CN115969464A - Piezoelectric impedance thrombolysis effect prediction method and system based on support vector machine regression - Google Patents
Piezoelectric impedance thrombolysis effect prediction method and system based on support vector machine regression Download PDFInfo
- Publication number
- CN115969464A CN115969464A CN202211678107.2A CN202211678107A CN115969464A CN 115969464 A CN115969464 A CN 115969464A CN 202211678107 A CN202211678107 A CN 202211678107A CN 115969464 A CN115969464 A CN 115969464A
- Authority
- CN
- China
- Prior art keywords
- wolf
- support vector
- target
- vector machine
- piezoelectric impedance
- 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.)
- Granted
Links
- 230000002537 thrombolytic effect Effects 0.000 title claims abstract description 105
- 238000012706 support-vector machine Methods 0.000 title claims abstract description 69
- 230000000694 effects Effects 0.000 title claims abstract description 41
- 238000000034 method Methods 0.000 title claims abstract description 33
- 241000282461 Canis lupus Species 0.000 claims abstract description 123
- 208000007536 Thrombosis Diseases 0.000 claims abstract description 63
- 238000012360 testing method Methods 0.000 claims abstract description 47
- 238000012549 training Methods 0.000 claims abstract description 41
- 239000008280 blood Substances 0.000 claims abstract description 35
- 210000004369 blood Anatomy 0.000 claims abstract description 35
- 239000003146 anticoagulant agent Substances 0.000 claims abstract description 26
- 238000012545 processing Methods 0.000 claims abstract description 17
- 230000006870 function Effects 0.000 claims description 69
- 238000003756 stirring Methods 0.000 claims description 15
- 238000010606 normalization Methods 0.000 claims description 9
- 238000005457 optimization Methods 0.000 claims description 9
- 238000004458 analytical method Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 6
- 238000003672 processing method Methods 0.000 claims description 3
- 239000000126 substance Substances 0.000 claims 1
- 229940079593 drug Drugs 0.000 description 7
- 239000003814 drug Substances 0.000 description 7
- 239000013078 crystal Substances 0.000 description 6
- 208000006011 Stroke Diseases 0.000 description 5
- 210000004204 blood vessel Anatomy 0.000 description 5
- 238000002474 experimental method Methods 0.000 description 5
- 239000003527 fibrinolytic agent Substances 0.000 description 4
- 239000013598 vector Substances 0.000 description 4
- 210000004556 brain Anatomy 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000013507 mapping Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 208000032382 Ischaemic stroke Diseases 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 2
- 230000006931 brain damage Effects 0.000 description 2
- 231100000874 brain damage Toxicity 0.000 description 2
- 208000029028 brain injury Diseases 0.000 description 2
- 230000006378 damage Effects 0.000 description 2
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 2
- 231100000518 lethal Toxicity 0.000 description 2
- 230000001665 lethal effect Effects 0.000 description 2
- 208000024891 symptom Diseases 0.000 description 2
- 206010008132 Cerebral thrombosis Diseases 0.000 description 1
- 206010018985 Haemorrhage intracranial Diseases 0.000 description 1
- 208000008574 Intracranial Hemorrhages Diseases 0.000 description 1
- 201000001429 Intracranial Thrombosis Diseases 0.000 description 1
- 241000219000 Populus Species 0.000 description 1
- 102000003990 Urokinase-type plasminogen activator Human genes 0.000 description 1
- 108090000435 Urokinase-type plasminogen activator Proteins 0.000 description 1
- 230000002490 cerebral effect Effects 0.000 description 1
- 238000011217 control strategy Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 238000004090 dissolution Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000013467 fragmentation Methods 0.000 description 1
- 238000006062 fragmentation reaction Methods 0.000 description 1
- 230000002427 irreversible effect Effects 0.000 description 1
- 230000000302 ischemic effect Effects 0.000 description 1
- 208000011977 language disease Diseases 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 235000015097 nutrients Nutrition 0.000 description 1
- 230000000704 physical effect Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000001225 therapeutic effect Effects 0.000 description 1
- 229960005356 urokinase Drugs 0.000 description 1
Images
Landscapes
- External Artificial Organs (AREA)
Abstract
The invention relates to a method and a system for predicting a piezoelectric impedance thrombolysis effect based on support vector machine regression, which comprises the following steps: testing piezoelectric impedance data corresponding to thrombolytic solutions with different concentrations by using a thrombus vibration actuator; carrying out standardization processing on the piezoelectric impedance data to obtain a training set and a test set; inputting the training set and the test set as characteristic parameters into a regression model of the support vector machine optimized by a wolf algorithm for training and testing to obtain a trained regression model of the support vector machine; predicting the blood concentration of the target to be detected based on the trained support vector machine regression model and the target piezoelectric impedance data; and predicting the thrombolysis effect of the target to be detected based on the blood concentration. The invention provides a piezoelectric impedance thrombolysis effect prediction method adopting support vector machine regression for the first time, and solves the technical problems of low reliability and low accuracy in the prior art.
Description
Technical Field
The invention relates to the technical field of thrombolysis effect prediction, in particular to a piezoelectric impedance thrombolysis effect prediction method and system based on support vector machine regression.
