A method and system for evaluating the interaction effects of traditional Chinese medicine and feed.

Abstract

This invention provides a method and system for evaluating the interaction effects of traditional Chinese medicine and feed formulations, relating to the field of computer technology. The method includes: acquiring first information; performing data association and fusion based on the first information to obtain an integrated feature set; constructing an interaction network based on the integrated feature set to obtain a compatibility effect network; perturbing and reconstructing the network based on the compatibility effect network to obtain a reconstruction result; predicting effects based on the reconstruction result, and obtaining a robustness score by comparing the state changes of network output nodes before and after perturbation and the network's ability to recover to a stable state; and performing an integrated evaluation based on the robustness score, outputting an evaluation result by combining the initial network effect strength with the robustness score after perturbation. This invention can systematically and quantitatively predict and evaluate the intrinsic interaction effects of different formulation schemes and their reliability in practical applications.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, in particular, to a Chinese herbal medicine-feed compatibility interaction effect evaluation method and system. BACKGROUND

[0002] In the field of livestock breeding and feed science, the development and application of Chinese herbal medicine as a feed additive is increasingly valued. Its purpose is to regulate animal physiological functions through natural active ingredients and replace part of antibiotics, thereby promoting healthy breeding. However, Chinese herbal medicine has complex components, and when it is mixed with basic feed, complex interactions may occur between multiple active ingredients and feed nutrients. The final effect of this interaction on the animal body (synergy, antagonism or masking) is difficult to predict and evaluate, which is a key technical bottleneck restricting its scientific and precise application.

[0003] Currently, the evaluation methods in this field mainly rely on observing changes in terminal phenotype indicators in animal feeding experiments or using in vitro cell models to screen the activity of a single component. These methods usually treat Chinese herbal medicine as a whole or only focus on individual components, ignoring the complex network relationships between multiple components, multiple nutrients and multiple physiological targets in the compatibility system. Such existing methods have obvious limitations: on the one hand, they are difficult to analyze the interaction mechanism between different input factors from a system level, and on the other hand, they cannot quickly and cost-effectively predict the effect and evaluate the stability of a large number of potential compatibility combinations before the experiment. Especially lacking the ability to quantitatively evaluate the stability of the expected effect of the compatibility scheme under conditions such as simulation of nutrient fluctuations or changes in herbal proportion, leading to long research and development cycle, high cost, and limited depth and reliability in guiding practical application.

[0004] Based on the above-mentioned shortcomings of the prior art, there is an urgent need for a Chinese herbal medicine-feed compatibility interaction effect evaluation method and system. SUMMARY

[0005] The purpose of the present application is to provide a Chinese herbal medicine-feed compatibility interaction effect evaluation method and system to improve the above-mentioned problems. In order to achieve the above-mentioned purpose, the technical solution adopted by the present application is as follows:

[0006] In a first aspect, the present application provides a Chinese herbal medicine-feed compatibility interaction effect evaluation method, comprising:

[0007] Obtaining first information, the first information including effective component content data of each Chinese herbal medicine under different Chinese herbal medicine compatibility schemes, nutritional component data of basic feed, and time series monitoring data of animal growth and health indicators after feeding different compatibility schemes;

[0008] According to the first information, data association and fusion are performed to obtain an integrated feature set;

[0009] According to the integrated feature set, an interaction network is constructed, the interaction intensity between different Chinese herbal medicine components, the interaction intensity between Chinese herbal medicine components and feed nutrients in influencing animal physiological indexes are quantified, and a compatibility effect network is obtained;

[0010] According to the compatibility effect network, network disturbance and reconstruction are performed, the increase or decrease of specific feed components or the adjustment of the proportion of Chinese herbal medicine compatibility is simulated, the propagation process of disturbance in the effect network is simulated, and the network connection is reconstructed, and a reconstruction result is obtained;

[0011] According to the reconstruction result, effect prediction is performed, the state change of the network output node before and after the disturbance is compared, and the ability of the network to recover to a stable state is compared, and a robustness score result is obtained;

[0012] According to the robustness score result, integrated evaluation is performed, the initial network effect intensity and the robustness score after disturbance are comprehensively evaluated, and an evaluation result is output.

[0013] In a second aspect, the application also provides an evaluation method and system for the compatibility interaction effect of Chinese herbal medicine and feed, comprising:

[0014] An acquisition module is configured to acquire first information, wherein the first information comprises effective component content data of each Chinese herbal medicine under different Chinese herbal medicine compatibility schemes, nutritional component data of basic feed, and time sequence monitoring data of animal growth and health indexes after feeding different compatibility schemes;

[0015] A fusion module is configured to perform data association fusion according to the first information, and obtain an integrated feature set;

[0016] A construction module is configured to perform interaction network construction according to the integrated feature set, quantify the interaction intensity between different Chinese herbal medicine components, the interaction intensity between Chinese herbal medicine components and feed nutrients in influencing animal physiological indexes, and obtain a compatibility effect network;

[0017] A reconstruction module is configured to perform network disturbance and reconstruction according to the compatibility effect network, simulate the propagation process of disturbance in the effect network by adjusting the increase or decrease of specific feed components or the proportion of Chinese herbal medicine compatibility, and reconstruct the network connection, and obtain a reconstruction result;

[0018] A prediction module is configured to perform effect prediction according to the reconstruction result, compare the state change of the network output node before and after the disturbance, and the ability of the network to recover to a stable state, and obtain a robustness score result;

[0019] An output module is configured to perform integrated evaluation according to the robustness score result, comprehensively evaluate the initial network effect intensity and the robustness score after disturbance, and output an evaluation result.

