Intelligent analysis method for electrochemical recognition of flavonoids

Abstract

The present application relates to the technical field of intelligent analysis of flavonoids, and particularly relates to an intelligent analysis method for electrochemical recognition of flavonoids. The method comprises the following steps: obtaining a response signal matrix in time-frequency dimension, and performing frequency domain compression and nonlinear interaction mapping to obtain a one-dimensional nonlinear compressed current response sequence; introducing a variable window integral singular response recognition mechanism based on the one-dimensional nonlinear compressed current response sequence to calculate a singular response strength; determining a singular response occurrence point based on the singular response strength, and performing weighted fusion on a local window tensor; performing asynchronous reconstruction processing on the obtained singular weighted fusion tensor to obtain a characteristic representation of the flavonoid target after asynchronous reconstruction processing, and performing intelligent analysis on the flavonoids. The method solves the problems of low resolution, poor stability and repeatability of traditional electrochemical recognition analysis methods for flavonoids, and the inability to obtain characteristics of various target objects in flavonoids, thereby improving the analysis accuracy.

Description

Technical Field

[0001] This invention relates to the field of intelligent analysis technology for flavonoids, and more particularly to an intelligent analysis method for the electrochemical identification of flavonoids. Background Technology

[0002] With the increasing demands for accuracy in identifying natural active substances in applications such as food functional testing, identification of traditional Chinese medicine components, and nutritional health assessment, flavonoids, as a class of functional components widely found in plant-derived substances, have become a research hotspot for qualitative identification and quantitative analysis. Currently, electrochemical analysis is widely used for the detection of flavonoids due to its advantages such as high sensitivity, rapid response, and no need for complex sample pretreatment. However, in practical applications, traditional electrochemical analysis techniques still face several technical bottlenecks when processing samples containing a mixture of multiple structurally similar flavonoids. For example, the response signals obtained by conventional voltammetry, potential step method, or cyclic voltammetry are mostly one-dimensional current-time or current-voltage curves, lacking the ability to express the structure of complex redox behavior under multi-frequency electrical excitation. This leads to severe peak overlap and unstable response amplitudes in the response curves of different flavonoid components, making it impossible to achieve high-resolution and effective identification. On the other hand, existing signal analysis methods mostly rely on manual feature extraction or low-dimensional statistical parameters (such as peak height, full width at half maximum, potential shift, etc.), making it difficult to capture higher-order features contained in the response signal, such as frequency interaction coupling, response abrupt changes, and nonlinear transitions. Furthermore, different flavonoids exhibit inconsistent start-up time delays and varying response rates during redox reactions, making it difficult for traditional machine learning methods based on static feature models to effectively model asynchronous behavior, further reducing classification accuracy and model generalization ability.

[0003] Therefore, traditional electrochemical identification and analysis of flavonoids still suffers from technical problems such as low resolution, poor stability and repeatability, and low analytical accuracy due to the inability to obtain the characteristics of various target compounds in flavonoids. Summary of the Invention

[0004] This invention provides an intelligent analysis method for the electrochemical identification of flavonoids, which solves the technical problems of low resolution, poor stability and repeatability in traditional electrochemical identification and analysis of flavonoids, as well as the low analytical accuracy caused by the inability to obtain the characteristics of various target compounds in flavonoids.

[0005] The intelligent analysis method for electrochemical recognition of flavonoids of the present invention specifically includes the following technical solutions:

[0006] A smart analytical method for the electrochemical recognition of flavonoids includes the following steps:

[0007] S1. Obtain the response signal matrix in the time-frequency dimension, and perform frequency domain compression and nonlinear interactive mapping through a nonlinear dual-frequency cross-spectrum compression algorithm to obtain a one-dimensional nonlinear compressed current response sequence; based on the one-dimensional nonlinear compressed current response sequence, introduce a variable window integral singular response identification mechanism to calculate the singular response intensity.