Background
According to Global disease burden research (GBD) data, cerebral stroke is the first three common lethal and disabling causes in the world and is also the first cause of lethal and disabling in China. The stroke of tens of millions of people all over the world is suffered every year, and a large part of the stroke belongs to ischemic stroke. The main reason for stroke in ischemic brain is cerebral thrombosis which obstructs blood vessels, reduces or cuts off nutrient supply of the brain, and causes brain damage, thereby causing language disorder, limb disorder and even death. In many cases, a stroke occurs suddenly and after a few hours it causes irreversible damage to the brain. Therefore, the thrombus should be removed and dissolved as soon as possible to reopen the occluded blood vessel to avoid brain damage.
There are two main international methods for treating ischemic stroke: drug thrombolysis and mechanical thrombus removal. The drug thrombolysis by using drugs such as rt-PA, urokinase and the like is a worldwide accepted treatment method. However, the use of drug thrombolysis has a therapeutic time window, typically 3-4.5 hours after the onset of symptoms. Most patients do not have a treatment with drug thrombolysis within a time window because the patients spend a lot of time to get a doctor and confirm the cause after the symptoms appear. Meanwhile, the drug thrombolysis has certain defects, such as long recanalization time of blood vessels, low recanalization rate, and intracranial hemorrhage easily caused by using a large amount of thrombolytic drugs. Therefore, mechanical interventional thrombus removal systems appear at home and abroad, and thrombus is grabbed at a fixed point in a blood vessel through a self-expansion spiral coil or a stent mechanism and taken out of the body. Compared with drug thrombolysis, the mechanical thrombus removal system has the advantages of longer time window, higher recanalization rate, faster recanalization time and the like. However, due to the characteristics of the self-expanding stent structure, it is prone to damage to the inner wall of the vessel and is not easily used in vessels having a small diameter.
In view of the above problems, zhongwei Jiang et al abroad proposed a method of driving mechanical vibration by a piezoelectric element to agitate and break up thrombus and accelerate the dissolution rate of the thrombus. The specific method is to inject a small amount of thrombolytic drug into the position blocked by thrombus, and the thrombus is stirred and broken by the vibration generated by the designed micro vibration actuator and is fully combined with thrombolytic drugs, so that the aim of quickly breaking and dissolving the thrombus is fulfilled. Based on this viewpoint, poplar crystal and others have designed miniature vibration actuators suitable for use in cerebral vessels.
The physical properties of thrombi are diversified due to different generation sites, formation reasons and formation time, and the required parameters of the fragmentation and thrombolysis and the dosage of thrombolytic drugs are different. In order to accelerate the thrombus clearing time and better make a control strategy of the vibration actuator, it is important to acquire the thrombolytic effect of the actuator in the blood vessel in real time. The traditional thrombolytic effect prediction method has the technical problems of low reliability and low accuracy.
Disclosure of Invention
In view of this, the present application provides a method and a system for predicting a piezo-impedance thrombolysis effect based on support vector machine regression, so as to alleviate the technical problems of low reliability and low accuracy in the prior art.
In a first aspect, an embodiment of the present invention provides a method for predicting a piezoelectric impedance thrombolysis effect based on support vector machine regression, including: testing piezoelectric impedance data corresponding to thrombolytic solutions of different concentrations by using a thrombus vibration actuator; carrying out standardization processing on the piezoelectric impedance data to obtain a training set and a test set; inputting the training set and the test set as characteristic parameters into a regression model of a support vector machine optimized by a wolf algorithm for training and testing to obtain a trained regression model of the support vector machine; predicting the blood concentration of the target to be measured based on the trained support vector machine regression model and the target piezoelectric impedance data; the target piezoelectric impedance data is blood piezoelectric impedance data acquired by the thrombus vibration actuator on the target to be detected; and predicting the thrombolysis effect of the target to be detected based on the blood concentration.
Further, the piezoelectric impedance data corresponding to the thrombolytic solutions with different concentrations are tested by using a thrombus vibration actuator, and the method comprises the following steps: mixing and simulating thrombolytic solutions with different concentrations, and inserting the thrombolytic solution into the thrombus vibration actuator; inputting a sine driving signal with a preset resonance frequency through a function generator and a power amplifier of the thrombus vibration actuator, so that the thrombus vibration actuator generates vibration and fully stirs the thrombolytic solutions with different concentrations; and measuring real part piezoelectric impedance data of the thrombolysis solutions with different concentrations after stirring by using an impedance analyzer of the thrombus vibration actuator to obtain piezoelectric impedance data corresponding to the thrombolysis solutions with different concentrations.
Further, the normalizing the piezoelectric impedance data includes: and carrying out standardization processing on the piezoelectric impedance data based on a mean variance normalization method.
Further, the training set and the testing set are used as characteristic parameters and input into a regression model of a support vector machine optimized by a wolf algorithm for training and testing, and the method comprises the following steps: performing mathematical analysis on the piezoelectric impedance data to construct a regression model of a support vector machine; establishing a gray wolf optimization algorithm and searching the optimal value of the regression model parameters of the support vector machine, wherein the method comprises the following steps: determining a target value function of a gray wolf algorithm; setting a search range of a penalty factor and a Gaussian kernel function parameter, a population size of the wolf and the maximum iteration times; randomly initializing a wolf population; classifying the initialized wolf populations, sorting according to the calculated numerical value of the target value function of each wolf according to the size, and sequentially naming the wolf populations as a, b, c and d from large to low according to the set population size; updating the positions of the a wolf, the b wolf and the c wolf, calculating the positions of the d wolf and the updated a wolf, b wolf and c wolf, and updating the current position of the d wolf according to the positions of the three leading wolfs closest to the d; and calculating and comparing the value of the optimal target value function of the new population, updating the value of the target value function if the value of the optimal target value function is better than the value of the previous target value function value, adding one to the iteration times, and continuously and circularly executing position updating until the iteration times reach the maximum iteration times.