[0020] The application has the following beneficial effects:

[0021] This invention achieves a leap from static effect prediction to dynamic robustness assessment by constructing a compatibility effect network that reflects the complex relationship between Chinese herbal medicine and feed components, and simulating its dynamic reconstruction process and system recovery capability under component perturbation. This enables the systematic and quantitative prediction and evaluation of the intrinsic interaction effects of different compatibility schemes and their reliability in practical applications. Attached Figure Description

[0022] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart illustrating a method for evaluating the interaction effects of traditional Chinese medicine and feed, as described in an embodiment of the present invention.

[0024] Figure 2 This is a schematic diagram of the structure of an evaluation system for the interaction effect of traditional Chinese medicine and feed as described in an embodiment of the present invention;

[0025] Figure 3 This is a schematic diagram of the structure of an evaluation device for the interaction effect of traditional Chinese medicine and feed as described in an embodiment of the present invention.

[0026] The diagram is labeled as follows: 800, an evaluation device for the interaction effect of traditional Chinese medicine and feed; 801, processor; 802, memory; 803, multimedia component; 804, I / O interface; 805, communication component; 901, acquisition module; 902, fusion module; 903, construction module; 904, reconstruction module; 905, prediction module; 906, output module. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0028] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0029] Example 1

[0030] This embodiment provides a method for evaluating the interaction effects of traditional Chinese medicine and feed.

[0031] See Figure 1 The figure shows that the method includes steps S100 to S600.

[0032] Step S100: Obtain first information, which includes data on the effective component content of each herb under different herbal compatibility schemes, nutritional component data of basic feed, and time-series monitoring data of animal growth and health indicators after feeding different compatibility schemes.

[0033] Specifically, as shown in Table 1, the data on the effective components of traditional Chinese medicine refers to the quantitative data of key active ingredients (such as polysaccharides, flavonoids, alkaloids, etc.) in each herb under a specific formulation, determined by analytical techniques such as high-performance liquid chromatography. The nutritional component data of the basic feed comes from the actual test results of the feed formula, including the precise content of conventional and micronutrients such as crude protein, crude fat, amino acids, vitamins, and minerals. The time-series monitoring data of animal growth and health indicators are continuous records of growth performance data such as body weight gain and feed conversion ratio, collected periodically from different groups of animals during the feeding cycle under strict feeding trial conditions, as well as health status data such as serum biochemical indicators, immunoglobulin content, and intestinal tissue morphology. The simultaneous acquisition of these three types of data ensures that the causal chain from chemical input to animal biological response can be traced and quantified.

[0034] Table 1. Examples of Traditional Chinese Medicine-Feed Formulation Schemes and Basic Data

[0035]

[0036] Step S200: Perform data association and fusion based on the first information to obtain an integrated feature set;

[0037] Understandably, this step aims to address the integration of multi-source heterogeneous data. The key lies not in simple data splicing, but in uncovering the inherent, biologically significant relationships between data from different dimensions. By causally aligning time-series animal physiological responses with static formulation component data—for example, correlating changes in blood parameters at specific post-feeding time points with the theoretical peak absorption period of the components—static data is transformed into a new dataset containing time-dynamic causal attributes. This provides a feature foundation for subsequently building models that reflect real physiological processes.

[0038] Step S300: Construct an interaction network based on the integrated feature set. By quantifying the interaction strength between different Chinese herbal medicine components and between Chinese herbal medicine components and feed nutrients in affecting animal physiological indicators, a compatibility effect network is obtained.

[0039] It should be noted that this study treats the components of traditional Chinese medicine and feed nutrients as nodes, and quantifies the complex relationships between them, such as synergy, antagonism, or conditional dependence, which they exhibit by jointly affecting animal physiological indicators, as the connection strength and direction between nodes. This constructs a weighted directed compatibility effect network, which intuitively represents the overall structure and potential action pathway of the compatibility system.

[0040] Step S400: Based on the compatibility effect network, perform network perturbation and reconstruction. By increasing or decreasing specific feed ingredients or adjusting the proportion of Chinese herbal medicines, simulate the propagation process of perturbation in the effect network and reconstruct the network connection to obtain the reconstruction result.

[0041] Understandably, this step introduces a dynamic and robust evaluation perspective, simulating formula fluctuations that may occur in actual production to verify the reliability of the formulation. The principle is to apply a virtual perturbation to the constructed static network and simulate the cascading propagation of this perturbation signal along the network connections, observing its impact on the overall network state. Finally, the connection weights are recalibrated based on the network's response, resulting in a new network state that has undergone adaptive adjustment after "stress testing."

[0042] Step S500: Based on the reconstruction results, predict the effects and obtain the robustness score by comparing the state changes of the network output nodes before and after the perturbation and the network's ability to recover to a stable state.

[0043] It should be noted that this step quantifies the degree of fluctuation (instability) of the expected effect of the matching scheme by comparing the state changes of the key output nodes of the network (representing the core animal physiological indicators) before and after the perturbation; at the same time, it evaluates its inherent buffering and self-organizing ability by simulating the ability and speed at which the network state recovers to the initial equilibrium after the perturbation is removed.

[0044] Step S600: Perform an integrated evaluation based on the robustness score results. By combining the initial network effect strength and the robustness score after perturbation, the evaluation results are output.

[0045] Understandably, by considering the initial effect strength as a "benefit" and the robustness score as a "risk control" indicator, and by integrating the two through multi-objective decision-making logic, a comprehensive evaluation result that reflects both effectiveness and reliability can be output, providing a direct and comprehensive decision-making basis for formula optimization.