[0008] S2. Based on singular response intensities, a singular response intensity sequence is constructed to determine the singular response occurrence points and obtain singular response weighting coefficients; a local window tensor is constructed, and combined with the singular response weighting coefficients, a singular weighted fusion tensor is obtained; an inversion mechanism based on the mapping between current response gradient and time delay is introduced to asynchronously reconstruct the singular weighted fusion tensor to obtain the feature representation of the flavonoid target after asynchronous reconstruction; based on the feature representation of the flavonoid target after asynchronous reconstruction, intelligent analysis of flavonoid compounds is performed.

[0009] Preferably, S1 specifically includes:

[0010] In the implementation of the nonlinear dual-frequency cross-spectral compression algorithm, based on the response signal matrix in the time-frequency dimension, a current derivative synergistic term is constructed to quantify the consistency and synergy of the reaction trends of flavonoids under different excitation frequencies. In addition, the frequency interaction weight is calculated by combining the frequency domain distance between excitation frequencies, and a compression kernel matrix is ​​constructed.

[0011] Preferably, S1 specifically includes:

[0012] Based on the compression kernel matrix, the response signal matrix in the time-frequency dimension is weighted in the frequency dimension to generate a spectral compression representation, resulting in a one-dimensional nonlinear compressed current response sequence.

[0013] Preferably, S1 specifically includes:

[0014] In the implementation of the variable window integral singular response identification mechanism, based on the one-dimensional nonlinear compressed current response sequence, local stationarity coefficients and normalized periodic sine mapping terms are constructed, and the steady-state jump response score of the time segment is calculated to obtain the singular response intensity.

[0015] Preferably, S2 specifically includes:

[0016] The singular response intensity sequence is processed to obtain a set of maxima. Based on the set of maxima and a preset threshold, the occurrence point of the singular response is determined.

[0017] Preferably, S2 specifically includes:

[0018] After normalizing the singular response intensity corresponding to the singular response occurrence point, singular response weighting coefficients are obtained. Based on the singular response weighting coefficients, the local window tensor is weighted and fused to obtain the singular weighted fused tensor. The local window tensor is obtained by feature extraction of the one-dimensional nonlinear compressive current response corresponding to the singular response occurrence point.

[0019] Preferably, S2 specifically includes:

[0020] In the implementation of the inversion mechanism based on the mapping of current response gradient and time delay, a time diffusion scale factor is introduced based on the singular weighted fusion tensor, and combined with the gradient penalty term, a sliding integral reconstruction is performed on each time step to obtain the asynchronous response window vector of the flavonoid target.

[0021] Preferably, S2 specifically includes:

[0022] Based on the asynchronous response window vectors of flavonoid targets, and combined with an attention mechanism, the association weights between flavonoid targets are obtained. Based on the association weights between flavonoid targets, the asynchronous response window vectors of flavonoid targets are weighted and accumulated to obtain the feature representations of flavonoid targets after asynchronous reconstruction.

[0023] Preferably, S2 specifically includes:

[0024] The feature representations of all flavonoid targets after asynchronous reconstruction are merged to generate a feature representation tensor in a unified semantic space. The feature representation tensor is then input into the inference network to output the classification probability of each type of flavonoid target, thereby realizing intelligent analysis of flavonoid compounds.

[0025] The beneficial effects of the technical solution of the present invention are:

[0026] 1. The nonlinear dual-frequency cross-spectral compression algorithm introduced in this invention, by constructing a compression kernel matrix controlled by a logistic normalization function, fully utilizes the inherent synergistic trend between the current responses of flavonoid target substances at different excitation frequencies. This effectively preserves the structural information contained in the cross-frequency coupling and eliminates the interference of low-correlation frequency bands. Compared to traditional Fourier or wavelet transform methods, this invention significantly improves the structural consistency between the spectral mapping and the physical response by introducing a synergistic term of the square of the frequency difference and the current derivative to jointly construct frequency interaction weights. This allows the final generated one-dimensional nonlinear compressed current response sequence to suppress redundant channel noise while maintaining information integrity, thereby enhancing the discriminability of the response behavior in subsequent analysis tasks.