Further, predicting the thrombolysis effect of the target to be detected based on the blood concentration comprises the following steps: determining a target concentration interval where the blood concentration is located; and determining the thrombolytic effect corresponding to the target concentration interval based on a preset concentration interval and thrombolytic effect corresponding relation table.
In a second aspect, an embodiment of the present invention further provides a system for predicting a thrombolysis effect of piezoelectric impedance based on support vector machine regression, including: the device comprises a testing module, a processing module, a training module, a first prediction module and a second prediction module; the test module is used for testing piezoelectric impedance data corresponding to thrombolytic solutions with different concentrations by using the thrombus vibration actuator; the processing module is used for carrying out standardization processing on the piezoelectric impedance data to obtain a training set and a test set; the training module is used for inputting the training set and the test set as characteristic parameters into a regression model of a support vector machine optimized by a wolf algorithm for training and testing to obtain a trained regression model of the support vector machine; the first prediction module is used for predicting the blood concentration of the target to be measured based on the trained support vector machine regression model and the target piezoelectric impedance data; the target piezoelectric impedance data is blood piezoelectric impedance data acquired by the thrombus vibration actuator on the target to be detected; and the second prediction module is used for predicting the thrombolysis effect of the target to be detected based on the blood concentration.
Further, the test module is further configured to: mixing and simulating thrombolytic solutions with different concentrations, and inserting the thrombolytic solution into the thrombus vibration actuator; inputting a sine driving signal with a preset resonance frequency through a function generator and a power amplifier of the thrombus vibration actuator, so that the thrombus vibration actuator generates vibration and fully stirs the thrombolytic solutions with different concentrations; and measuring real part piezoelectric impedance data of the thrombolysis solutions with different concentrations after stirring by using an impedance analyzer of the thrombus vibration actuator to obtain piezoelectric impedance data corresponding to the thrombolysis solutions with different concentrations.
Further, the training module is further configured to: performing mathematical analysis on the piezoelectric impedance data to construct a regression model of a support vector machine; establishing a gray wolf optimization algorithm and searching the optimal value of the regression model parameters of the support vector machine, wherein the method comprises the following steps: determining a target value function of a gray wolf algorithm; setting a search range of a penalty factor and a Gaussian kernel function parameter, a population size of the wolf and the maximum iteration number; randomly initializing a wolf population; classifying the initialized wolf populations, sorting according to the calculated numerical value of the target value function of each wolf according to the size, and sequentially naming the wolf populations as a, b, c and d from large to low according to the set population size; updating the positions of the a wolf, the b wolf and the c wolf, calculating the positions of the d wolf and the updated a wolf, b wolf and c wolf, and updating the current position of the d wolf according to the positions of the three leading wolfs closest to the d; and calculating and comparing the value of the optimal target value function of the new population, updating the value of the target value function if the value of the optimal target value function is better than the value of the previous target value function value, adding one to the iteration times, and continuously and circularly executing position updating until the iteration times reach the maximum iteration times.
Further, the second prediction module is further configured to: determining a target concentration interval where the blood concentration is located; and determining the thrombolytic effect corresponding to the target concentration interval based on a preset concentration interval and thrombolytic effect corresponding relation table.
In a third aspect, an embodiment of the present invention further provides an electronic device, including: a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the processing method according to the first aspect as described above when executing the computer program.
The invention provides a method and a system for predicting a piezoelectric impedance thrombolysis effect based on support vector machine regression, wherein the thrombolysis effect of a thrombus vibration actuator is affected by the outside and shows nonlinear change, a thrombus thrombolysis effect prediction model constructed by the support vector machine regression optimized by a Hui wolf algorithm is used for training concentration obtained by an experiment and corresponding piezoelectric impedance data, the blood concentration of the broken thrombolysis actuator can be accurately predicted through fed-back piezoelectric impedance, the prediction precision is obviously improved compared with that of a traditional regression support vector machine, and the technical problems of low reliability and low accuracy in the prior art are solved.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the detailed description or the prior art description will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a flowchart of a method for predicting a piezoelectric impedance thrombolysis effect based on support vector machine regression according to an embodiment of the present invention;
FIG. 2 is a schematic view of a thrombus vibration actuator according to an embodiment of the present invention;
FIG. 3 is a diagram illustrating a comparison between a predicted value and a true value of a regression of a conventional SVM according to an embodiment of the present invention;
FIG. 4 is a comparison graph of the predicted value and the true value of the thrombus stirring effect based on the support vector machine regression optimized by the Grey wolf algorithm provided by the embodiment of the present invention;
fig. 5 is a schematic diagram of a system for predicting a thrombolysis effect of piezo-impedance based on support vector machine regression according to an embodiment of the present invention.
Detailed Description
To further illustrate the technical means and effects of the present invention adopted to achieve the predetermined objects, the following detailed description of the embodiments, structures, features and effects according to the present invention will be made with reference to the accompanying drawings and preferred embodiments.
The first embodiment is as follows:
fig. 1 is a flowchart of a method for predicting a thrombolysis effect of piezoelectric impedance based on support vector machine regression according to an embodiment of the present invention. As shown in fig. 1, the method specifically includes the following steps:
and S102, testing piezoelectric impedance data corresponding to thrombolytic solutions with different concentrations by using a thrombus vibration actuator.