[0046] Further, step S200 includes steps S210 to S230.

[0047] Step S210: Align the time-series monitoring data in the first information with the effective ingredient content data and nutrient component data. By associating the time point changes of animal indicators with the metabolic absorption cycle of the feeding components, a dynamic dataset with time causal attributes is obtained.

[0048] Step S220: Perform cross-modal association mining based on the dynamic dataset. By introducing the synergistic or antagonistic logic of traditional Chinese medicine and feed components in biological metabolic pathways as constraints, deep feature association pairs are obtained.

[0049] Step S230: Construct a feature space based on deep feature association pairs. By mapping the association pairs to vectors in the space and adjusting the spatial topology according to biological semantic distance, an integrated feature set is obtained.

[0050] Specifically, the data alignment process in step S210 is a prerequisite for the biological rationality of subsequent analyses. The specific operation involves setting an expected biological effect window for each component based on the metabolic kinetics of different types of components (such as easily absorbed carbohydrates and proteins that require degradation) in the animal body, and dynamically correlating fluctuations in animal indicators (such as blood glucose or hormone levels at specific time points after feeding) with the corresponding component's window period. For example, the peak effect period of rapidly absorbed flavonoids in hawthorn is correlated with short-term changes in antioxidant indicators after feeding, while components like Poria cocos polysaccharides, which require fermentation by intestinal microorganisms to be effective, are correlated with medium- to long-term changes in immune indicators. This processing integrates previously independent time-series data into a dynamic dataset with causal direction, meaning that changes in effects at a certain time point are more likely to be driven by components that have entered their effect window.

[0051] Based on this, step S220, cross-modal association mining, further introduces prior knowledge from the fields of traditional Chinese medicine and animal feed science as constraints on this dynamic dataset to uncover deep feature association pairs with physiological significance. The "synergistic or antagonistic logic" is the core constraint here. For example, given that calcium ions may affect the absorption of certain flavonoids, when mining the relationship between "hawthorn flavonoids" and "blood calcium," the existence of antagonistic patterns will be prioritized. Similarly, based on the physiological knowledge that "astragalus polysaccharides and amino acids synergistically promote protein synthesis," when analyzing the combined effect of "astragalus polysaccharides" and "feed lysine" on "daily weight gain," the focus will be on finding signals of synergistic enhancement. This method avoids purely mathematical spurious correlations, ensuring that the uncovered "feature association pairs" (such as the ternary combination of "astragaloside A-vitamin E-serum IgG") are backed by reasonable biological pathway hypotheses.

[0052] Step SS230, feature space construction, aims to transform these abstract association pairs into structured data that can be computed by the network model. This process maps each association pair to a feature vector in a high-dimensional space, where the dimension of the vector represents different attributes of the association (such as association strength, direction, and specificity). Crucially, the spatial topology is "adjusted based on biological semantic distance," meaning that the similarity of the biological functions they represent is considered when arranging the positions of these vector points. For example, all feature vectors associated with "gut health" indicators will cluster in close proximity, while vectors associated with "immune regulation" will form another cluster. This semantic-based topological structure ensures that the final integrated feature set not only contains quantitative relationships but also embeds the domain's knowledge structure.

[0053] Further, step S300 includes steps S310 to S330.

[0054] Step S310: Evaluate the influence of network nodes based on the integrated feature set. By analyzing the contribution and independence of each herbal medicine component and nutrient to the trend of changes in animal physiological indicators, a set of key nodes is selected.

[0055] Step S320: Define the interaction relationship between nodes based on the set of key nodes. By calculating the functional synergy, effect antagonism, or conditional dependence of any two nodes when affecting the same physiological indicator, the strength of the directional interaction relationship between nodes is obtained.

[0056] Step S330: Generate a dynamic network structure based on the strength of the directional interaction relationship. By mapping the interaction relationship to network edges and constructing a weighted directed graph containing multiple feedback loops based on the interaction strength and direction, a matching effect network is obtained.

[0057] Specifically, step S310, the network node influence assessment, is the screening stage for constructing the interaction network. Its purpose is to identify the core factors that truly have a significant impact on animal physiological indicators from numerous traditional Chinese medicine components and feed nutrients. This process comprehensively analyzes the contribution of each component to the changing trends of multiple physiological indicators, with particular consideration of the independence of their effects. For example, a certain traditional Chinese medicine component may have a strong positive contribution to daily weight gain, but its effect may be highly covariant with the content of a certain amino acid in the feed, and its independent contribution will be reduced. Conversely, a trace component may not have a high contribution on its own, but its unique influence pattern and high potential for synergistic or antagonistic effects with other factors may lead to its retention due to its uniqueness. Through this assessment that balances contribution strength and independence, a set of "key action nodes" that is both critical and representative can be screened, thereby significantly simplifying the complexity of subsequent networks and focusing on core interactions.

[0058] After identifying the key nodes, the interaction definition process in step S320 aims to precisely characterize the intrinsic connections between any two nodes. Its core lies in identifying more complex patterns such as "functional synergy," "effect antagonism," or "conditional dependence" between nodes. For example, by analyzing the combined effect of "Astragalus polysaccharide" and "Vitamin E" on increasing "serum IgG," if the combined effect of both is significantly greater than the sum of their individual effects, it is considered functional synergy; if the presence of "andrographolide" significantly weakens the growth-promoting effect of a certain feed component, it is considered effect antagonism; if the improvement effect of "total hawthorn flavonoids" on "feed conversion ratio" only becomes apparent when "dietary fiber" reaches a certain level, it is considered conditional dependence. By performing this intensity calculation based on biological effect patterns on all pairs of key nodes, the strength of directional interaction relationships with clear biological orientations is obtained.

