with reference to fig. 5 in the order of the preparation phase, the training phase, and the use phase.
1. Preparation phase-historical CGM blood sugar data acquisition and pretreatment
Continuous blood glucose monitoring system CGM (Continuous Glucose Monitoring) is one of the key components of an artificial pancreas by which a patient's blood glucose level can be continuously monitored to help a type one diabetes (T1 DM) patient maintain blood glucose concentration within a safe range.
1.1CGM blood sugar data acquisition
The CGM indirectly reflects the blood sugar level by monitoring the glucose concentration of subcutaneous interstitial fluid through a glucose sensor, and can provide continuous, comprehensive and reliable all-day blood sugar information. The obtained historical blood glucose data should comprise complete three-day data, wherein blood glucose values are collected every five minutes, and total blood glucose values are 3 x 288, wherein the blood glucose values should comprise a plurality of PISA typical fault information obtained by experimental pressing besides normal physiological activities such as meal, exercise, sleep and the like, and each PISA typical fault information can comprise an accurate pressing label for establishing a follow-up Semi-KNN model.
1.2CGM blood sugar data pretreatment
The historical blood glucose data acquired by the device (CGM) is stored in the storage device, and can be preprocessed in a digital signal analysis mode, including filtering, filling of missing values, marking and the like. The filtering is to remove isolated noise points on the CGM blood glucose sequence, wherein the isolated noise points mean data with overlarge blood glucose value deviation before and after the moment, the sliding window is used for traversing the sequence, when the sliding window is in a tiny area, the tiny area is processed averagely, the step is an optional step, and the filtered blood glucose sequence is smoother and is convenient for subsequent processing. The missing value filling is to prevent the situation that the blood sugar value is missing, and the blood sugar sequence is usually used after the missing value is filled, otherwise, faults of blood sugar curves in subsequent processing are easy to occur, so that the model output result is inaccurate.
In the actual processing, marking processing is needed to be carried out on the filled blood glucose sequence, the experimental interval with pressing action is marked, and the rest time is not marked, so that PISA events and other unknown events are distinguished, and therefore the accuracy of the Semi-supervised Semi-KNN model can be effectively ensured.
The above is an explanation of the preprocessing process of the blood glucose data collected by the CGM. And preprocessing the continuously acquired CGM blood glucose sequence containing the PISA compression experimental event to obtain a filtered blood glucose sequence filled with the missing value and a PISA time stamp label.
2. Preparation phase-constructing features and adding initial seeds based on a priori knowledge
As known from the PISA event time period of the experimental compression, the partial data in the blood glucose sequence may be subjected to marking processing, and the PISA timestamp label obtained by the preprocessing is, for example, 9:00-9: and in the interval 45, 9 blood glucose values in the blood glucose sequence are obtained based on the PISA timestamp label, and the corresponding first-order differential features are corresponding to PISA event features.
2.1 constructional features
In general, the blood glucose sequence has a large amount of blood glucose data, and the whole sequence cannot be expressed by using several values. Specifically, in the blood glucose sequence x= { X1, X2, …, xn } several sub-sequences qi= { X are formed after a sliding window of size w i ,x i+1 ,…,x i+k A subset of sequences is defined as d= { q1, q2, …, qm }, for each childThe sequence qi performs first order difference calculation, and a calculation formula (1) is as follows, wherein h is the change amount of the first order difference formula, and in the embodiment, the value of h is 0.8 to 1.2, preferably 1 in the blood glucose sequence characteristic structure;
and after the first-order difference is calculated on all the subsequences, the first-order difference is used as the input of a follow-up model, namely a Semi-supervised Semi-KNN model. In the embodiment, the first-order difference characteristic is adopted, so that on one hand, the influence on model universality caused by different forms of the original blood sugar sequence is avoided, and on the other hand, the first-order difference characteristic has a good inhibition effect on the fluctuation of the time sequence, and the Semi-supervised Semi-KNN model can be input more simply.