[0027] 2. This invention introduces a variable window integral singular response identification mechanism based on a one-dimensional nonlinear compressed current response sequence. Its key feature is that by constructing an integral combination of local stationarity coefficients and normalized periodic sine mapping terms, it quantifies and enhances the transient current peaks formed by the short-term strong responses of flavonoids. This not only accurately determines the time segment of redox transitions but also effectively avoids the noise sensitivity problem in traditional differential-threshold transition detection. Furthermore, it can adapt to the inconsistent response time scales of different flavonoids, thus ensuring robust capture capability of the short-term response behavior of various flavonoid targets.

[0028] 3. To address the issue of response delay differences in flavonoids under electrochemical excitation, this invention designs an inversion mechanism based on the mapping between current response gradient and time delay. A kernel function with a time diffusion scale factor centered on the response delay is constructed to nonlinearly reconstruct different response time windows. By introducing an independent reaction start-up time delay for each type of flavonoid target, the aggregation mode of different response behaviors on the time axis can be automatically adjusted, thereby uniformly representing their reaction trajectory in logical time. This breaks through the limitations of "fixed window, synchronous sampling" in traditional convolutional neural networks, establishing a highly adaptable feature reconstruction mechanism for electrochemical detection tasks with strong time asynchronicity. Attached Figure Description

[0029] Figure 1 This is a flowchart of the intelligent analysis method for electrochemical recognition of flavonoids described in this invention. Detailed Implementation

[0030] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0032] The following description, in conjunction with the accompanying drawings, details the specific scheme of the intelligent analysis method for electrochemical recognition of flavonoids provided by this invention.

[0033] See attached document Figure 1 The diagram illustrates a flowchart of an intelligent analysis method for electrochemical recognition of flavonoids provided by an embodiment of the present invention. The method includes the following steps:

[0034] S1. Obtain the response signal matrix in the time-frequency dimension, and perform frequency domain compression and nonlinear interactive mapping through a nonlinear dual-frequency cross-spectrum compression algorithm to obtain a one-dimensional nonlinear compressed current response sequence; based on the one-dimensional nonlinear compressed current response sequence, introduce a variable window integral singular response identification mechanism to calculate the singular response intensity.

[0035] The time-frequency response signal matrix was obtained from the flavonoid solution using an electrochemical sensor at a set excitation frequency. Response signal matrix in the time-frequency dimension It is a two-dimensional dynamic current response data structure, in which, It is a sampling time series. For the first Each excitation potential corresponds to a frequency, i.e., the excitation frequency; the electrochemical sensor is a microelectrode sensing platform based on a three-electrode system, including a working electrode, a reference electrode, and an auxiliary electrode, and has excitation control capability, response acquisition capability, and frequency division resolution capability.

[0036] Because different flavonoids exhibit weak but specific redox response peaks in specific frequency ranges, the response modes implicit in the time-frequency response signal matrix exhibit frequency-dimensional cross-coupling and time-dimensional drift randomness. Therefore, a nonlinear dual-frequency cross-spectral compression algorithm is needed to perform frequency domain compression and nonlinear interactive mapping on the time-frequency response signal matrix to retain physically meaningful spectral cooperative modes, thereby extracting physically identifiable feature time series and obtaining a one-dimensional nonlinear compressed current response sequence. The specific implementation process of the nonlinear dual-frequency cross-spectral compression algorithm is as follows: Based on the logistic normalization function, a compression kernel matrix is ​​constructed, where any element of the compression kernel matrix... Represented as:

[0037]