[0059] Functional synergy refers to the effect of two components acting together being greater than the sum of their individual effects. The formula for calculating the strength of functional synergy is:

[0060] ;

[0061] In the formula, S syn E represents the quantitative value of functional synergy strength. A E represents the effect value of herbal ingredient A on the target physiological indicator when acting alone; B This indicates the effect value of feed nutrient B acting alone.

[0062] Effect antagonism refers to the weakening of the effect of one component by another. The formula for calculating the strength of effect antagonism is:

[0063] ;

[0064] In the formula, S ant This represents the quantitative value of the antagonistic strength of the effect.

[0065] Conditional dependence refers to the phenomenon where the effect of one component is only apparent when the effect of another component reaches a certain level. The formula for calculating the strength of conditional dependence is:

[0066] ;

[0067] In the formula, This indicates the effect value of herbal medicine component A when the level of feed nutrient B is greater than or equal to the threshold T. This represents the effect value of herbal ingredient A when the level of feed nutrient B is less than or equal to the threshold T; T is the threshold, representing the key level of feed nutrient B (e.g., dietary fiber at 50 g / kg), which is set by domain knowledge.

[0068] Step S330, the generation of the dynamic network structure, concretizes the abstract interaction relationships into a computable and simulable system model. This process treats each key node as a point (vertices) in the network, and quantifies the "directed interaction strength" calculated in step S320 as edges with direction and weights connecting these points. Crucially, the construction of "multiple feedback loops" reflects the complexity of real-world biological systems. For example, node A (a component of a traditional Chinese medicine) may positively influence node B (a physiological indicator), while an increase in the level of node B may in turn inhibit the absorption or efficacy of node A, forming a negative feedback loop. By constructing this weighted directed graph based on the directed interaction strengths of all node pairs, the resulting compatibility effect network is no longer a simple list of components, but a dynamic system structure capable of characterizing the direct, indirect, and cyclical interactions between components.

[0069] Further, step S400 includes steps S410 to S430.

[0070] Step S410: Perform directional perturbation simulation based on the compatibility effect network. By virtually introducing the absence of specific feed components or the dose multiplication of key components of traditional Chinese medicine, an intervention signal is generated for the initial nodes of the network.

[0071] Step S420: Perform disturbance cascade propagation simulation based on the intervention signal, calculate the cumulative change of the state of subsequent nodes by iteratively calculating along the directed edges of the network, and identify the state oscillation or saturation effect caused by the feedback loop to obtain the dynamic response trajectory of the network.

[0072] Step S430: Recompile the network connection weights according to the dynamic response trajectory. Recalibrate the weights of the input connections based on the final steady-state deviation of the node state to obtain a new equilibrium network that reflects the adaptive adjustment of the system as the reconstruction result.

[0073] Specifically, steps S410 to S430 together constitute the core process for dynamic robustness stress testing of the static compatibility effect network. Step S410 is based on specific scenarios that may occur in actual breeding or formulation practices, such as simulating a shortage of a key vitamin in feed or an accidental overdose of a core herbal ingredient. This simulation generates a specific intervention signal by directly changing the initial state values ​​of the corresponding nodes in the network, thereby setting clear initial conditions for subsequent propagation analysis. This makes the evaluation work closely aligned with the risks that may be faced in actual application scenarios. The directional perturbation simulation is expressed as:

[0074] ;

[0075] In the formula, D represents the initial state value of node i (representing a traditional Chinese medicine ingredient or feed nutrient) before it is disturbed; i This represents the perturbation factor applied to node i, with a coefficient preset by the perturbation type, simulating the absence of a component (such as a critical vitamin deficiency), where 0 ≤ D. i ≤1, when simulating dose doubling (e.g., excessive dosage of key components in traditional Chinese medicine), take the corresponding multiplier; D when there is no perturbation. i =1; This represents the change in the state of node i at the start of the simulation.

[0076] Next, the S420 perturbation cascade propagation simulation characterizes the diffusion effect of the intervention signal throughout the network. The process involves iterative calculations along the influence path defined by the directed edges in the network, simulating how the initial perturbation gradually affects the state of downstream nodes and calculating its cumulative effect. This process can identify and handle complex dynamic behaviors caused by "feedback loops" in the network structure. For example, a change in the state of a node may feed back to itself through a loop, potentially causing periodic oscillations in its state or limiting its influence to a certain range until saturation. By simulating this process, the final dynamic response trajectory is no longer a snapshot of a single time point, but a complete record of the entire dynamic process from being impacted by the perturbation to reaching a new transient equilibrium, revealing the vulnerabilities and stabilization mechanisms of the interaction chains within the matching scheme. The perturbation cascade propagation iterative formula is expressed as:

[0077] ;

[0078] In the formula, S j (t) represents the state value of node j (such as an indicator node representing the daily weight gain of an animal) at time step t; S represents the upper limit of saturation for node j; j (t-1) represents the state value of node j at the previous time step t-1; w ijThis represents the connection weight from node i to node j; denoted as t-1, representing the change in the state of node i relative to its initial steady state; Input(j) represents the set of nodes that have direct input connections to all nodes j; η represents the propagation rate coefficient, which is usually taken as a value less than 1, such as 0.1.