And step S104, carrying out standardization processing on the piezoelectric impedance data to obtain a training set and a test set.
Optionally, in the embodiment of the present invention, the solution with different concentrations and the piezoelectric impedance data are normalized based on a mean variance normalization method.
And S106, inputting the training set and the test set as characteristic parameters into a regression model of the support vector machine optimized by the Grey wolf algorithm for training and testing to obtain a trained regression model of the support vector machine.
Step S108, predicting the blood concentration of the target to be detected based on the trained regression model of the support vector machine and the target piezoelectric impedance data; the target piezoelectric impedance data is blood piezoelectric impedance data acquired by the thrombus vibration actuator on the target to be detected.
And step S110, predicting the thrombolysis effect of the target to be detected based on the blood concentration.
The invention provides a piezoelectric impedance thrombolysis effect prediction method based on support vector machine regression, which is characterized in that the thrombolysis effect of a thrombus vibration actuator is affected by the outside and shows nonlinear change, a thrombus thrombolysis effect prediction model constructed by the support vector machine regression optimized by a Hui wolf algorithm is used for training concentration obtained by an experiment and corresponding piezoelectric impedance data, the blood concentration of the broken thrombolysis actuator can be accurately predicted through fed-back piezoelectric impedance, the prediction precision is obviously improved compared with that of a traditional regression support vector machine, and the technical problems of low reliability and low accuracy in the prior art are solved.
Specifically, step S102 further includes the steps of:
and step S1021, matching and simulating thrombolytic solutions with different concentrations, and inserting a thrombus vibration actuator.
Alternatively, FIG. 2 is a schematic view of a thrombus vibration actuator provided according to an embodiment of the invention. As shown in fig. 2, in the embodiment of the present invention, the thrombus vibration actuator includes an impedance analyzer, an upper computer, a power amplifier and a signal generator; wherein the impedance analyzer includes a piezoelectric crystal.
Step S1022, inputting a sine driving signal with a preset resonance frequency through a function generator and a power amplifier of the thrombus vibration actuator, so that the thrombus vibration actuator generates vibration and fully stirs thrombolytic solutions with different concentrations;
and step S1023, measuring the real part piezoelectric impedance data of the thrombolysis solutions with different concentrations after stirring by an impedance analyzer of the thrombus vibration actuator to obtain the piezoelectric impedance data corresponding to the thrombolysis solutions with different concentrations.
For example, in the embodiment of the invention, a micro thrombus vibration actuator is designed, an experimental platform simulating thrombolysis of the actuator is built, a corresponding thrombolysis effect test is carried out in an optimal frequency range obtained by the experiment, 60 groups of real part piezoelectric impedance values fed back by piezoelectric elements under different blood concentrations are obtained, all data are subjected to standardization processing, 2/3 of experimental data are randomly selected from the data to serve as a training set, and the other 1/3 of data serve as a test set.
Specifically, in the process of collecting 60 groups of experimental data, solution with well-proportioned concentration is adopted, the tip of the actuator is inserted into the solution, a 50V sinusoidal signal with the resonance frequency of 659.8Hz is input through the signal generator and the amplifier, and the actuator is vibrated to generate vibration and enable the solution to be uniformly stirred. After stirring uniformly, inputting a 1V sweep frequency signal by adopting an impedance analyzer, setting the sweep frequency to be 600Hz-730Hz, the step length to be 1.3Hz, setting the acquisition point to be 101 points, and taking the piezoelectric impedance of the real part fed back by the piezoelectric crystal as output. In the experiment, the impedance analyzer connected with the piezoelectric crystal can record and output the real part piezoelectric impedance value of the piezoelectric crystal under different blood concentrations in real time and store the real part piezoelectric impedance value to the upper computer. 2/3 of the tested 60 groups of experimental data are randomly selected as a training set, and the rest 1/3 of the groups of data are taken as a testing set. Meanwhile, aiming at the dimension problem among all variables in the sample, the mean variance normalization processing is carried out on the training and testing data.
Specifically, step S106 further includes the steps of:
step S1061, performing mathematical analysis on the piezoelectric impedance data to construct a regression model of a support vector machine;
step S1062, establishing a gray wolf optimization algorithm and searching for an optimal value of a regression model parameter of a support vector machine, including:
determining a target value function of a gray wolf algorithm;
setting a search range of a penalty factor and a Gaussian kernel function parameter, a population size of the wolf and the maximum iteration times;
randomly initializing a wolf population;
classifying the initialized wolf populations, sorting according to the calculated numerical value of the target value function of each wolf according to the size, and sequentially naming the wolf populations as a, b, c and d from large to low according to the set population size;
updating the positions of the a wolf, the b wolf and the c wolf, calculating the positions of the d wolf and the updated a wolf, b wolf and c wolf, and updating the current position of the d wolf according to the positions of the three leading wolfs closest to the d;
and calculating and comparing the value of the optimal target value function of the new population, updating the value of the target value function if the value of the optimal target value function is better than the value of the previous target value function value, adding one to the iteration times, and continuously and circularly executing position updating until the iteration times reach the maximum iteration times.
