CN116226605B - Fitting method and system for stimulator parameters - Google Patents
Fitting method and system for stimulator parameters Download PDFInfo
- Publication number
- CN116226605B CN116226605B CN202310450578.6A CN202310450578A CN116226605B CN 116226605 B CN116226605 B CN 116226605B CN 202310450578 A CN202310450578 A CN 202310450578A CN 116226605 B CN116226605 B CN 116226605B
- Authority
- CN
- China
- Prior art keywords
- data
- target
- clustering
- class
- rule
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
Abstract
The invention relates to the technical field of parameter fitting, in particular to a fitting method and system of stimulator parameters, wherein the method comprises the steps of obtaining test data, wherein the test data are multiple groups of data, each group of data comprises multiple data, and the data types of the multiple data in each group of data are different; expanding the test data according to the test data, the data type and a preset data expansion rule to obtain expanded data; based on the correlation calculation rule, calculating correlation coefficients among the expansion data of different data types; processing the expansion data according to the correlation coefficient and the data processing rule to determine target data; clustering the target data based on a preset clustering rule, and determining target clustering data; and fitting the target cluster data of each class based on a preset fitting rule to obtain a target fitting result. The method has the effect of improving the accuracy of the fitting curve.
Description
Technical Field
The application relates to the technical field of parameter fitting, in particular to a method and a system for fitting parameters of a stimulator.
Background
In many fields of science and technology, it is often necessary to find their functional relational expressions from a series of data obtained by actual tests. In theory, the polynomial f (x) of degree n can be constructed according to the interpolation principle, so that the function value of f (x) at each test point just passes through the real test point. However, in many cases, in order to reflect the real situation as much as possible, a plurality of sample points are acquired, resulting in a high degree of the interpolation polynomial f (x), which not only increases the calculation amount of the function, but also affects the approximation degree; and then, as the interpolation polynomial passes through each measured sample point, the measurement error is reserved, so that the accuracy of the fitting function is affected, and the function relation of actual data is not easily reflected. Therefore, generally, a functional relation between the tested quantities is found according to the known actual tested sample points, so that the fitted functional curve can fully reflect the relation between the actual tested quantities.
Currently, in the field of nerve treatment, a nerve stimulator is used for treating a patient, and an external energy controller is required to simultaneously supply electric energy to the implanted nerve stimulator and transmit stimulation parameters, and also is required to monitor the working state of the implanted nerve stimulator in real time. The doctor can adjust the transmitted stimulation parameters through the change of the working state of the implantable neural stimulator, and the problem of untimely adjustment of the stimulation parameters can occur. It will be appreciated that the external energy controller needs to adjust the transmitted stimulation parameters according to the change of the operating state of the implantable neural stimulator, so that a relationship between the change of the operating state of the implantable neural stimulator and the transmitted stimulation parameters needs to be established, and the expression form of the change of the operating state of the implantable neural stimulator is the change of the stimulation parameters, so that a data change curve between the stimulation parameters, namely, fitting of the stimulation parameters needs to be established. By establishing a data change curve between the stimulation parameters, the external energy controller can timely adjust the output stimulation parameters according to the change of the stimulation parameters, and the treatment effect of the nerve stimulator on the patient can be improved to a certain extent, so that the treatment experience of the patient is improved. The corresponding data change curve is obtained by analyzing various data of the implanted nerve stimulator, and the external energy controller adjusts the stimulation parameters transmitted to the implanted nerve stimulator according to the data change curve. When all data are fitted, the obtained fitted curve is influenced by other abnormal data, so that the accuracy of the fitted curve is low.
The prior art solutions described above have the following drawbacks: the accuracy of the fitting curve is low.
Disclosure of Invention
In order to solve the problem of low accuracy of a fitted curve, the application provides a fitting method and a fitting system of stimulator parameters.
In a first aspect of the present application, a method of fitting stimulator parameters is provided. The method comprises the following steps:
obtaining test data of a stimulator, wherein the test data are multiple groups of data, each group of data comprises multiple data, and the data types of the multiple data in each group of data are different;
expanding the test data according to the test data, the data type and a preset data expansion rule to obtain expanded data;
based on the correlation calculation rule, calculating correlation coefficients among the expansion data of different data types;
processing the expansion data according to the correlation coefficient and the data processing rule to determine target data;
clustering the target data based on a preset clustering rule, and determining target clustering data;
and fitting the target cluster data of each class based on a preset fitting rule to obtain a target fitting result.
According to the technical scheme, test data are obtained, and the test data are expanded according to the test data, the data type and the preset data expansion rule, so that expanded data are obtained; calculating correlation coefficients among the expansion data of different data types; processing the expansion data through judging the correlation coefficient to determine target data; and clustering the target data, and fitting the clustered target clustering data to obtain a target fitting result. The final fitting result is more accurate by data processing before data clustering fitting, and the accuracy of the fitting curve is improved.