[0079] Ultimately, the repetition of network connection weights in S430 aims to quantify the adaptive structural changes that occur in the network after experiencing a perturbation. This is based on the dynamic response trajectory obtained from S420, particularly focusing on the final deviation of each node's state from its initial steady state. The core logic of this process is that a robust network system buffers perturbations through self-adjustment of its internal interactions. Therefore, if a node exhibits a large steady-state deviation after a perturbation, it means that the connection weights pointing to that node may need to be weakened, as this indicates that the original connections amplified the perturbation too strongly; conversely, if a node quickly returns to or approaches its initial state, its input connections may be strengthened, thus playing a stabilizing role. Through this recalibration of connection weights based on the final node state, the resulting new balanced network essentially reflects the adaptive reconstruction of the original matching effect network's internal interactions after undergoing this specific stress test. The adaptive adjustment formula for network connection weights is expressed as:

[0080] ;

[0081] In the formula, This represents the connection weight from node i to node j before the perturbation simulation; This indicates the new weight after weight adjustment; This represents the final deviation of node j from its initial state after the disturbance propagation reaches a new steady state; α represents the initial state value of node j; α represents the adjustment intensity coefficient, which is a positive number and can be 0.05.

[0082] Further, step S500 includes steps S510 to S530.

[0083] Step S510: Measure the steady-state deviation of nodes based on the reconstruction results. By calculating the state differences and variances of nodes corresponding to key animal physiological indicators before and after the disturbance, the degree of instability of network functional output is measured.

[0084] Step S520: Perform system recovery trajectory analysis based on the degree of instability measurement. By simulating the path and rate of node states converging towards the initial steady state along the network structure after the disturbance is removed, the self-organizing recovery capability of the network is evaluated.

[0085] Step S530: Perform robustness integration quantification based on the degree of instability measurement and self-organizing recovery capability assessment. Generate robustness score results by integrating the stability of functional output and the agility of system recovery.

[0086] Specifically, the focus of the node steady-state deviation quantification process in step S510 is to evaluate the difference in state values ​​of network output nodes representing key animal physiological indicators (such as daily weight gain, feed conversion ratio, and immune level) before and after undergoing perturbation simulation. This difference calculation not only focuses on the mean magnitude of the deviation, but also characterizes the volatility of the output by analyzing its variance. This expands the impact on the system function from a single numerical value to a measure of instability that includes volatility characteristics. This quantitative result reflects the initial ability of the compatibility scheme to maintain the stability of the expected effect, that is, the attribute of the system to resist perturbation and maintain static stability.

[0087] Based on the quantification of the unstable state, the S520 system recovery trajectory analysis further explores the network's dynamic self-organizing capability. This process removes the applied disturbance signal in the model and simulates how the state of each node traces back from the disturbed state to the initial equilibrium point according to the inhibition, promotion, and feedback relationships existing in the current network structure. By analyzing the smoothness of this convergence path (such as whether repeated oscillations occur) and the convergence rate, the efficiency of the network's internal interaction mechanism in self-correcting and returning to normal is evaluated, thereby obtaining an assessment of the self-organizing recovery capability. This qualitative capability is transformed into a quantitative index about recovery speed and path stability.

[0088] Ultimately, the robustness integration quantification process of S530 aims to fuse the static stability and dynamic resilience attributes obtained from the first two steps. Its core logic is that an ideal compatibility scheme should simultaneously possess a low degree of instability (indicating strong anti-interference ability) and good resilience (indicating good elasticity). This process establishes a trade-off mechanism to map the instability measurement and resilience assessment into a unified decision framework. For example, it assigns a weight to the ability to resist initial disturbances as a robust foundation, while assigning a weight to the ability to recover quickly as a resilient adaptation. Finally, a single robustness score is generated through a comprehensive algorithm. This score no longer considers a single characteristic of the system in isolation, but rather makes an integrated judgment on its overall robustness.

[0089] Further, step S600 includes steps S610 to S630.

[0090] Step S610: Quantify the initial performance baseline based on the robustness score and the compatibility effect network. By calculating the magnitude of the state vector of the network output node before the introduction of disturbance, the initial effect strength characterizing the theoretical expected effect of the compatibility scheme is obtained.

[0091] Step S620: Perform multi-dimensional performance fusion based on the initial effect strength and robustness score results. Construct a decision plane with effect strength as expected return and robustness score as risk coefficient, and locate the Pareto front position of the matching scheme in the plane to obtain the comprehensive performance index.

[0092] Step S630: Perform application-oriented level mapping based on the comprehensive performance index. By discretizing the continuous performance index intervals and mapping them to predefined evaluation levels, and associating them with decision-making suggestions under different livestock production scenarios, the evaluation results are output.

[0093] Specifically, the initial efficacy baseline quantification process in step S610 first establishes a benchmark for efficacy evaluation; using the undisturbed original compatibility effect network constructed in step S300, the overall effect level that the compatibility scheme can achieve under ideal conditions is comprehensively characterized by calculating the magnitude of the vector composed of all output node states representing key animal physiological indicators; this initial effect strength is a theoretical efficacy peak, reflecting the maximum potential of the scheme.