2.2 adding initial seed based on expert experience (priori knowledge)
A representative sample of several data points in the dataset of the current semi-supervised operation is called Seed, which may be represented as the sample point itself, or in a constrained form. The pairwise constraint consists of a must-link (ML) and a cannot-link (CL), two data points in the ML constraint must be in the same cluster, and two data points declared by the CL constraint must be in different clusters.
In this embodiment, a known PISA event label is defined as an initial seed of a semi-supervised model, and CL constraints are constructed by different PISA events than by blood glucose fluctuations caused by other physiological events (e.g., eating, exercise).
In this embodiment, features are constructed on the blood glucose sequence and an initial seed portion is added, that is, after preprocessing is performed on the originally acquired blood glucose sequence and the PISA timestamp label, the first-order differential features of the blood glucose sequence and the pair constraint contained in the PISA are obtained.
That is, the preprocessing described above is followed by a training data set and a PISA constraint set.
3. Training stage-training Semi-supervised Semi-KNN model
Inputting the preprocessed training data set and the pairwise constraints contained in the PISA into a Semi-KNN model to obtain a Semi-supervised fault detection model;
3.1 basic model KNN
The existing KNN is a nonparametric supervised classifier that determines the class of test samples from the majority class of the closest training samples. Taking the two classification problems under the blood glucose sequence as an example, formalized definition of the two classification problems and the solving process is as follows: the sample set of known blood glucose sequences is s= (x) 1 ,y 1 ),(x 2 ,y 2 ),…,(x N ,y N ) Wherein x is i ∈R 2 Is the point in the blood glucose measurement, y i ∈{c 1 ,c 2 And (c) indicating the category to which the blood glucose level sample belongs. For a new blood glucose sample x, the class y of the blood glucose sample can be solved by equation (2):
wherein N is K (x) Represents a set of K samples nearest to the blood glucose sample x, f being the value of y i Is a function of the indication of (2)
The problem of using supervised classification algorithms for anomaly detection is that: 1) The proportion of the abnormal sample to the normal sample is extremely unbalanced, so that the classification result is unbalanced; 2) The variety of the abnormal samples is various, the unknown abnormality and the known abnormality are classified into one type without interpretation, and the simple classification can not meet the requirements; 3) The data annotation is too costly. Thus, semi-supervised models requiring only a small amount of a priori knowledge are more suitable for the needs of the application.
3.2 Semi-KNN model
In this embodiment, the semi-supervised anomaly detection model is more suitable for PISA fault identification, and the semi-supervised anomaly detection model is based on a basic model, and the label input is changed into the partial PISA constraint set input obtained by the priori knowledge, and the data object is the preprocessed CGM blood glucose sequence data.
3.2.1 traversing all data objects of the training dataset, constructing an offline K-dimensional search binary tree of the training dataset
Through fitting samples in a training data set, a K-dimensional search binary tree (K-D tree) is constructed, the specific construction process is as follows, and simple two-dimensional sequence data is assumed: { (2, 3), (4, 7), (5, 4), (7, 2), (8, 1), (9, 6) }, first, the variance of the data in the x, y direction is calculated to know that the variance in the x direction is large, so the divided domain is determined as the x-axis direction; secondly, sorting and selecting a median value of 7 according to the value 2,5,9,4,8,7 of the x-axis direction, so that the split hyperplane is (7, 2) and is perpendicular to the x-axis; then, the left subspace and the right subspace are determined, and the split hyperplane divides the entire space into two parts, as shown in fig. 2 (a). The left subspace contains 3 nodes { (2, 3), (4, 7), (5, 4) }; the right subspace contains 2 nodes { (8, 1), (9, 6) }, and then recursion continues until each space contains only one data point, generating the final K-d tree, as shown in FIG. 2 (b).