Specifically, in the embodiment of the present invention, the blood concentration and the real piezoelectric impedance value fed back by the piezoelectric wafer are mathematically analyzed, the support vector machine regression is adopted to predict the corresponding piezoelectric impedance thrombolysis effect, and the gaussian kernel function is selected as the kernel function of the support vector machine regression; first, assume that a sample set containing n samples is given
Wherein x i For the ith feature vector, R represents the real number set.
Mapping the sample set from a low-dimensional space to a high-dimensional space through nonlinearity, wherein a support vector regression expression of the nonlinearity mapping is expressed as:
in the formula: k (x) i ,x)＝φ(x i )φ(x j ) Is a kernel function, b is an intercept, α i And
is a Lagrangian multiplier, where x i And x j Are the ith and jth feature vectors.
In order to support the high prediction precision of the vector machine regression, a Gaussian kernel function is introduced as a kernel function of the support vector machine regression, and the kernel function is expressed as:
compared with the traditional regression model, the prediction is judged to be correct only when the predicted value f (x) is equal to the true value y, namely the loss is zero, the regression of the support vector machine tolerates a loss deviation epsilon between the true value and the predicted value, when the predicted value is in the deviation zone, the prediction is also considered to be correct, and then the SVR problem can be formalized as follows:
wherein C is a penalty factor, l ε Is an epsilon-insensitive loss function, which can be expressed as:
introducing a relaxation variable xi i And
equation (3) can be written as: />
Where ε (ε > 0) is the maximum error allowed, and introducing the kernel function and the Lagrangian multiplier, equation (5) can be expressed as the following equation:
wherein alpha is i And
expressed as lagrangian multiplier, when the lagrangian function takes the minimum value, the expression of the support vector machine after high-dimensional mapping can be expressed as:
wherein K (x, x) i )＝φ(x i )φ(x j ) Is the kernel function and b is the intercept.
Aiming at the problem of inconsistent dimensions in data, preprocessing the data by using mean variance normalization, wherein a formula of the mean variance normalization is as follows:
where x is the value to be normalized, μ is the sample mean, and S is the standard deviation of the sample.
Considering that the penalty coefficient C and the kernel function parameter gamma have larger influence on the prediction model, the penalty factor C and the kernel function parameter gamma are optimized by utilizing a wolf optimization algorithm.
After the prediction model is established, a real part piezoelectric impedance signal of a piezoelectric crystal of the thrombus vibration actuator is used as an input sample, a blood concentration value is used as an output sample, and 2/3 of 60 groups of experimental data are used as large sample training data to improve the accuracy of prediction.
Specifically, in the embodiment of the present invention, the gray wolf optimization algorithm is established and the optimization process of the regression model parameters of the support vector machine is performed as follows:
optimizing the penalty coefficient C and the kernel function parameter gamma, wherein the process comprises the following steps: 1) Determining a target value function of a gray wolf algorithm; 2) Setting a penalty factor C and a search range ori _ w _ range _ C, ori _ w _ range _ gamma, the population size num of the wolf and the maximum iteration number total _ num of the Gaussian kernel function parameter gamma; 3) Randomly initializing a wolf population; 4) Classifying the initialized wolf populations, sorting according to the calculated fitness function value of each wolf, and sequentially naming a, b, c and d from large to low according to the set population size; 5) Updating the positions of a, b, c and the like wolfs, calculating d and the positions of the updated a, b, c and the like wolfs, and updating the current position of the d wolf according to the positions of three leading wolfs closest to d; 6) And calculating and comparing the optimal target value function value of the new population, updating the target value function value if the optimal target value function value is better than the previous target value function value, and simultaneously, turning to 5) to continue execution. If the value is worse than the previous target value function value, the set counter is increased by 1, and when the maximum iteration number total _ nums is reached or the count of the counter reaches 20, the algorithm exits; 7) And optimizing the parameters of the regression type support vector machine by utilizing a wolf algorithm.
Optionally, the method provided in the embodiment of the present invention further includes: using a decision coefficient R 2 And the mean square error E judges the model prediction accuracy of the established support vector machine regression (GWOO-SVR) prediction model optimized by the Grey wolf algorithm, and determines a coefficient R 2 And the mean square error E is expressed as:
FIG. 3 is a drawing showingFig. 4 is a comparison graph of a predicted value and a true value of a thrombus stirring effect of support vector machine regression optimized based on a grayish wolf algorithm provided by the embodiment of the present invention, wherein horizontal and vertical coordinates of fig. 3 and fig. 4 are dimensionless data after normalization processing. As shown in fig. 3 and fig. 4, compared with the conventional SVR, the GWO-SVR prediction algorithm in the embodiment of the present invention significantly improves accuracy. The mean square error E of GWOO-SVR is reduced from 0.05483498191998703 to 0.02810084622159569, the reduction is 48.75%, and a coefficient R is determined 2 The increase from 94.52% to 97.19%.
Specifically, step S110 further includes the steps of:
step S1101, determining a target concentration interval where the blood concentration is;
step S1102, determining a thrombolysis effect corresponding to the target concentration interval based on the preset concentration interval and thrombolysis effect correspondence table.
In the embodiment of the invention, after the concentration of the thrombolytic solution is accurately predicted, an evaluation standard of the thrombolytic effect can be preliminarily made according to the prediction result, the concentration of the solution is predicted to be 0% -20% of low thrombolytic effect, the concentration is predicted to be 20% -50% of medium thrombolytic effect, and the concentration is predicted to be 50% -70% of high thrombolytic effect, as shown in table 1.