[0094] After obtaining the performance benchmark and robustness score, the S620's multi-dimensional performance fusion processing directly addresses the core contradiction in decision-making—that efficient solutions may be unstable, while stable solutions may have mediocre results. This processing constructs the decision problem as a two-dimensional optimization space, where the "initial effect strength" is considered the expected return, and the reciprocal of the "robustness score" is treated as a risk coefficient. By identifying the distribution of all matching solutions to be evaluated in this decision plane and finding those Pareto optimal solutions that cannot be comprehensively surpassed by other solutions in any dimension, a "comprehensive performance index" that balances returns and risks is calculated for each solution. This index objectively reflects the comprehensive competitiveness of the solution in a real, complex environment. The formula for calculating the comprehensive performance index is:

[0095] ;

[0096] In the formula, PI k I represents the overall performance index of the k-th matching scheme; k R represents the initial effect strength of the k-th option; k I represents the robustness score of the k-th solution; max I min R represents the maximum and minimum initial effect strengths among all options; max R min This represents the maximum and minimum robustness scores among all options; denoted as the average of the initial effect strength and robustness scores of all options; w1 and w2 are the weighting coefficients of effectiveness and robustness, respectively, w1+w2=1; γ is the effectiveness sensitivity coefficient, γ>0; λ is the risk aversion coefficient, λ>0; δ is the synergy enhancement coefficient, 0≤δ<1, used to control the amplification of the synergistic effect between effectiveness and robustness. It is the hyperbolic tangent function.

[0097] Ultimately, the S630's application-oriented level mapping process transforms continuous indices into actionable guidance. Instead of simple sorting, it presets different performance index threshold ranges based on different priorities in livestock production (e.g., fattening farms pursuing maximum growth efficiency versus breeding farms focusing on stable herd health), and maps each solution to specific levels such as "preferred," "suitable for certain conditions," and "use with caution." At the same time, the output results are associated with specific decision-making recommendations, such as noting "recommended for use in farms with meticulous nutrition management" for solutions that are highly efficient but less robust, thus achieving the transformation from complex calculations to clear application guidelines.

[0098] Example 2

[0099] like Figure 2 As shown, this embodiment provides an evaluation system for the interaction effects of traditional Chinese medicine and feed formulation. The system includes:

[0100] The acquisition module 901 is used to acquire first information, which includes the effective component content data of each Chinese herbal medicine under different Chinese herbal medicine compatibility schemes, the nutritional component data of the basic feed, and the time-series monitoring data of animal growth and health indicators after feeding different compatibility schemes.

[0101] The fusion module 902 is used to perform data association and fusion based on the first information to obtain an integrated feature set;

[0102] Module 903 is used to construct an interaction network based on the integrated feature set. By quantifying the interaction strength between different Chinese herbal medicine components and between Chinese herbal medicine components and feed nutrients in affecting animal physiological indicators, a compatibility effect network is obtained.

[0103] The reconstruction module 904 is used to perform network perturbation and reconstruction based on the compatibility effect network. By increasing or decreasing specific feed ingredients or adjusting the proportion of Chinese herbal medicines, it simulates the propagation process of perturbation in the effect network and reconstructs the network connection to obtain the reconstruction result.

[0104] The prediction module 905 is used to predict the effect based on the reconstruction result. By comparing the state changes of the network output nodes before and after the perturbation and the network's ability to recover to a steady state, a robustness score is obtained.

[0105] Output module 906 is used to perform integrated evaluation based on robustness score results. It outputs evaluation results by combining the initial network effect strength and the robustness score after perturbation.

[0106] In one specific embodiment of this application, the fusion module 902 includes:

[0107] The first fusion unit is used to align the time-series monitoring data in the first information with the effective ingredient content data and nutrient component data. By associating the time point changes of animal indicators with the metabolic absorption cycle of feeding components, a dynamic dataset with time causal attributes is obtained.

[0108] The second fusion unit is used to perform cross-modal association mining based on dynamic datasets. It introduces the synergistic or antagonistic logic of the biological metabolic pathways of Chinese herbal medicine and feed ingredients as constraints to obtain deep feature association pairs.

[0109] The third fusion unit is used to construct a feature space based on deep feature association pairs. By mapping the association pairs to vectors in the space and adjusting the spatial topology according to the biological semantic distance, an integrated feature set is obtained.

[0110] In one specific embodiment of this application, the construction module 903 includes:

[0111] The first building unit is used to evaluate the influence of network nodes based on the integrated feature set. By analyzing the contribution and independence of each Chinese herbal medicine component and nutrient to the trend of changes in animal physiological indicators, a set of key nodes is screened.

[0112] The second building unit is used to define the interaction relationship between nodes based on the set of key nodes. By calculating the functional synergy, effect antagonism or conditional dependence of any two nodes when affecting the same physiological index, the strength of the directional interaction relationship between nodes is obtained.

[0113] The third building unit is used to generate a dynamic network structure based on the strength of directional interaction relationships. By mapping the interaction relationships to network edges and constructing a weighted directed graph containing multiple feedback loops based on the interaction strength and direction, a matching effect network is obtained.

[0114] In one specific embodiment of this application, the reconstruction module 904 includes:

[0115] The first reconstruction unit is used to perform directional perturbation simulation based on the compatibility effect network. By virtually introducing the absence of specific feed components or the dose multiplication of key components of traditional Chinese medicine, intervention signals are generated for the initial nodes of the network.

[0116] The second reconstruction unit is used to simulate the cascade propagation of disturbances based on the intervention signal. It calculates the cumulative changes in the state of subsequent nodes by iteratively calculating along the directed edges of the network and identifies state oscillations or saturation effects caused by feedback loops to obtain the dynamic response trajectory of the network.

[0117] The third reconstruction unit is used to re-establish the network connection weights based on the dynamic response trajectory. By recalibrating the weights of the input connections based on the final steady-state deviation of the node state, a new equilibrium network reflecting the adaptive adjustment of the system is obtained as the reconstruction result.