In one possible implementation manner, expanding the test data according to the test data, the data type and a preset data expansion rule to obtain expanded data, including:
multiplying and adding any two data of different data types in each group of data to obtain multiplied index data and added index data; the test data, the multiplied index data and the added index data form expansion data;
or alternatively, the first and second heat exchangers may be,
dividing and subtracting any two data of different data types in each group of data to obtain divided index data and subtracted index data; and the test data, the division index data and the subtraction index data form expansion data.
In one possible implementation manner, the processing the expansion data according to the correlation coefficient and the data processing rule to determine target data includes:
screening the expansion data according to the correlation coefficient and the data screening rule to determine first target data;
and carrying out normalization processing on the first target data to obtain second target data, wherein the second target data is the target data.
In one possible implementation manner, the clustering the target data based on a preset clustering rule, and determining target clustering data includes:
clustering the target data to obtain first clustering data, wherein the first clustering data comprises a plurality of classes of data, and the data is key value pairs (x, y);
judging the data of each class, and when the values of x in the key value pair (x, y) of each class are the same or the values of y are the same, the class is an abnormal class; when there are one or more differences in the x or y values in the key value pair (x, y) for each class, the class is a feature class;
determining clustering parameters according to preset data calculation rules and the feature classes;
classifying the feature classes into a first feature class and a second feature class based on a preset data judging rule;
clustering again according to the clustering parameters, the second feature class, the abnormal class and the clustering rule, and determining pending clustering data;
when the pending cluster data accords with a preset cluster result, the pending cluster data is the target cluster data.
In one possible implementation manner, the determining the clustering parameter according to the preset data calculation rule and the feature class includes:
summing key value pairs (x, y) in the feature class respectively to obtain a plurality of sum values, wherein the sum value = x+y; the minimum value of the plurality of sum values is a clustering parameter.
In one possible implementation manner, the classifying the feature class into the first feature class and the second feature class based on a preset data judging rule includes:
calculating the median of the data in the feature class;
judging the ratio of the number of data exceeding the median to the total number of data in the feature class;
when the ratio exceeds a preset value, the feature class is a first feature class;
and when the ratio is smaller than a preset value, the feature class is a second feature class.
In one possible implementation, the method further includes:
acquiring actual data or an actual data range;
and carrying the actual data or the actual data range into the target fitting result to obtain a target data value or a target data range.
In a second aspect of the present application, a fitting system for stimulator parameters is provided. The system comprises:
the data acquisition module is used for acquiring test data, wherein the test data are multiple groups of data, each group of data comprises multiple data, and the data types of the multiple data in each group of data are different;
the first data processing module is used for expanding the test data according to the test data, the data type and a preset data expansion rule to obtain expanded data;
the data calculation module is used for calculating the correlation coefficient between the expansion data of different data types based on the correlation calculation rule;
the second data processing module is used for processing the expansion data according to the correlation coefficient and the data processing rule and determining target data;
the data clustering module is used for clustering the target data based on a preset clustering rule and determining target clustering data;
the data fitting module is used for fitting the target cluster data of each class based on a preset fitting rule to obtain a target fitting result.
In summary, the present application includes at least one of the following beneficial technical effects:
acquiring test data and expanding the test data to obtain expanded data; then calculating correlation coefficients among the expansion data of different data types; processing the expansion data through judging the correlation coefficient to determine target data; after the target data are obtained, clustering the target data, and fitting the clustered target data to obtain a target fitting result. The final fitting result is more accurate by data processing before data clustering fitting, and the accuracy of the fitting curve is improved.
Drawings
Fig. 1 is a flow chart of a fitting method of stimulator parameters provided in the present application.
Fig. 2 is a schematic structural diagram of a fitting system for stimulator parameters provided herein.
Fig. 3 is a schematic structural diagram of an electronic device provided in the present application.
In the figure, 200, a fitting system of stimulator parameters; 201. a data acquisition module; 202. a first data processing module; 203. a data calculation module; 204. a second data processing module; 205. a data clustering module; 206. and a data fitting module.
Description of the embodiments
For the purposes of making the objects, technical solutions and advantages of the embodiments of the present application more clear, the technical solutions of the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is apparent that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art based on the embodiments herein without making any inventive effort, are intended to be within the scope of the present application.
In addition, the term "and/or" herein is merely an association relationship describing an association object, and means that three relationships may exist, for example, a and/or B may mean: a exists alone, A and B exist together, and B exists alone. In this context, unless otherwise specified, the term "/" generally indicates that the associated object is an "or" relationship.