TABLE 1 corresponding relationship table between preset concentration interval and thrombolytic effect
In conclusion, the accuracy of predicting the thrombolysis degree by the GWO-SVR-based prediction algorithm is higher than that of the traditional SVR prediction algorithm which is the regression of the support vector machine, and the solution concentration can be accurately predicted by the piezoelectric impedance to be used as an important basis for evaluating the thrombolysis effect of the actuator.
Example two:
fig. 5 is a schematic diagram of a system for predicting a thrombolysis effect of a piezoelectric impedance based on support vector machine regression according to an embodiment of the present invention. As shown in fig. 5, the system includes: test module 10, processing module 20, training module 30, first prediction module 40, and second prediction module 50.
Specifically, the testing module 10 is configured to test piezoelectric impedance data corresponding to thrombolytic solutions of different concentrations by using the thrombus vibration actuator.
And the processing module 20 is configured to perform normalization processing on the piezoelectric impedance data to obtain a training set and a test set.
Optionally, in an embodiment of the present invention, the piezoelectric impedance data is normalized based on a mean variance normalization method.
And the training module 30 is configured to input the training set and the test set as characteristic parameters into a support vector machine regression model optimized by the grayish wolf algorithm for training and testing to obtain a trained support vector machine regression model.
The first prediction module 40 is used for predicting the blood concentration of the target to be detected based on the trained support vector machine regression model and the target piezoelectric impedance data; the target piezoelectric impedance data is blood piezoelectric impedance data acquired by the thrombus vibration actuator on the target to be detected.
And the second prediction module 50 is used for predicting the thrombolysis effect of the target to be detected based on the blood concentration.
The invention provides a piezoelectric impedance thrombolysis effect prediction system based on support vector machine regression, which is characterized in that the thrombolysis effect of a thrombus vibration actuator is affected by the outside and shows nonlinear change, a thrombus thrombolysis effect prediction model constructed by the support vector machine regression optimized by a Hui wolf algorithm is used for training the concentration obtained by the experiment and corresponding piezoelectric impedance data, the blood concentration of the broken thrombolysis actuator can be accurately predicted through the fed-back piezoelectric impedance, the prediction precision is obviously improved compared with that of the traditional regression support vector machine, and the technical problems of low reliability and low accuracy in the prior art are solved.
Specifically, the test module 10 is further configured to:
mixing and simulating thrombolytic solutions with different concentrations, and inserting a thrombus vibration actuator;
inputting a sine driving signal with a preset resonance frequency through a function generator and a power amplifier of the thrombus vibration actuator, so that the thrombus vibration actuator generates vibration and fully stirs thrombolytic solutions with different concentrations;
and measuring the real part piezoelectric impedance data of the thrombolysis solutions with different concentrations after stirring by using an impedance analyzer of the thrombus vibration actuator to obtain the piezoelectric impedance data corresponding to the thrombolysis solutions with different concentrations.
Specifically, the training module 30 is further configured to:
performing mathematical analysis on the piezoelectric impedance data to construct a regression model of a support vector machine;
establishing a gray wolf optimization algorithm and searching an optimal value of a regression model parameter of a support vector machine, wherein the method comprises the following steps:
determining a target value function of a gray wolf algorithm;
setting a search range of a penalty factor and a Gaussian kernel function parameter, a population size of the wolf and the maximum iteration times;
randomly initializing a wolf population;
classifying the initialized wolf populations, sorting according to the calculated numerical value of the target value function of each wolf, and sequentially naming the wolf populations as a, b, c and d from large to low according to the set population size;
updating the positions of the a wolf, the b wolf and the c wolf, calculating the positions of the d wolf and the updated a wolf, the b wolf and the c wolf, and updating the current position of the d wolf according to the positions of the three leading wolfs closest to the d;
and calculating and comparing the value of the optimal target value function of the new population, updating the value of the target value function if the value of the optimal target value function is better than the value of the previous target value function value, adding one to the iteration times, and continuously and circularly executing position updating until the iteration times reach the maximum iteration times.
Specifically, the second prediction module 50 is further configured to:
determining a target concentration interval where the blood concentration is;
and determining the thrombolytic effect corresponding to the target concentration interval based on the corresponding relation table of the preset concentration interval and the thrombolytic effect.
An embodiment of the present invention further provides an electronic device, including: the device comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor executes the computer program to realize the processing method in the first embodiment.
Although the present invention has been described with reference to the preferred embodiments, it will be understood by those skilled in the art that various changes may be made and equivalents may be substituted for elements thereof without departing from the scope of the present invention.
Claims (10)
1. A piezoelectric impedance thrombolysis effect prediction method based on support vector machine regression is characterized by comprising the following steps:
testing piezoelectric impedance data corresponding to thrombolytic solutions with different concentrations by using a thrombus vibration actuator;
carrying out standardization processing on the piezoelectric impedance data to obtain a training set and a test set;
inputting the training set and the test set as characteristic parameters into a regression model of a support vector machine optimized by a wolf algorithm for training and testing to obtain a trained regression model of the support vector machine;
predicting the blood concentration of the target to be measured based on the trained support vector machine regression model and the target piezoelectric impedance data; the target piezoelectric impedance data is blood piezoelectric impedance data acquired by the thrombus vibration actuator on the target to be detected;
and predicting the thrombolysis effect of the target to be detected based on the blood concentration.