[0118] In one specific embodiment of this application, the prediction module 905 includes:

[0119] The first prediction unit is used to measure the steady-state deviation of nodes based on the reconstruction results. By calculating the state differences and variances of nodes corresponding to key animal physiological indicators before and after the disturbance, the degree of instability of network functional output is measured.

[0120] The second prediction unit is used to perform system recovery trajectory analysis based on the degree of instability. By simulating the path and rate of node states converging towards the initial steady state along the network structure after the disturbance is removed, the self-organizing recovery capability of the network is evaluated.

[0121] The third prediction unit is used to perform robustness integration quantification based on instability measurement and self-organizing recovery capability assessment. It generates a robustness score by fusing two dimensions: the stability of the functional output and the agility of system recovery.

[0122] In one specific embodiment of this application, the output module 906 includes:

[0123] The first output unit is used to quantify the initial performance baseline based on the robustness score and the compatibility effect network. By calculating the magnitude of the state vector of the network output node before the introduction of perturbation, the initial effect strength characterizing the theoretical expected effect of the compatibility scheme is obtained.

[0124] The second output unit is used to perform multi-dimensional performance fusion based on the initial effect strength and robustness score results. It constructs a decision plane with the effect strength as the expected return and the robustness score as the risk coefficient, and locates the Pareto front position of the matching scheme in the plane to obtain the comprehensive performance index.

[0125] The third output unit is used to perform application-oriented level mapping based on the comprehensive performance index. It discretizes the continuous performance index range and maps it to a predefined evaluation level, and associates it with decision-making suggestions under different livestock production scenarios to output the evaluation results.

[0126] Example 3

[0127] Corresponding to the above method embodiments, this embodiment also provides an evaluation device for the interaction effect of traditional Chinese medicine-feed formulation. The evaluation device for the interaction effect of traditional Chinese medicine-feed formulation described below and the evaluation method for the interaction effect of traditional Chinese medicine-feed formulation described above can be referred to in correspondence.

[0128] Figure 3 This is a block diagram illustrating an evaluation device 800 for the interaction effects of traditional Chinese medicine and feed formulation, according to an exemplary embodiment. Figure 3 As shown, the device 800 for evaluating the interaction effect of traditional Chinese medicine and feed may include: a processor 801 and a memory 802. The device 800 may also include one or more of the following: a multimedia component 803, an I / O interface 804, and a communication component 805.

[0129] The processor 801 controls the overall operation of the herbal medicine-feed compatibility interaction evaluation device 800 to complete all or part of the steps in the aforementioned herbal medicine-feed compatibility interaction evaluation method. The memory 802 stores various types of data to support the operation of the herbal medicine-feed compatibility interaction evaluation device 800. This data may include, for example, instructions for any application or method operating on the herbal medicine-feed compatibility interaction evaluation device 800, and application-related data such as contact data, sent and received messages, images, audio, video, etc. The memory 802 can be implemented using any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touchscreen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in the memory 802 or transmitted via the communication component 805. The audio component also includes at least one speaker for outputting audio signals. I / O interface 804 provides an interface between processor 801 and other interface modules, such as keyboards, mice, and buttons. These buttons can be virtual or physical. Communication component 805 is used for wired or wireless communication between the herbal medicine-feed compatibility interaction evaluation device 800 and other devices. Wireless communication includes Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, or 4G, or a combination thereof. Therefore, the corresponding communication component 805 may include a Wi-Fi module, a Bluetooth module, or an NFC module.

[0130] In an exemplary embodiment, an evaluation device 800 for the interaction effect of traditional Chinese medicine-feed formulation can be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the aforementioned evaluation method for the interaction effect of traditional Chinese medicine-feed formulation.

[0131] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided. When executed by a processor, these program instructions implement the steps of the above-described method for evaluating the interaction effects of traditional Chinese medicine and feed. For example, the computer-readable storage medium may be the memory 802 including the program instructions, which may be executed by a processor 801 of a device 800 for evaluating the interaction effects of traditional Chinese medicine and feed to complete the above-described method for evaluating the interaction effects of traditional Chinese medicine and feed.

[0132] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for evaluating the effect of Chinese herbal medicine-feed compatibility interaction, characterized in that, The method comprises the following steps: acquiring first information, which comprises effective component content data of each Chinese herbal medicine under different compatibility schemes, nutritional component data of basic feed, and time series monitoring data of animal growth and health indicators after feeding different compatibility schemes; performing data correlation fusion according to the first information to obtain an integrated feature set; performing interaction network construction according to the integrated feature set, quantifying the interaction intensity between different Chinese herbal medicine components and between Chinese herbal medicine components and feed nutrients in influencing animal physiological indicators to obtain a compatibility effect network; performing network perturbation and reconstruction according to the compatibility effect network, simulating the propagation process of perturbation in the effect network and reconstructing network connections by adjusting the increase or decrease of specific feed components or the adjustment of Chinese herbal medicine compatibility ratio to obtain a reconstruction result; performing effect prediction according to the reconstruction result, comparing the state changes of network output nodes before and after perturbation and the ability of the network to recover to a stable state to obtain a robustness score result; performing integrated evaluation according to the robustness score result, outputting an evaluation result by comprehensively evaluating the initial network effect intensity and the robustness score after perturbation; wherein the network perturbation and reconstruction according to the compatibility effect network, simulating the propagation process of perturbation in the effect network and reconstructing network connections by adjusting the increase or decrease of specific feed components or the adjustment of Chinese herbal medicine compatibility ratio to obtain a reconstruction result, comprises: performing directed perturbation simulation according to the compatibility effect network, generating intervention signals for the initial nodes of the network by virtually introducing the absence of specific feed components or the dose multiplication of key components of Chinese herbal medicine; performing perturbation cascade propagation simulation according to the intervention signals, obtaining the dynamic response trajectory of the network by iteratively calculating the cumulative changes of the state of subsequent nodes along the directed edges of the network and identifying the state oscillation or saturation effect caused by feedback loops; performing network connection weight reoriginating according to the dynamic response trajectory, obtaining a new balanced network reflecting the adaptive adjustment of the system as the reconstruction result by recalibrating the weights of the input connections of the nodes according to the final stable state deviation of the node state.