[0038] in, It is a frequency interaction weight, representing the excitation frequency. With excitation frequency The coupling strength between the corresponding current response signals is used for compression mapping and is an element of the compression kernel matrix; It is the first Each excitation potential corresponds to a frequency, i.e., the excitation frequency; It is the square of the frequency difference, used to measure the frequency domain distance between excitation frequencies; This is the frequency compression sensitivity coefficient, used to control the influence of frequency differences on the compression result. It is determined through cross-validation and is measured in Hz. The reference value range is [value missing]. The value can be 500. Cross-validation is a well-known technique in the field of science and will not be elaborated here. , These represent the excitation frequency, respectively. and Next, time The current response signal collected at the location; , These are the first derivatives of the current response signal, representing the values ​​at the excitation frequency and the excitation frequency, respectively. and The rate of change of the current response signal over time can reflect the redox dynamic trend, and is obtained through numerical difference calculation. It is the squared term of the normalized frequency difference, essentially a frequency spatial distance factor of the Gaussian kernel function. Its physical meaning is to determine the excitation frequency when constructing a frequency interaction weight. and "Proximity" in the frequency domain; It is a concerted term of the current derivative, reflected in time. At two different excitation frequencies and Whether the electrochemical response changes under different excitation frequencies are simultaneous, rapid and consistent in direction, thus achieving a measure of the consistency and synergy of the reaction trends of flavonoids under different excitation frequencies; It is a criterion for weighting, used to control the compressibility between frequency pairs, and is subsequently converted into the final frequency interaction weights by the sigmoid function;

[0039] Furthermore, the time-frequency response signal matrix is ​​weighted in the frequency dimension using a compression kernel matrix, and a spectral compressed representation is generated to obtain a one-dimensional nonlinear compressed current response sequence:

[0040]

[0041] in, It is a compressed and fused current response time series, i.e., a one-dimensional nonlinear compressed current response series, which is a one-dimensional time function obtained by fusing the current response signals under all frequency channels according to the frequency interaction weight. It is the total number of excitation frequencies; At the excitation frequency The response signal matrix below;

[0042] The significance of the above formula lies in mapping the time-frequency dimension response signal matrix into a one-dimensional nonlinear compressive current response sequence. The value at each time point is a combination of the autocorrelation and cross-correlation behaviors among multiple frequencies. It retains the physical response characteristics while having the ability to remove unstructured interference, serving as the input basis for subsequent singular steady-state extraction.

[0043] Based on one-dimensional nonlinear compression current response sequence To identify the instantaneous current jump point during a redox reaction, i.e., the steady-state jump point, the phenomenon of transiently high response current generated by most flavonoids under specific excitation is crucial. This phenomenon is typically extremely short-lived and easily masked by conventional smoothing. Therefore, a variable window integral singular response identification mechanism is introduced, using the variable window integral method to calculate the singular response intensity at each time point. The integral formula is as follows:

[0044]

[0045] in, It is the first The steady-state jump response score for each time segment, i.e., the singular response intensity; It is the current starting time of integration, indicating the time of the first integration. Each sampling time; This is the length of the integration time window, used to enclose a local signal segment. It is extracted based on the rise time determined by the extreme derivative. The reference value range is... The value here can be 0.5; It is in time The one-dimensional nonlinear compression current response at the point; It is a time integral variable; It is the first derivative of the one-dimensional nonlinear compressive current response with respect to time, i.e., the rate of change of current, which is calculated through numerical difference. This represents the maximum value of the one-dimensional nonlinear compressed current response across the entire signal segment, used for normalization. This is the exponential decay coefficient, used to control the weight decay rate at distant time points. It is automatically optimized using the gradient descent method, and the reference value range is [value range missing]. The value here can be 5; It is a time decay kernel, used to weaken the contribution far from the center time point, so as to ensure that the contribution is greater the closer to the start of integration; It is the local stationarity coefficient. When the current changes slowly, this term is close to 1, and when it changes abruptly, this term approaches 0. It is a normalized periodic sine map term used to detect the response of periodic or short-term surge signals; It is a time-domain gradient suppression term used to suppress sudden disturbances at points of drastic change.