At present, an implanted nerve stimulator in a nerve stimulation system based on radio frequency control performs radio frequency communication and energy transmission with an external energy controller, and the external energy controller provides electric stimulation pulses in real time to drive stimulation electrodes of the implanted nerve stimulator so as to apply stimulation signals to a treatment part of a patient; and the external energy controller supplies radio frequency electric energy to the implanted nerve stimulator to maintain the operation of the implanted nerve stimulator.
The stimulator is implanted in the body when in use, the antenna of the stimulator is implanted in the subcutaneous fat layer, and the specific implantation position of the stimulator is according to the type of stimulator (such as the spinal nerve stimulator SCS is implanted in the epidural space of the spinal cord). Comprising the following steps: n (typically 8 or 16) electrodes, a processor MCU, a memory, an antenna, bluetooth, and various detection modules, such as a temperature detection module, a voltage detection module, an impedance detection module, etc. The electric energy is obtained through the antenna and transmitted to the stimulation chip, and the stimulation chip is controlled to generate stimulation waveforms and transmit the stimulation waveforms to the stimulation electrodes. Pulse signals are sent to nerves in the body through the electrodes so as to realize the therapeutic effect. The parameters of the pulse signals come from the energy controller and are received through Bluetooth or radio frequency antennas. The power of the stimulator is also from the energy controller. The stimulator can also realize detection of temperature, voltage, impedance and the like, and the detection result is fed back to the energy controller through Bluetooth.
The energy controller is positioned outside the body, and the antenna of the energy controller is aligned with the stimulator antenna in the body so as to achieve higher transmission efficiency. The energy controller comprises: the device comprises a processor MCU, a memory, an antenna, bluetooth, an external communication module and a battery. The wireless power supply is carried out on the stimulator through the radio frequency antenna, wireless communication interaction is carried out on the stimulator through the Bluetooth or the radio frequency antenna, the wireless power supply comprises parameters for sending pulse signals to the stimulator and detection results fed back by the stimulator are received.
The external energy controller is used for simultaneously providing electric energy for the implanted nerve stimulator and transmitting stimulation parameters, and also is required to monitor the working state of the implanted nerve stimulator in real time. Currently, various data of an implanted neural stimulator, such as frequency, pulse width, voltage, etc., are collected in real time. The data of the implanted nerve stimulator is analyzed to obtain a corresponding data change curve, and the energy controller adjusts the implanted nerve stimulator and transmits stimulation parameters according to the data change curve. When fitting is performed on all acquired data, namely, one type of data is used as a dependent variable, and all other types of data are used as independent variables, on the one hand, the calculated amount is large, and on the other hand, the accuracy of the obtained fitting result can be influenced when the data with poor correlation are used as the independent variables. The fitting curve obtained when fitting a plurality of items of data is affected by other items of data, so that the accuracy of the fitting curve is low. When fitting data of two data types, abnormal data exists in the fitted data, and the accuracy of a fitted curve is also affected.
Embodiments of the present application are described in further detail below with reference to the drawings attached hereto.
The embodiment of the application provides a fitting method of stimulator parameters, and the main flow of the method is described as follows.
As shown in fig. 1:
step S101: test data is obtained.
Specifically, the test data is that after the stimulator is implanted into a human body, various data such as frequency, pulse width, voltage and the like are collected in real time. The test data is a plurality of groups of data, each group of data comprises a plurality of data, and the data types of the plurality of data in each group of data are different, namely the data types of the test data comprise but are not limited to frequency, pulse width and voltage. A set of data represents the frequency, pulse width and voltage acquired at a certain point in time, and the data types of a plurality of data in the set of data are different. The plurality of sets of data refer to frequency, pulse width, voltage and other data acquired at a plurality of time points.
In this embodiment, the test data includes multiple sets of data, each set of data including a frequency and a voltage.
After the stimulator is implanted into a human body, the energy controller can provide electric energy for the stimulator through radio frequency communication and transmit stimulation parameters, the stimulator generates pulse signals after receiving the signals, and the energy controller acquires various data such as frequency, pulse width, voltage and the like generated in the working process of the stimulator. The stimulator is located inside the human body, the energy controller is arranged outside the human body corresponding to the stimulator, and the stimulator and the energy controller have the characteristic of small volume, so that the data calculation capacity of the stimulator and the energy controller is weak, and the energy controller sends various received data to other equipment with strong data calculation capacity through wireless communication to perform corresponding data analysis and calculation.
Step S102: and expanding the test data according to the test data, the data type and a preset data expansion rule to obtain expanded data.
Specifically, when the data amount of the acquired test data is small, expansion of the test data is required. In one embodiment, any two data of different data types in each group of data are multiplied and added to obtain multiplication index data and addition index data; the test data, the multiplication index data, and the addition index data constitute expansion data. For example, the frequencies and voltages in a group of data are multiplied to obtain multiplied index data, the frequencies and voltages in a group of data are added to obtain added index data, the added index data and the multiplied index data are expanded data, the test data, the added index data and the multiplied index data form expanded data, and further analysis and calculation are performed on the expanded data.