2. The method of claim 1, wherein the step of testing piezoelectric impedance data corresponding to different concentrations of thrombolytic solutions by using a thrombus vibration actuator comprises:
mixing and simulating thrombolytic solutions with different concentrations, and inserting the thrombolytic solution into the thrombus vibration actuator;
inputting a sine driving signal with a preset resonance frequency through a function generator and a power amplifier of the thrombus vibration actuator, so that the thrombus vibration actuator generates vibration and fully stirs the thrombolytic solutions with different concentrations;
and measuring real part piezoelectric impedance data of the thrombolysis solutions with different concentrations after stirring by using an impedance analyzer of the thrombus vibration actuator to obtain piezoelectric impedance data corresponding to the thrombolysis solutions with different concentrations.
3. The method of claim 1, wherein normalizing the piezoelectric impedance data comprises: and normalizing the piezoelectric impedance data based on a mean variance normalization method.
4. The method of claim 1, wherein the training and testing are performed by inputting the training set and the testing set as feature parameters into a regression model of a support vector machine optimized by a grayish wolf algorithm, comprising:
performing mathematical analysis on the piezoelectric impedance data to construct a regression model of a support vector machine;
establishing a gray wolf optimization algorithm and searching the optimal value of the regression model parameters of the support vector machine, wherein the method comprises the following steps:
determining a target value function of a gray wolf algorithm;
setting a search range of a penalty factor and a Gaussian kernel function parameter, a population size of the wolf and the maximum iteration times;
randomly initializing a wolf population;
classifying the initialized wolf populations, sorting according to the calculated numerical value of the target value function of each wolf according to the size, and sequentially naming the wolf populations as a, b, c and d from large to low according to the set population size;
updating the positions of the a wolf, the b wolf and the c wolf, calculating the positions of the d wolf and the updated a wolf, the b wolf and the c wolf, and updating the current position of the d wolf according to the positions of the three leading wolfs closest to the d;
and calculating and comparing the value of the optimal target value function of the new population, updating the value of the target value function if the value of the optimal target value function is better than the value of the previous target value function value, adding one to the iteration times, and continuously and circularly executing position updating until the iteration times reach the maximum iteration times.
5. The method of claim 1, wherein predicting the thrombolytic effect of the test object based on blood concentration comprises:
determining a target concentration interval where the blood concentration is located;
and determining the thrombolytic effect corresponding to the target concentration interval based on a preset concentration interval and thrombolytic effect corresponding relation table.
6. A piezoelectric impedance thrombolysis effect prediction system based on support vector machine regression is characterized by comprising the following components: the device comprises a testing module, a processing module, a training module, a first prediction module and a second prediction module; wherein the content of the first and second substances,
the testing module is used for testing piezoelectric impedance data corresponding to thrombolytic solutions with different concentrations by using the thrombus vibration actuator;
the processing module is used for carrying out standardization processing on the piezoelectric impedance data to obtain a training set and a test set;
the training module is used for inputting the training set and the test set as characteristic parameters into a support vector machine regression model optimized by a wolf algorithm for training and testing to obtain a trained support vector machine regression model;
the first prediction module is used for predicting the blood concentration of the target to be measured based on the trained support vector machine regression model and the target piezoelectric impedance data; the target piezoelectric impedance data is blood piezoelectric impedance data acquired by the thrombus vibration actuator on the target to be detected;
and the second prediction module is used for predicting the thrombolysis effect of the target to be detected based on the blood concentration.
7. The system of claim 6, wherein the testing module is further configured to:
mixing and simulating thrombolytic solutions with different concentrations, and inserting the thrombolytic solution into the thrombus vibration actuator;
inputting a sine driving signal with a preset resonance frequency through a function generator and a power amplifier of the thrombus vibration actuator, so that the thrombus vibration actuator generates vibration and fully stirs the thrombolytic solutions with different concentrations;
and measuring the real part piezoelectric impedance data of the thrombolysis solutions with different concentrations after stirring by using an impedance analyzer of the thrombus vibration actuator to obtain the piezoelectric impedance data corresponding to the thrombolysis solutions with different concentrations.
8. The system of claim 6, wherein the training module is further configured to:
performing mathematical analysis on the piezoelectric impedance data to construct a regression model of a support vector machine;
establishing a gray wolf optimization algorithm and searching the optimal value of the regression model parameters of the support vector machine, wherein the method comprises the following steps:
determining a target value function of a gray wolf algorithm;
setting a search range of a penalty factor and a Gaussian kernel function parameter, a population size of the wolf and the maximum iteration times;
randomly initializing a wolf population;
classifying the initialized wolf populations, sorting according to the calculated numerical value of the target value function of each wolf according to the size, and sequentially naming the wolf populations as a, b, c and d from large to low according to the set population size;
updating the positions of the a wolf, the b wolf and the c wolf, calculating the positions of the d wolf and the updated a wolf, the b wolf and the c wolf, and updating the current position of the d wolf according to the positions of the three leading wolfs closest to the d;
and calculating and comparing the value of the optimal target value function of the new population, updating the value of the target value function if the value of the optimal target value function is better than the value of the previous target value function value, adding one to the iteration times, and continuously and circularly executing position updating until the iteration times reach the maximum iteration times.
9. The system of claim 6, wherein the second prediction module is further configured to:
determining a target concentration interval where the blood concentration is located;
and determining the thrombolytic effect corresponding to the target concentration interval based on a preset concentration interval and thrombolytic effect corresponding relation table.