3.2.2 traversing the PISA constraint set, determining the PISA anomaly threshold
The training data set contains PISA event data caused by experimental pressing, so that the PISA constraint set is traversed, PISA abnormalities and other normal physiological events of the training data set can be distinguished, and therefore, an abnormal threshold sigma of the Semi-supervised Semi-KNN model can be obtained, wherein the boundary threshold sigma of the abnormal threshold is sigma 1 and sigma 2, and is expressed as sigma= [ sigma 1, sigma 2].
The average distance of each PISA event from other normal physiological events is calculated, which is typically a euclidean distance, but to fit the blood glucose sequence data type, a DTW similarity metric function that distorts in the time axis to achieve better alignment is used to better express the shape similarity of the time sequence. After all average distances are obtained, the upper quartile Q3 and the lower quartile Q1 of the distance set are calculated, and then the anomaly threshold σ1=q3+1.5 (Q3-Q1), σ2=q1-1.5 (Q3-Q1) of the model, and when the distance dist > σ1 or dist < σ2 from the normal sample, the sample is considered as an anomaly sample.
That is, if the distance dist < sigma 2 of the sample of the ML relationship in the PISA constraint is the blood glucose data to be measured, determining the blood glucose data to be measured as an abnormal sample; and determining that the blood glucose data to be measured is an abnormal sample if the sample distance dist > sigma 1 of the CL relation in the PISA constraint.
4. Use phase-obtaining new to blood glucose data, performing constraint propagation
After Semi-supervised Semi-KNN model training is completed, blood glucose data collected in real time can be checked, and accurate identification can be achieved when the new blood glucose sequence contains PISA fault events (namely non-PISA events).
4.1 blood glucose data to be measured
And obtaining continuous blood glucose values to be analyzed, namely 9 blood glucose measuring points, preprocessing, such as filtering, removing isolated noise points in blood glucose information to be analyzed and filling missing values in a sliding window mode, and obtaining a blood glucose sequence to be tested of the blood glucose information to be tested. 4.2 constraint propagation
And executing a constraint propagation algorithm on the blood glucose sequence to be tested, and expanding a PISA constraint set in the original training data set, so that the judgment is assisted when the Semi-supervised Semi-KNN model is used, and the result is more accurate.
The specific implementation process is as follows:
4.2.1 traversing the PISA constraint set in the original training dataset, calculating the SBD distance
When a subset of sequences is d= { q1, q2, …, qm }, any qi in D treats the other sequences as nearest neighbors, meaning that the distance from the current sequence to the other subsequences ql is less than a threshold λ, which may be typically specified artificially, the pair-wise constraint described above contains the following properties:
1) Any q e D, where r e D, can generate ML (q, r);
2) Given ML (p, q), (q, r), ML (p, r) can be obtained;
3) Given ML (p, q), ML (c, p) and ML (p, r) can be generated, where c ε D, r ε D;
4) Given CL (p, q), CL (c, p) and CL (p, r) can be generated, where c εD, r εD.
The distance metric selected in this embodiment is the SBD distance, and the SBD algorithm is a cross-correlation based shape similarity metric, whose efficient and parameter-free characteristics are incomparable with DTW metrics, which are a measurement method with high accuracy but high computational cost. And the accuracy of the SBD algorithm is close to that of the DTW algorithm, so that the SBD can be better used for measuring the similarity between CGM curves and conveniently realize online similarity calculation.
4.2.2 performing constraint propagation
When the SBD distance between the blood glucose data A of each sequence in the blood glucose sequence to be tested and a certain PISA event B in the training data set is smaller than the threshold lambda, namely f SBD (A,B)<λ, a constraint relationship ML (a, B) can be determined, from which the constraint propagation properties can be updated.
When the SBD distance of the CL constraint relation between the blood glucose data A of each sequence in the blood glucose sequence to be tested and a certain PISA event B in the training data set is smaller than the threshold lambda, namely f SBD (A,B)<λ, a constraint relation CL (a, B) can be determined, and the constraint set can be updated by the constraint propagation property as described above.