In a first implementation manner corresponding to the present embodiment, the extension data includes multiple sets of data, where each set of data includes a frequency, a voltage, a sum of the frequency and the voltage, and a product of the frequency and the voltage.
In another embodiment, any two data of different data types in each group of data are divided and subtracted to obtain divided index data and subtracted index data; the test data, the division index data and the subtraction index data form expansion data.
In a second implementation manner corresponding to the present embodiment, the expansion data includes multiple sets of data, where each set of data includes a frequency, a voltage, a difference value between the frequency and the voltage, and a quotient value between the frequency and the voltage.
In the data processing process of the stimulator, the problem that data indexes are limited exists, for example, only three types of indexes including voltage, frequency and pulse width exist, the influence of the frequency on the data change of the voltage is small, the inaccuracy condition of the fitting relation between the analysis voltage and the frequency exists, the data are expanded through a data expansion rule, and the sum value and the product of the frequency and the pulse width are used as new indexes to participate in data analysis. It may occur that the sum or product of the frequency and the pulse width has a greater effect on the data change of the voltage than the single data indicator of the frequency or pulse width. Therefore, the accuracy of the fitting curve can be improved to a certain extent by analyzing the expanded data.
Step S103: based on the correlation calculation rule, calculating correlation coefficients among the expansion data of different data types.
Specifically, a correlation coefficient between the expansion data of different data types is calculated, for example, a correlation coefficient between frequency and voltage is calculated, and a correlation coefficient between voltage and pulse width is calculated. The correlation coefficient may be a spearman correlation coefficient or a Pi Erman correlation coefficient, or may be other data representing a correlation between two data, which is not limited herein. The above-mentioned method for calculating the correlation coefficient, i.e. the correlation calculation rule, is a technical means known to those skilled in the art, and will not be described herein.
After the expansion data are acquired, a correlation coefficient between the expansion data needs to be calculated, and the correlation coefficient can reflect the correlation between the data of two different data types. For example, there are n sets of data each including a pulse width, a voltage, and a frequency, for which a correlation coefficient between the pulse width and the voltage, a correlation coefficient between the pulse width and the frequency, and a correlation coefficient between the voltage and the frequency are calculated, respectively.
The calculation formula of the correlation coefficient ρ between the voltage and the frequency:
；
wherein x is i Indicating the value of the i-th voltage,representing the average of n voltage values, y i Represents the i-th frequency value,/->Representing n frequenciesAverage of the values. The calculation modes of other data correlation coefficients are the same, and are not described in detail herein.
In the present embodiment, a first correlation coefficient of frequency and voltage, a second correlation coefficient of frequency and sum of frequency and voltage, a third correlation coefficient of frequency and sum of frequency and voltage, a fourth correlation coefficient of voltage and sum of frequency and voltage, a fifth correlation coefficient of voltage and sum of frequency and voltage, and a sixth correlation coefficient of frequency and voltage.
Step S104: and processing the expansion data according to the correlation coefficient and the data processing rule, and determining target data.
Specifically, the expansion data is screened according to the correlation coefficient and the data screening rule, and first target data is determined; and carrying out normalization processing on the first target data to obtain second target data, wherein the second target data is the target data. And screening the data according to the correlation coefficient among the test data of different data types, and selecting the data type with the highest correlation with the data type for the data of a certain data type, namely, for the data type of the data type, the data of the data type with the highest correlation has the highest influence on the data, and the data of the data type with the highest correlation needs to be modified or adjusted. The data type is used as a dependent variable, and the data type with the highest correlation with the data type is used as the independent variable for fitting. After the screening is finished, the screened expansion data is normalized, and the normalized data is the target data. By judging the correlation coefficient, the data can be paired in pairs, for example, when the correlation coefficient between the frequency and the voltage is larger than the correlation coefficient between the voltage and the pulse width, the influence of the frequency on the pulse width is larger than the influence of the voltage on the pulse width, so that the pulse width data and the frequency data are paired for subsequent processing.
The normalization processing is to convert dimensionless data into scalar quantity through transformation. Methods of normalization include, but are not limited to, linear normalization, which is a linear transformation of the raw data, mapping the data values between [0,1], and Z-score normalization. The above-mentioned Z-score normalization refers to a process of dividing the difference between a number and an average by the standard deviation. Normalization of data is a well known technique for those skilled in the art, and is not described in detail herein.