10. An electronic device, comprising: memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the processing method according to any one of claims 1 to 5 when executing the computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211678107.2A CN115969464B (en) | 2022-12-26 | 2022-12-26 | Method and system for predicting thrombolysis effect of piezoelectric impedance based on regression of support vector machine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211678107.2A CN115969464B (en) | 2022-12-26 | 2022-12-26 | Method and system for predicting thrombolysis effect of piezoelectric impedance based on regression of support vector machine |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115969464A true CN115969464A (en) | 2023-04-18 |
| CN115969464B CN115969464B (en) | 2024-05-10 |
Family
ID=85969523
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211678107.2A Active CN115969464B (en) | 2022-12-26 | 2022-12-26 | Method and system for predicting thrombolysis effect of piezoelectric impedance based on regression of support vector machine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115969464B (en) |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5014715A (en) * | 1988-11-22 | 1991-05-14 | Chapolini Robert J | Device for measuring the impedance to flow of a natural or prosthetic vessel in a living body |
| JPH0422370A (en) * | 1990-05-17 | 1992-01-27 | Olympus Optical Co Ltd | Thrombus dissolutive therapy device |
| CN110956610A (en) * | 2019-10-31 | 2020-04-03 | 广州市番禺区中心医院(广州市番禺区人民医院、广州市番禺区心血管疾病研究所) | Method and system for predicting thrombolytic curative effect of lower limb deep venous thrombosis based on sparse representation |
| CN111145901B (en) * | 2019-12-04 | 2021-02-09 | 深圳大学 | Deep venous thrombosis thrombolytic curative effect prediction method and system, storage medium and terminal |
| US20240055122A1 (en) * | 2020-12-18 | 2024-02-15 | The Johns Hopkins University | Methods, systems and related aspects for real-time prediction of adverse outcomes using machine learning and high-dimensional clinical data |
| CN113331912B (en) * | 2021-07-20 | 2022-09-06 | 昆明理工大学 | Vibration actuator for removing cerebral thrombosis and thrombus removing equipment |
| CN115101162A (en) * | 2022-06-16 | 2022-09-23 | 温州大学 | Thrombolysis efficiency prediction method and system based on biological fluid mechanics |
| CN115276452A (en) * | 2022-08-29 | 2022-11-01 | 苏州谱洛医疗科技有限公司 | Thrombolysis system circuit, ultrasonic energy adjusting method and ultrasonic thrombolysis system |
| CN115440383B (en) * | 2022-09-30 | 2023-03-21 | 中国医学科学院北京协和医院 | System for predicting curative effect of PD-1/PD-L1 monoclonal antibody of advanced cancer patient |
-
2022
- 2022-12-26 CN CN202211678107.2A patent/CN115969464B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN115969464B (en) | 2024-05-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11341404B2 (en) | Analysis-data analyzing device and analysis-data analyzing method that calculates or updates a degree of usefulness of each dimension of an input in a machine-learning model | |
| Zhaomin et al. | Batch process monitoring based on multisubspace multiway principal component analysis and time-series Bayesian inference | |
| Heatley | Nuclear magnetic relaxation and models for backbone motions of macromolecules in solution | |
| Fan et al. | Effects of data quality and quantity on deep learning for protein-ligand binding affinity prediction | |
| CN112164426A (en) | Drug small molecule target activity prediction method and device based on TextCNN | |
| US20230274789A1 (en) | Media, methods, and systems for protein design and optimization | |
| JP2018040787A (en) | Method for predicting chemical shift values of nmr spin systems in sample of fluid class, in particular in sample of biofluid | |
| Mello-Roman et al. | KPLS optimization with nature-inspired metaheuristic algorithms | |
| Born et al. | Trends in deep learning for property-driven drug design | |
| CN115969464B (en) | Method and system for predicting thrombolysis effect of piezoelectric impedance based on regression of support vector machine | |
| Zhao et al. | Discriminative learning for protein conformation sampling | |
| Feng et al. | A Novel Genetic Algorithm‐Based Optimization Framework for the Improvement of Near‐Infrared Quantitative Calibration Models | |
| Chen et al. | Prediction of dMRI signals with neural architecture search | |
| Abseher et al. | NMR cross-relaxation investigated by molecular dynamics simulation: a case study of ubiquitin in solution | |
| Correa et al. | A knowledge-based artificial bee colony algorithm for the 3-D protein structure prediction problem | |
| Goutham et al. | Artificial intelligence for compressive strength prediction of concrete | |
| CN114359428A (en) | Magnetic resonance fingerprint imaging dictionary resolution optimization method and device | |
| EP4105672A1 (en) | Systems and methods for provisioning training data to enable neural networks to analyze signals in nmr measurements | |
| Ertefaie et al. | Comment: Outcome-Wide Individualized Treatment Strategies | |
| Showalter | Intrinsically disordered proteins: methods for structure and dynamics Studies | |
| US20230116904A1 (en) | Selecting a cell line for an assay | |
| Pawar et al. | Exploring Aromatase Inhibitors to Breast Cancer with Machine Learning Approach Based Quantitative Structure-Activity Relationships Model Technique | |
| Peng et al. | Performance analysis between different decision trees for uncertain data | |
| Terashima et al. | Neural networks for a quick access to a digital twin of scanning physical property measurements | |
| Geng et al. | Computer Simulation of Gene Knockout in Bacteria Accelerated by AI |
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 |