5. Stage of use-detection of anomalies Using Semi-supervised Semi-KNN model, and judgment of anomaly class
After the step 4 is executed, inputting the blood glucose data of each sequence in the blood glucose sequence to be detected and the constraint relation expanding the PISA constraint set of the Semi-supervised Semi-KNN model, performing abnormality detection by using the Semi-supervised Semi-KNN model established in the step 3.2.2, judging the abnormality level according to the constraint relation, and specifically executing the steps as follows:
5.1 search for k nearest neighbors and calculate average distance
The blood glucose data (i.e. new blood glucose data) of each sequence in the blood glucose sequences to be tested is input into the Semi-supervised Semi-KNN model, and k data points nearest to the blood glucose data of each sequence can be obtained through the k-dimensional search binary tree established in the step 3.2.1, for example, a k-dimensional search binary tree sample established according to the step 3.2.1 is obtained, and the blood glucose data of each sequence is assumed to be (2.1, 3.1), as shown in fig. 3. Firstly, finding nearest neighbor approximate points (2, 3) through binary search, and calculating a distance 0.1414; then backtracking to the father node (5, 4), and finding out that the father node is not intersected with the hyperplane with y=4 by taking (2.1, 3.1) as a circle center and 0.1414 as a radius, so that the father node does not need to enter into a right subspace search; then trace back to the parent node (7, 2), the circle also does not intersect the hyperplane of x=7, and therefore does not need to enter the right subspace search of (7, 2); and (3) obtaining the nearest sample (2, 3) after backtracking is completed, and obtaining the nearest k sample points which are called k neighbors through loop iteration for k times. After k-nearest neighbor is obtained, calculating the average distance between the sample point and the blood glucose data of each sequence, using the distance as a score for abnormality determination, and considering the sample point as a PISA fault event when the score meets a 3.2.2 threshold.
It should be noted that, in the iteration process, when the new sample has a constraint relation with the PISA, the k neighbor sample should be satisfied and no CL relation is included, that is, if the distance between the sample encountering the CL relation and the new sample in the iteration process satisfies the requirement of k neighbor, the sample should be discarded, the iteration is continued downwards, and the guiding effect of the constraint relation on the abnormal detection iteration process of KNN can be represented by fig. 4.
5.2 outputting the abnormal grade according to the constraint relation
The step can consider the constraint relation of counting whether the k neighbor contains the PISA while outputting the abnormal detection result. That is, after obtaining the distance between the blood glucose data of each sequence in the Semi-supervised Semi-KNN model, it is necessary to consider whether there is a constraint relationship, that is, the result obtained in the constraint propagation stage described above, in addition to comparing with the abnormal threshold.
The specific operation is as follows: when no constraint relation is contained, judging as abnormal level 1; when ML relationships are included with the expanded PISA constraint set, a higher anomaly level of 2,3, … n is assigned depending on the number of ML relationships that include the expanded PISA constraint set. The more ML constraints there are in the extended PISA constraint set, the higher the anomaly level.
The DTW (Dynamic Time Warping) algorithm described above is used to detect the degree of similarity of two time series, stretching or compressing the time series to align it as much as possible. In most cases, the two sequences have very similar shapes overall, but these shapes are not aligned on the x-axis. Before comparing the similarities, one (or both) of the sequences needs to be warped in the time axis to achieve better alignment. DTW is an effective way to implement this warping.
The SBD algorithm is a cross-correlation (cross-correlation) based shape similarity measure, whose efficient and parameter-free characteristics are incomparable with DTW, which is a measurement method with high accuracy but high computational cost. And the accuracy of the SBD algorithm is close to that of the DTW algorithm, so that the SBD can be better used for measuring the similarity between CGM curves and conveniently realize online similarity calculation.