[0046] S2. Based on singular response intensities, a singular response intensity sequence is constructed to determine the singular response occurrence points and obtain singular response weighting coefficients; a local window tensor is constructed, and combined with the singular response weighting coefficients, a singular weighted fusion tensor is obtained; an inversion mechanism based on the mapping between current response gradient and time delay is introduced to asynchronously reconstruct the singular weighted fusion tensor to obtain the feature representation of the flavonoid target after asynchronous reconstruction; based on the feature representation of the flavonoid target after asynchronous reconstruction, intelligent analysis of flavonoid compounds is performed.

[0047] Based on singular response intensity, local maximum detection algorithms, such as non-maximum suppression with a suppression threshold, are used to analyze singular response intensity sequences. The process is performed to obtain a set of maxima, and the maxima greater than a threshold are then selected. The point is taken as the point where the singular response occurs, and the threshold is... Based on the average intensity of historical singular responses, the total number of singular response occurrence points is denoted as... ;

[0048] The sum of the singular response intensities corresponding to all singular response occurrence points is used as a normalization term. This normalization process is then applied to the singular response intensities corresponding to each singular response occurrence point to obtain the singular response weighting coefficients. And ultimately generate a singular weighted fusion tensor:

[0049]

[0050] in, At a certain point in time The feature representation tensor after fusion is the singular weighted fusion tensor; It is the index of the point where the singular response occurs; Therefore, the first The features of the one-dimensional nonlinear compressed current response corresponding to each singular response point are extracted, including the current response signal, rate of change (first derivative), local energy, peak symmetry index, positive and negative direction jump markers, etc. The length of the local window is obtained by calculating half the ratio of the total length of the sampling time series to the total number of singular response points and rounding down.

[0051] Furthermore, to avoid the problem of different flavonoids initiating responses at different times—that is, due to the time delay of electrochemical reactions, different substances may respond with different delays after excitation—asynchronous reconstruction processing of the aforementioned singular weighted fusion tensor is required. An inversion mechanism based on the mapping between current response gradient and time delay is introduced. Sliding integral reconstruction is performed at each time step to obtain asynchronous response window vectors for various flavonoid targets. Combined with existing attention mechanisms, the feature representations of flavonoid targets after asynchronous reconstruction are obtained. The specific calculation formula for the asynchronous response window vectors of flavonoid targets is as follows:

[0052]

[0053] in, At a certain point in time No. The asynchronous response window vector for flavonoid targets, in physical terms, represents the response window vector considering the reaction initiation delay. Under the premise of [missing information], the time-polymerization inversion value of the electrochemical expression of flavonoid target at a certain moment; It is the first The estimated reaction start-up time delay for flavonoid target compounds was determined through backpropagation optimization, with a reference range of [value missing]. The possible value is 1.0; This is the time diffusion scaling factor, used to control the sharpness or smoothness of the integral kernel function within the time window. It is determined based on the existing kernel-weighted moving integral model, and the reference value range is [value missing]. The possible value is 0.1; It is the weighting function in the integral kernel, used to simulate the maximum response of flavonoids near the reaction initiation time, to ensure decay at a more distant time. It is a one-dimensional nonlinear compressive current response at time point The rate of change of current, used to reflect the drasticness of current change, is determined by numerical difference; It is a gradient penalty term, which avoids gradient explosion caused by drastic current jumps by performing a logarithmic smoothing transformation on the rate of change of current. At a certain point in time The singular weighted fusion tensor on; It is the weighted core part of the asynchronous time-compensated integral structure, which describes the electrochemical response energy of the flavonoid target near its characteristic reaction initiation time.