In this embodiment, for the frequency, the first correlation coefficient, the second correlation coefficient, and the third correlation coefficient are compared and a maximum value (in this embodiment, the maximum value represents the maximum data correlation of the two data types) is obtained, and when the first correlation coefficient is the maximum, the frequency and the voltage are combined into a data pair, and the difference between the frequency and the voltage and the quotient between the frequency and the voltage are combined into a data pair. And carrying out normalization processing on the data forming the data pairs, wherein the obtained target data are still two groups of data pairs, namely, the data pairs of frequency and voltage, the difference value of the frequency and the voltage and the data pair of the quotient value of the frequency and the voltage.
Step S105: and clustering the target data based on a preset clustering rule, and determining target clustering data.
Specifically, the target data is clustered, and the clustering rule may be gaussian clustering, K-means clustering, or other clustering methods, which are not limited herein. And clustering the target data to obtain first clustered data, wherein the target data comprises a plurality of groups of data pairs, for example, the data pairs formed by frequency and voltage are a group of data pairs, and the data pairs formed by the difference value of the frequency and the voltage and the quotient value of the frequency and the voltage are a group of data pairs. The first cluster data also comprises a plurality of groups, and one group of target data corresponds to one group of first cluster data. The group of first cluster data comprises a plurality of classes of data, wherein the data is a key value pair (x, y); judging the data of each class, and when the values of x in the key value pair (x, y) of each class are the same or the values of y are the same, the class is an abnormal class; a class is a feature class when there are one or more differences in the x or y values in the key value pairs (x, y) of each class. The number of the feature classes is one or more, and the number of the abnormal classes is one or more.
And determining clustering parameters according to preset data calculation rules and the characteristic classes. The above-mentioned clustering parameter refers to the particles of the cluster, and the key value pairs (x, y) in the feature class are summed to obtain a plurality of sum values, where the sum value=x+y; the minimum value of the sum values is a cluster parameter, i.e. a particle.
And classifying the characteristic classes into a first characteristic class and a second characteristic class based on a preset data judging rule. Calculating the median of the data in the feature class, wherein the median is the median x of the x in the feature class media And the median y of y media Constituent key-value pairs (x media ，y media ). Acquiring the ratio of the number of the data in the feature class exceeding the median to the total number of the data; when the ratio exceeds a preset value, the characteristic class is a first characteristic class; and when the ratio is smaller than a preset value, the characteristic class is a second characteristic class. In this embodiment, the preset value is 95%, that is, when the ratio is greater than or equal to 95%, the feature class is a first feature class, and when the ratio is less than 95%, the feature class is a second feature class.
And clustering again according to the clustering parameters, the second feature class, the abnormal class and the clustering rule, and determining pending clustering data. And clustering the data in the second characteristic class and the abnormal class again by taking the clustering parameters as the mass centers of the clustering clusters to obtain pending clustering data. When the pending cluster data accords with a preset cluster result, the pending cluster data is the target cluster data. When the distance between the mass center of each class in the pending cluster data and all the data in the class is smaller than or equal to the minimum preset value, the completion of the cluster is indicated, and the pending cluster data is the target cluster data; when the distances between the mass center of each class in the pending cluster data and all the data in the class are larger than a minimum preset value, the pending cluster data are not clustered, and the data need to be clustered continuously until the pending cluster data accord with a preset clustering result.
In this embodiment, each set of data corresponds to a pair of data of one frequency and voltage, and the plurality of sets of data corresponds to a plurality of pairs of data. And clustering the plurality of data pairs of the frequency and the voltage to obtain data pairs of a plurality of classes, and taking the class as an abnormal class when the frequency values or the voltage values in one class are the same. Classes that are not exception classes are taken as feature classes. And further judging the feature class, calculating the median of the feature class, namely, the median of the voltage data and the median of the frequency data to form a median key value pair, and when the ratio of the number of the data pairs exceeding the median key value in the feature class to the total number of the data pairs in the feature class is more than or equal to 95%, marking the feature class as a first feature class, and when the ratio is less than 95%, marking the feature class as a second feature class.
Summing key value pairs (x, y) in the feature class to obtain a plurality of sum values, wherein the sum value=x+y; the minimum value of the plurality of sum values is a clustering parameter. And clustering the data pairs in the abnormal class and the second characteristic class again, taking the clustering parameters as particles of the new primary clustering, recording each class obtained by clustering as pending clustering data, and indicating that the clustering is completed when the distance between the mass center of each class in the pending clustering data and all data in the class is smaller than or equal to a minimum preset value, wherein the pending clustering data is target clustering data. The minimum preset value is set by a user according to fitting requirements.
The processing of the data pair consisting of the difference values of the plurality of frequencies and the voltages and the quotient values of the frequencies and the voltages in the plurality of groups of data is the same as the processing mode of the data pair of the plurality of frequencies and the voltages, and is not described herein.
Step S106: and fitting the target cluster data of each class based on a preset fitting rule to obtain a target fitting result.