The sliding inner product between two pieces of time series data is calculated through cross correlation, and the sliding inner product has original robustness for phase offset. For a given two-bar time sequence x= (x) 1 ,x 2 ,…,x m ) And y= (y) 1 ,y 2 ,…,y m ) And given the corresponding phase difference s, the inner product of the two curves results as follows:
the standard cross-correlation NCC and the distance metric SBD can be calculated as follows:
according to the embodiment, experiments are carried out on the similarity measurement mode, and experimental results show that the SBD similarity measurement algorithm has strong anti-noise capability, so that the waveform gap between a PISA fault event and a normal sequence can be effectively distinguished, and the noise difference between other sequences can be effectively avoided; the DTW similarity measurement algorithm is sensitive to shape characteristics, and can amplify small differences of shapes after the distance output is normalized; euclidean distance is sensitive to blood glucose curve amplitude, i.e. absolute distance differences in the original sequence distribution can be fully represented.
Example III
The embodiment also provides an electronic device, including: a memory and a processor; the processor is configured to execute the computer program stored in the memory, so as to implement the steps of executing the PISA fault identification method based on the Semi-supervised Semi-KNN model according to any of the first and second embodiments.
It should be noted that in the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The invention may be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In the claims enumerating several means, several of these means may be embodied by one and the same item of hardware. The use of the terms first, second, third, etc. are for convenience of description only and do not denote any order. These terms may be understood as part of the component name.
Furthermore, it should be noted that in the description of the present specification, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to a specific feature, structure, material, or characteristic described in connection with the embodiment or example being included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms are not necessarily directed to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, the different embodiments or examples described in this specification and the features of the different embodiments or examples may be combined and combined by those skilled in the art without contradiction.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art upon learning the basic inventive concepts. Therefore, the appended claims should be construed to include preferred embodiments and all such variations and modifications as fall within the scope of the invention.
It will be apparent to those skilled in the art that various modifications and variations can be made to the present invention without departing from the spirit or scope of the invention. Thus, the present invention should also include such modifications and variations provided that they come within the scope of the following claims and their equivalents.
Claims (8)
1. The PISA fault identification method based on the Semi-supervised Semi-KNN model is characterized by comprising the following steps of:
s10, obtaining blood glucose information to be measured in a preset time period, and preprocessing the blood glucose information to be measured to obtain preprocessed blood glucose information to be measured;
s20, based on a pre-established detection pressure sensor attenuation PISA constraint set and the preprocessed blood glucose information to be detected, adopting a similarity measurement processing mode to acquire a constraint relation of the blood glucose information to be detected;
the PISA constraint set is a set which is constructed based on priori knowledge and has ML constraint and CL constraint in the Semi-supervised Semi-KNN model training stage, and each element in the set is information of first-order differential characteristics of a blood glucose subsequence;
the S20 includes: when the SBD distance between the blood glucose data A of each sequence in the blood glucose sequence to be measured and one PISA event B in the PISA constraint set is smaller than the threshold lambda, namely f SBD (A,B)<Lambda, determining a constraint relation ML (A, B), and updating the PISA constraint set; lambda is a preset value greater than 0; f (f) SBD A function representing the calculated SBD distance of the two sequences;
when each of the blood glucose sequences is to be testedWhen the SBD distance of the CL constraint relation between the blood glucose data A of the column and one PISA event B in the PISA constraint set is smaller than the threshold lambda, namely f SBD (A,B)<Lambda, determining a constraint relation CL (A, B), and updating the PISA constraint set;
traversing each sequence in the blood glucose sequence to be measured, and taking the updated PISA constraint set as a constraint relation of the blood glucose information to be measured;
s30, inputting the preprocessed blood glucose information to be measured and the constraint relation into a pre-trained Semi-supervised Semi-KNN model, and outputting a classification result of the blood glucose information to be measured by the Semi-supervised Semi-KNN model;
the Semi-supervised Semi-KNN model is a model of a Semi-supervised mode for identifying abnormal blood glucose information, which is obtained by training the KNN model by adopting a training data set and the PISA constraint set, and the training data set comprises blood glucose data processed through first-order difference;
the S30 includes: based on the K-D tree and the blood glucose sequence to be measured, the nearest K data points from each data in the blood glucose sequence to be measured are acquired in a cyclic iteration mode, the K-D tree in the using stage is obtained,
traversing the constraint relation based on the K-D tree in the using stage to obtain a classification result of PISA abnormal information of the blood glucose sequence to be detected;
calculating the actual distance between each PISA event and other events in the constraint relation by adopting a DTW similarity measurement function, and comparing the actual distance with an abnormal threshold value sigma= [ sigma 1, sigma 2] to obtain classification results belonging to the PISA event and the non-PISA event; sigma is an abnormal threshold value of the Semi-supervised Semi-KNN model, and boundary threshold values of the abnormal threshold value sigma are sigma 1 and sigma 2.