[0054] Furthermore, based on the asynchronous response window vectors of flavonoid targets, an existing attention mechanism is used to determine the association weights between the two classes of flavonoid targets. Then, based on these association weights, the asynchronous response window vectors of the flavonoid targets are weighted and accumulated to obtain the [followed by the previous step]. Flavonoid target at time point Feature representation after asynchronous reconstruction Furthermore, the feature representations of all flavonoid target compounds after asynchronous reconstruction are merged to generate a feature representation tensor in a unified semantic space. The feature representation tensor is then input into an existing inference network, such as a temporal convolutional network, to output classification probabilities and obtain the classification probability of each type of flavonoid target, thereby achieving intelligent analysis of flavonoid compounds.

[0055] In summary, an intelligent analytical method for the electrochemical recognition of flavonoids has been developed.

[0056] The order of the embodiments is for illustrative purposes only and does not represent the superiority or inferiority of the embodiments. The processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0057] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0058] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A smart analytical method for the electrochemical recognition of flavonoids, characterized in that, Includes the following steps: S1. Obtain the response signal matrix in the time-frequency dimension, and introduce a nonlinear dual-frequency cross-spectral compression algorithm to perform frequency domain compression and nonlinear interactive mapping. Based on the response signal matrix in the time-frequency dimension, construct a current derivative synergistic term to quantify the consistency and synergy of the reaction trends of flavonoids under different excitation frequencies. Combine the frequency domain distance between excitation frequencies to calculate the frequency interaction weight and construct a compression kernel matrix. Based on the compression kernel matrix, the response signal matrix in the time-frequency dimension is weighted in the frequency dimension to generate a spectral compression representation, resulting in a one-dimensional nonlinear compressed current response sequence. Based on a one-dimensional nonlinear compressed current response sequence, a variable window integral singular response identification mechanism is introduced to construct local stationarity coefficients and normalized periodic sine mapping terms, and the steady-state jump response score of time segments is calculated to obtain the singular response intensity. S2. Based on singular response intensities, a singular response intensity sequence is constructed to determine the singular response occurrence points and obtain singular response weighting coefficients. A local window tensor is constructed, and combined with the singular response weighting coefficients, a singular weighted fusion tensor is obtained. An inversion mechanism based on the mapping between current response gradient and time delay is introduced. Based on the singular weighted fusion tensor, a time diffusion scale factor is introduced, and combined with a gradient penalty term, a sliding integral reconstruction is performed for each time step to obtain the asynchronous response window vector of the flavonoid target. Based on the asynchronous response window vector of the flavonoid target, combined with an attention mechanism, the correlation weights between flavonoid target compounds are obtained. Based on the correlation weights between flavonoid target compounds, the asynchronous response window vectors of the flavonoid target compounds are weighted and accumulated to obtain the feature representation of the flavonoid target compounds after asynchronous reconstruction. Based on the feature representation of the flavonoid target compounds after asynchronous reconstruction, intelligent analysis of flavonoid compounds is performed.

2. The intelligent analysis method for electrochemical recognition of flavonoids according to claim 1, characterized in that, S2 specifically includes: The singular response intensity sequence is processed to obtain a set of maxima. Based on the set of maxima and a preset threshold, the occurrence point of the singular response is determined.

3. The intelligent analysis method for electrochemical recognition of flavonoids according to claim 2, characterized in that, S2 specifically includes: After normalizing the singular response intensity corresponding to the singular response occurrence point, singular response weighting coefficients are obtained. Based on the singular response weighting coefficients, the local window tensor is weighted and fused to obtain the singular weighted fused tensor. The local window tensor is obtained by feature extraction of the one-dimensional nonlinear compressive current response corresponding to the singular response occurrence point.

4. The intelligent analysis method for electrochemical recognition of flavonoids according to claim 1, characterized in that, S2 specifically includes: The feature representations of all flavonoid targets after asynchronous reconstruction are merged to generate a feature representation tensor in a unified semantic space. The feature representation tensor is then input into the inference network to output the classification probability of each type of flavonoid target, thereby realizing intelligent analysis of flavonoid compounds.

Images