Specifically, after clustering is completed, target cluster data are obtained, fitting is carried out on each type of target cluster data, a pending fitting result is obtained, whether the pending fitting result is identical to a theoretical fitting result or not is judged, if so, the pending fitting result is the target fitting result, and if not, the pending fitting result is not adopted. For example, theoretically, the theoretical fitting result of a certain pair of variables is a quadratic polynomial, and if the pending fitting result is a first order polynomial or a third order polynomial, the pending fitting result is not used as the target fitting result, i.e. the pending fitting result is not used; when the pending fitting result is a quadratic polynomial, the pending fitting result is a target fitting result.
In this embodiment, a plurality of classes formed by data pairs of a plurality of frequencies and voltages are fitted, each class corresponds to a fitted curve of a corresponding relationship between the frequencies and the voltages, whether the fitted curve is the same as a theoretical fitted result is determined, if so, the fitted curve is a target fitted result of the frequencies and the voltages, and if not, the fitted curve is not adopted.
And fitting the data consisting of the difference values of the frequencies and the voltages and the quotient values of the frequencies and the voltages to a plurality of classes consisting of the differences, and judging the fitting result to be the same as the fitting result of the frequencies and the voltages. As can be seen from the above procedure, the target fitting result includes one or more fitting curves.
The fitting method of the stimulator parameters further comprises the following steps:
acquiring actual data or an actual data range; and carrying the actual data or the actual data range into the target fitting result to obtain a target data value or a target data range.
Specifically, the target fitting result represents a correspondence between an independent variable and a dependent variable, the actual data is the independent variable, the actual data range represents a variation range of the independent variable, and the actual data or the actual data range is input by a user according to an actual requirement. And (3) the actual data or the actual data range is brought into the target fitting result, so that the corresponding target data value or target data range can be obtained. The device is adjusted when the value of the device at the time of actual use differs from the target data value or the value at the time of actual use is not within the target data range.
The target fitting result between the corresponding data is obtained by analyzing each item of data of the stimulator, when the corresponding relation between the two items of data does not accord with the target fitting result, the energy controller obtains the target data value of the dependent variable by inputting the actual value of the independent variable into the target fitting result, and then the stimulation parameter is transmitted to the stimulator according to the target data value, so that the purpose of correcting the stimulator parameter is achieved.
An embodiment of the present application provides a fitting system 200 of stimulator parameters, referring to fig. 2, the fitting system 200 of stimulator parameters includes:
the data acquisition module 201 is configured to acquire test data of the stimulator, where the test data is a plurality of sets of data, each set of data includes a plurality of data, and data types of the plurality of data in each set of data are different;
the first data processing module 202 is configured to expand the test data according to the test data, the data type and a preset data expansion rule, so as to obtain expanded data;
the data calculation module 203 is configured to calculate correlation coefficients between the expanded data of different data types based on a correlation calculation rule;
the second data processing module 204 is configured to process the expansion data according to the correlation coefficient and the data processing rule, and determine target data;
the data clustering module 205 is configured to cluster the target data based on a preset clustering rule, and determine target cluster data;
the data fitting module 206 is configured to fit each type of target cluster data based on a preset fitting rule, so as to obtain a target fitting result.
It will be clear to those skilled in the art that, for convenience and brevity of description, reference may be made to the corresponding process in the foregoing method embodiment for the specific working process of the described module, which is not described herein again.
The foregoing description is only of the preferred embodiments of the present application and is presented as a description of the principles of the technology being utilized. It will be appreciated by persons skilled in the art that the scope of the application referred to in this application is not limited to the specific combinations of features described above, but it is intended to cover other embodiments in which any combination of features described above or their equivalents is possible without departing from the spirit of the application. Such as the above-mentioned features and the technical features having similar functions (but not limited to) applied for in this application are replaced with each other.