2. The method according to claim 1, wherein prior to S10, the method further comprises:
s01, acquiring a plurality of historical blood glucose data by means of a continuous blood glucose monitoring system CGM, preprocessing each historical blood glucose data, and obtaining a blood glucose sequence; each blood glucose sequence comprises blood glucose data with PISA timestamp labels and blood glucose data without PISA timestamp labels;
s02, dividing each blood glucose sequence into a plurality of subsequences, and performing first-order difference calculation on each subsequence to obtain a training data set;
s03, forming rules according to semi-supervision constraint conditions based on priori knowledge and training data with PISA timestamp labels in a training data set, and generating a PISA constraint set;
s04, training the Semi-supervised Semi-KNN model by using the training data set and the PISA constraint set to obtain a trained Semi-supervised Semi-KNN model;
the Semi-supervised Semi-KNN model is an improved KNN model and is constructed in a Semi-supervised mode.
3. The method according to claim 2, wherein S04 comprises:
s04-1, traversing all subsequences of a training data set, and constructing an offline K-dimensional search binary tree to obtain a K-D tree;
s04-2, traversing a PISA constraint set based on the K-D tree to obtain an abnormal threshold sigma of the Semi-supervised Semi-KNN model, wherein boundary thresholds of the abnormal threshold sigma 1 and sigma 2 are expressed as sigma= [ sigma 1, sigma 2];
calculating the average distance between each PISA event and other events in the PISA constraint set by adopting a DTW similarity measurement function to obtain a distance set;
the abnormality threshold σ= [ σ1, σ2] is obtained according to the following equation (1);
σ1=q3+1.5 (Q3-Q1), equation (1)
σ2＝Q1-1.5(Q3-Q1)，
Determining that the blood glucose data to be measured has an abnormal sample when the sample distance dist < sigma 2 of the ML relation in the PISA constraint; determining that a sample distance dist > sigma 1 of the CL relation in the PISA constraint of the blood glucose data to be measured is an abnormal sample;
q3 is the upper quartile in the distance set and Q1 is the lower quartile in the distance set.
4. The method according to claim 2, wherein said S02 comprises:
sliding window processing is carried out on each blood sugar sequence, and in the blood sugar sequences X= { X1, X2, … and xn }, a plurality of subsequences qi= { X are formed after a sliding window with the size of w i ,x i+1 ,…,x i+k }，
A sequence subset is D= { q1, q2, …, qm }, first-order differential calculation is carried out on each subsequence qi according to a formula (2), n represents the total length of the blood glucose sequence, i is any value from 1 to n-w, the i-th sequence is represented, and qi is the i-th sequence in the sequence subset;
h is the change amount of the first-order difference formula, and the value of h is 0.8-1.2;
after the first order difference is calculated for all the sub-sequences, the first order difference value of each sub-sequence is used as a training data set.
5. The method according to claim 1, wherein S10 comprises:
acquiring blood glucose information to be measured for 30-45 minutes or more by means of CGM;
and performing filtering treatment, and preprocessing the blood glucose information to be detected in a sliding window mode to remove isolated noise points in the blood glucose information to be detected and realize filling of missing values, so as to obtain a blood glucose sequence to be detected of the blood glucose information to be detected.
6. A method according to claim 1, characterized in that the actual distance is compared with an anomaly threshold value, the amount of data belonging to the ML constraint in the constraint relation is determined, and the anomaly class value belonging to the PISA event is determined from the amount of data.
7. An electronic device comprising a memory and a processor, the memory storing a computer program, the processor executing the computer program stored in the memory and performing the steps of the PISA fault identification method based on the Semi-supervised Semi-KNN model as set forth in any of the preceding claims 1 to 6.
8. A computer readable storage medium, characterized in that it has stored thereon a computer program which, when executed by a processor, implements the steps of the PISA fault identification method based on Semi-supervised Semi-KNN model as set forth in any of the preceding claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210152424.4A CN114548259B (en) | 2022-02-18 | 2022-02-18 | PISA fault identification method based on Semi-supervised Semi-KNN model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210152424.4A CN114548259B (en) | 2022-02-18 | 2022-02-18 | PISA fault identification method based on Semi-supervised Semi-KNN model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114548259A CN114548259A (en) | 2022-05-27 |
| CN114548259B true CN114548259B (en) | 2023-10-10 |
Family
ID=81676044
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210152424.4A Active CN114548259B (en) | 2022-02-18 | 2022-02-18 | PISA fault identification method based on Semi-supervised Semi-KNN model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114548259B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102281816A (en) * | 2008-11-20 | 2011-12-14 | 人体媒介公司 | Method and apparatus for determining critical care parameters |
| CN107133642A (en) * | 2017-04-25 | 2017-09-05 | 东北大学 | A kind of priori method for diagnosing faults based on Tennessee Yi Siman processes |
| CN110021397A (en) * | 2019-02-01 | 2019-07-16 | 捷普科技（上海）有限公司 | Method and storage medium based on human body physiological parameter prediction dosage |
| CN110379503A (en) * | 2019-07-30 | 2019-10-25 | 东北大学 | A kind of online fault detection and diagnosis system based on continuous blood sugar monitoring system |
| CN110448306A (en) * | 2019-07-30 | 2019-11-15 | 东北大学 | A kind of online fault detection and diagnosis method based on continuous blood sugar monitoring system |
| CN110689954A (en) * | 2019-10-10 | 2020-01-14 | 王月娥 | Multifunctional endocrine detector control system and control method |
| CN111027421A (en) * | 2019-11-26 | 2020-04-17 | 西安宏规电子科技有限公司 | Graph-based direct-push type semi-supervised pedestrian re-identification method |
| CN112819059A (en) * | 2021-01-26 | 2021-05-18 | 中国矿业大学 | Rolling bearing fault diagnosis method based on popular retention transfer learning |
| FR3109455A1 (en) * | 2020-04-20 | 2021-10-22 | Universite De Paris | Determination of health states of systems equipped with sensors |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI502537B (en) * | 2013-12-13 | 2015-10-01 | Ind Tech Res Inst | Management systems and methods for managing physiology data measurement |
| US20160029966A1 (en) * | 2014-07-31 | 2016-02-04 | Sano Intelligence, Inc. | Method and system for processing and analyzing analyte sensor signals |
| US11872371B2 (en) * | 2017-04-20 | 2024-01-16 | The Feinstein Institutes For Medical Research | Systems and methods for real-time monitoring of physiological biomarkers through nerve signals and uses thereof |
-
2022
- 2022-02-18 CN CN202210152424.4A patent/CN114548259B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102281816A (en) * | 2008-11-20 | 2011-12-14 | 人体媒介公司 | Method and apparatus for determining critical care parameters |
| CN107133642A (en) * | 2017-04-25 | 2017-09-05 | 东北大学 | A kind of priori method for diagnosing faults based on Tennessee Yi Siman processes |
| CN110021397A (en) * | 2019-02-01 | 2019-07-16 | 捷普科技（上海）有限公司 | Method and storage medium based on human body physiological parameter prediction dosage |
| CN110379503A (en) * | 2019-07-30 | 2019-10-25 | 东北大学 | A kind of online fault detection and diagnosis system based on continuous blood sugar monitoring system |