Claims (5)
1. A method of fitting stimulator parameters, comprising:
obtaining test data of a stimulator, wherein the test data are multiple groups of data, each group of data comprises multiple data, and the data types of the multiple data in each group of data are different;
expanding the test data according to the test data, the data type and a preset data expansion rule to obtain expanded data;
expanding the test data according to the test data, the data type and a preset data expansion rule to obtain expanded data, wherein the expanding data comprises the following steps:
multiplying and adding any two data of different data types in each group of data to obtain multiplied index data and added index data; the test data, the multiplied index data and the added index data form expansion data;
or alternatively, the first and second heat exchangers may be,
dividing and subtracting any two data of different data types in each group of data to obtain divided index data and subtracted index data; the test data, the division index data and the subtraction index data form expansion data;
based on the correlation calculation rule, calculating correlation coefficients among the expansion data of different data types;
processing the expansion data according to the correlation coefficient and the data processing rule to determine target data;
the processing the expansion data according to the correlation coefficient and the data processing rule to determine target data comprises the following steps:
screening the expansion data according to the correlation coefficient and the data screening rule to determine first target data;
normalizing the first target data to obtain second target data, wherein the second target data is target data;
clustering the target data based on a preset clustering rule, and determining target clustering data;
the clustering of the target data based on the preset clustering rule, and the determining of the target clustering data comprise the following steps:
clustering the target data to obtain first clustering data, wherein the first clustering data comprises a plurality of classes of data, the data are key value pairs (x, y), and x and y in the key value pairs are target data with different data types;
judging the data of each class, and when the values of x in the key value pair (x, y) of each class are the same or the values of y are the same, the class is an abnormal class; when there are one or more differences in the x or y values in the key value pair (x, y) for each class, the class is a feature class;
determining clustering parameters according to preset data calculation rules and the feature classes;
classifying the feature classes into a first feature class and a second feature class based on a preset data judging rule;
clustering again according to the clustering parameters, the second feature class, the abnormal class and the clustering rule, and determining pending clustering data;
when the pending cluster data accords with a preset cluster result, the pending cluster data is target cluster data;
and fitting the target cluster data of each class based on a preset fitting rule to obtain a target fitting result.
2. The method of fitting stimulator parameters according to claim 1, wherein said determining cluster parameters according to preset data calculation rules and said feature classes comprises:
summing key value pairs (x, y) in the feature class respectively to obtain a plurality of sum values, wherein the sum value = x+y; the minimum value of the plurality of sum values is a clustering parameter.
3. The fitting method of stimulator parameters according to claim 1, wherein the classifying the feature classes into a first feature class and a second feature class based on a preset data judgment rule comprises:
calculating the median of the data in the feature class;
judging the ratio of the number of data exceeding the median to the total number of data in the feature class;
when the ratio exceeds a preset value, the feature class is a first feature class;
and when the ratio is smaller than a preset value, the feature class is a second feature class.
4. The method of fitting stimulator parameters according to claim 1, further comprising:
acquiring actual data or an actual data range;
and carrying the actual data or the actual data range into the target fitting result to obtain a target data value or a target data range.
5. A fitting system for stimulator parameters, comprising:
the data acquisition module (201) is used for acquiring test data, wherein the test data is a plurality of groups of data, each group of data comprises a plurality of data, and the data types of the plurality of data in each group of data are different;
the first data processing module (202) is used for expanding the test data according to the test data, the data type and a preset data expansion rule to obtain expanded data; expanding the test data according to the test data, the data type and a preset data expansion rule to obtain expanded data, wherein the expanding data comprises the following steps: multiplying and adding any two data of different data types in each group of data to obtain multiplied index data and added index data; the test data, the multiplied index data and the added index data form expansion data; or dividing and subtracting any two data of different data types in each group of data to obtain divided index data and subtracted index data; the test data, the division index data and the subtraction index data form expansion data;
the data calculation module (203) is used for calculating correlation coefficients among the expansion data of different data types based on the correlation calculation rule;
the second data processing module (204) is used for processing the expansion data according to the correlation coefficient and the data processing rule and determining target data; the processing the expansion data according to the correlation coefficient and the data processing rule to determine target data comprises the following steps: screening the expansion data according to the correlation coefficient and the data screening rule to determine first target data; normalizing the first target data to obtain second target data, wherein the second target data is target data;
the data clustering module (205) is used for clustering the target data based on a preset clustering rule to determine target clustering data; the clustering of the target data based on the preset clustering rule, and the determining of the target clustering data comprise the following steps: clustering the target data to obtain first clustering data, wherein the first clustering data comprises a plurality of classes of data, the data are key value pairs (x, y), and x and y in the key value pairs are target data with different data types; judging the data of each class, and when the values of x in the key value pair (x, y) of each class are the same or the values of y are the same, the class is an abnormal class; when there are one or more differences in the x or y values in the key value pair (x, y) for each class, the class is a feature class; determining clustering parameters according to preset data calculation rules and the feature classes; classifying the feature classes into a first feature class and a second feature class based on a preset data judging rule; clustering again according to the clustering parameters, the second feature class, the abnormal class and the clustering rule, and determining pending clustering data; when the pending cluster data accords with a preset cluster result, the pending cluster data is target cluster data;
and the data fitting module (206) is used for fitting the target cluster data of each class based on a preset fitting rule to obtain a target fitting result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310450578.6A CN116226605B (en) | 2023-04-25 | 2023-04-25 | Fitting method and system for stimulator parameters |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310450578.6A CN116226605B (en) | 2023-04-25 | 2023-04-25 | Fitting method and system for stimulator parameters |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116226605A CN116226605A (en) | 2023-06-06 |
| CN116226605B true CN116226605B (en) | 2023-07-25 |
Family
ID=86582870
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310450578.6A Active CN116226605B (en) | 2023-04-25 | 2023-04-25 | Fitting method and system for stimulator parameters |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116226605B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2007131066A2 (en) * | 2006-05-02 | 2007-11-15 | Physiostream, Inc. | Decentralized physiological data collection and analysis system and process |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5995868A (en) * | 1996-01-23 | 1999-11-30 | University Of Kansas | System for the prediction, rapid detection, warning, prevention, or control of changes in activity states in the brain of a subject |
| EP1558131A4 (en) * | 2002-10-15 | 2008-05-28 | Medtronic Inc | Timed delay for redelivery of treatment therapy for a medical device system |
| US8078273B2 (en) * | 2007-03-30 | 2011-12-13 | Neurometrix, Inc. | Detection and classification of neuromuscular late wave activity for the assessment of neuromuscular function |
| WO2012030987A2 (en) * | 2010-08-31 | 2012-03-08 | Rehabilitation Institute Of Chicago | Method for reflex threshold estimation in spastic muscles |
| EP3873335A4 (en) * | 2018-11-02 | 2022-07-27 | Saluda Medical Pty Ltd | Characterisation of neurostimulation therapeutic efficacy |
| US11413457B2 (en) * | 2019-01-22 | 2022-08-16 | Boston Scientific Neuromodulation Corporation | Fitting algorithm for recruiting of neural targets in a spinal cord stimulator system |
| CN114219169A (en) * | 2021-12-23 | 2022-03-22 | 上海颖幡技术有限公司 | Script banner supply chain sales and inventory prediction algorithm model and application system |
| CN114870250B (en) * | 2022-04-20 | 2022-11-25 | 浙江帝诺医疗科技有限公司 | Nerve regulation system and nerve regulation stimulator based on same |
-
2023
- 2023-04-25 CN CN202310450578.6A patent/CN116226605B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2007131066A2 (en) * | 2006-05-02 | 2007-11-15 | Physiostream, Inc. | Decentralized physiological data collection and analysis system and process |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116226605A (en) | 2023-06-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102274022B (en) | Sleep state monitoring method based on electroencephalogram signals | |
| Zardoshti-Kermani et al. | EMG feature evaluation for movement control of upper extremity prostheses | |
| Phinyomark et al. | Feature extraction and reduction of wavelet transform coefficients for EMG pattern classification | |
| CN100482155C (en) | Instant detection system and detection method for state of attention based on interaction between brain and computer | |
| Coelho et al. | Assessing fractal dimension methods as feature extractors for EMG signal classification | |
| US20100168819A1 (en) | Rf telemetry link quality assessment system and method | |
| CN113052113B (en) | Depression identification method and system based on compact convolutional neural network | |
| Treves et al. | Firing rate distributions and efficiency of information transmission of inferior temporal cortex neurons to natural visual stimuli | |
| CN102930284A (en) | Surface electromyogram signal pattern recognition method based on empirical mode decomposition and fractal | |
| CN111476158B (en) | Multi-channel physiological signal somatosensory gesture recognition method based on PSO-PCA-SVM | |
| CN109452938B (en) | HFECG signal characteristic frequency detection method based on multi-scale multi-fractal | |
| Wu et al. | Hybrid BF–PSO and fuzzy support vector machine for diagnosis of fatigue status using EMG signal features | |
| Del Boca et al. | Myoelectric signal recognition using fuzzy clustering and artificial neural networks in real time | |
| CN110598676A (en) | Deep learning gesture electromyographic signal identification method based on confidence score model | |
| CN109512390A (en) | Sleep stage method and wearable device based on EEG time domain various dimensions feature and M-WSVM | |
| CN106923820A (en) | A kind of electrocardiosignal artifact identification method and electrocardiosignal artifact identifying device | |
| CN116226605B (en) | Fitting method and system for stimulator parameters | |
| CN108281192B (en) | Human body component prediction method and system based on ensemble learning algorithm | |
| Julkunen | Mobile application for adaptive threshold hunting in transcranial magnetic stimulation | |
| CN115804601A (en) | N-Back subtask electroencephalogram test method based on multi-complexity reducer assembly scene | |
| CN109214431B (en) | Sample training method, classification method, identification method, device, medium and system | |
| Kramer et al. | Real-time measurement of muscle fatigue related changes in surface EMG | |
| Meinel et al. | Tikhonov regularization enhances EEG-based spatial filtering for single trial regression | |
| Fatchan et al. | Determination of Nutritional Status Using Classification Method Datamining Using K-Nearst Neighbord (KNN) Algorithm | |
| Smith et al. | Evaluation of simple algorithms for spectral parameter analysis of the electroencephalogram |
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 |