| CN110448306A (en) * | 2019-07-30 | 2019-11-15 | 东北大学 | A kind of online fault detection and diagnosis method based on continuous blood sugar monitoring system |
| CN110689954A (en) * | 2019-10-10 | 2020-01-14 | 王月娥 | Multifunctional endocrine detector control system and control method |
| CN111027421A (en) * | 2019-11-26 | 2020-04-17 | 西安宏规电子科技有限公司 | Graph-based direct-push type semi-supervised pedestrian re-identification method |
| FR3109455A1 (en) * | 2020-04-20 | 2021-10-22 | Universite De Paris | Determination of health states of systems equipped with sensors |
| CN112819059A (en) * | 2021-01-26 | 2021-05-18 | 中国矿业大学 | Rolling bearing fault diagnosis method based on popular retention transfer learning |
Non-Patent Citations (4)
| Title |
|---|
| Semi-Supervised Classification of Graph Convolutional Networks with Laplacian Rank Constraints;Haiqi Zhang等;Neural Processing Letters;第2645-2656页 * |
| 一种基于KNN的半监督分类改进算法;陆广泉等;广西师范大学学报(自然科学版);第30卷(第1期);第45-49页 * |
| 基于闭包准则和成对约束的半监督聚类算法;向力宏等;佛山科学技术学院学报(自然科学版)(第2期);第39-49页 * |
| 小鼠持续葡萄糖监测技术的建立及其血糖时间序列的多尺度熵分析;李成等;上海交通大学学报(医学版);第41卷(第2期);第134-139页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114548259A (en) | 2022-05-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP3258333A1 (en) | Method and system for monitoring sensor data of rotating equipment | |
| Liu et al. | Clinical time series prediction: Toward a hierarchical dynamical system framework | |
| US11692910B2 (en) | Abnormality diagnostic device, abnormality diagnostic method, and program | |
| US12119114B2 (en) | Missing medical diagnosis data imputation method and apparatus, electronic device and medium | |
| CN107137093B (en) | A kind of continuous blood sugar monitoring device comprising abnormal plasma glucose probability alarm | |
| CN111122162B (en) | Industrial system fault detection method based on Euclidean distance multi-scale fuzzy sample entropy | |
| CN111128327B (en) | Hypoglycemia early warning method and device | |
| US20210321956A1 (en) | Determination of health status of systems equipped with sensors | |
| EP3422222A1 (en) | Method and state machine system for detecting an operation status for a sensor | |
| CN112905380A (en) | System anomaly detection method based on automatic monitoring log | |
| Ou et al. | A deep sequence multi-distribution adversarial model for bearing abnormal condition detection | |
| Banaee et al. | Data-driven rule mining and representation of temporal patterns in physiological sensor data | |
| CN112949735A (en) | Liquid hazardous chemical substance volatile concentration abnormity discovery method based on outlier data mining | |
| CN118035926B (en) | Model training and water detection method and system based on multivariate data diffusion | |
| CN114004258B (en) | Semi-supervised electrocardiographic abnormality detection method | |
| Saranya et al. | Diagnosing chronic kidney disease using KNN algorithm | |
| CN114443338A (en) | Sparse negative sample-oriented anomaly detection method, model construction method and device | |
| CN114548259B (en) | PISA fault identification method based on Semi-supervised Semi-KNN model | |
| Aldi et al. | Standardscaler's Potential in Enhancing Breast Cancer Accuracy Using Machine Learning | |
| Feng et al. | Multi‐model sensor fault detection and data reconciliation: A case study with glucose concentration sensors for diabetes | |
| Ghosh et al. | Risk prediction for acute hypotensive patients by using gap constrained sequential contrast patterns | |
| Thabit et al. | Detecting diabetes using machine learning algorithms | |
| Luca et al. | Point process models for novelty detection on spatial point patterns and their extremes | |
| Wu et al. | Research on abnormal detection of one-class support vector machine based on ensemble cooperative semi-supervised learning | |
| Funkner et al. | Motif identification in vital signs of chronic patients |
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 |