2. The method for evaluating the Chinese herbal medicine-feed compatibility interaction effect according to claim 1, characterized in that, The data correlation fusion according to the first information comprises: aligning the time series monitoring data with the effective component content data and the nutritional component data in the first information, obtaining a dynamic data set with time causal properties by correlating the time point changes of animal indicators with the metabolic absorption period of feeding components; performing cross-modal correlation mining according to the dynamic data set, obtaining deep feature correlation pairs by introducing the synergistic or antagonistic logic of Chinese herbal medicine-feed components in biological metabolic pathways as a constraint condition; performing feature space construction according to the deep feature correlation pairs, obtaining an integrated feature set by mapping the correlation pairs into vectors in space and adjusting the space topology according to biological semantic distance.

3. The method for evaluating the Chinese herbal medicine-feed compatibility interaction effect according to claim 1, characterized in that, The interaction network construction according to the integrated feature set, quantifying the interaction intensity between different Chinese herbal medicine components and between Chinese herbal medicine components and feed nutrients in influencing animal physiological indicators to obtain a compatibility effect network, comprises: According to the integration feature set, network node influence evaluation is performed, by analyzing the contribution degree and independence of each Chinese herbal medicine ingredient and nutrient to the change trend of animal physiological indicators, a key role node set is screened out; According to the key role node set, the interaction relationship between nodes is defined, by calculating the functional synergy, effect antagonism or conditional dependence of any two nodes in influencing the same physiological indicator, the directional interaction relationship strength between nodes is obtained; According to the directional interaction relationship strength, a dynamic network structure is generated, by mapping the interaction relationship to the network edge, and constructing a weighted directed graph containing multiple feedback loops according to the interaction strength and direction, a compatibility effect network is obtained.

4. The method for evaluating Chinese herbal medicine-feed compatibility interaction effects according to claim 1, characterized in that, According to the reconstruction result, effect prediction is performed, by comparing the state change of the network output node before and after the disturbance and the ability of the network to recover to the stable state, a robustness score result is obtained, including: According to the reconstruction result, the node steady state deviation is quantified, by calculating the state difference and variance of the corresponding node of the key animal physiological indicator before and after the disturbance, the instability degree of the network function output is obtained; According to the instability degree, system recovery trajectory analysis is performed, by simulating the path and rate of the node state converging to the initial stable state along the network structure after the disturbance is removed, the self-organizing recovery ability evaluation of the network is obtained; According to the instability degree and the self-organizing recovery ability evaluation, the robustness integration is quantified, by integrating the stability of the functional output and the agility of the system recovery in two dimensions, a robustness score result is generated.

5. A system for evaluating the effect of Chinese herbal medicine-feed compatibility interaction, characterized by, Including: An acquisition module is configured to acquire first information, the first information including effective component content data of each Chinese herbal medicine under different Chinese herbal medicine compatibility schemes, nutrient component data of basic feed, and time sequence monitoring data of animal growth and health indicators after feeding different compatibility schemes; A fusion module is configured to perform data correlation fusion according to the first information to obtain an integrated feature set; A construction module is configured to construct an interaction network according to the integrated feature set, by quantifying the interaction strength between different Chinese herbal medicine components, between Chinese herbal medicine components and feed nutrients in influencing animal physiological indicators, to obtain a compatibility effect network; A reconstruction module is configured to perform network disturbance and reconstruction according to the compatibility effect network, by adjusting the increase or decrease of specific feed components or the proportion of Chinese herbal medicine compatibility, to simulate the propagation process of disturbance in the effect network and reconstruct the network connection, to obtain a reconstruction result; A prediction module is configured to perform effect prediction according to the reconstruction result, by comparing the state change of the network output node before and after the disturbance and the ability of the network to recover to the stable state, to obtain a robustness score result; An output module is configured to perform integrated evaluation according to the robustness score result, by integrating the initial network effect strength and the robustness score after disturbance, to output an evaluation result; The reconstruction module includes: A first reconstruction unit is configured to perform directional disturbance simulation according to the compatibility effect network, by virtually introducing the absence of specific feed components or the dose multiplication of key components of Chinese herbal medicine, to generate an intervention signal to the initial node of the network; A second reconstruction unit is configured to perform disturbance cascade propagation simulation according to the intervention signal, to obtain a dynamic response trajectory of the network by iteratively calculating cumulative changes of subsequent node states along network directed edges, and to identify state oscillation or saturation effects caused by feedback loops. A third reconstruction unit is configured to perform network connection weight recalibration according to the dynamic response trajectory, to obtain a new balanced network reflecting adaptive adjustment of the system as a reconstruction result by recalibrating weights of input connections of node states according to their final steady-state deviation degrees.

6. The system for evaluating Chinese herbal medicine-feed compatibility interaction effects according to claim 5, wherein, The fusion module comprises: A first fusion unit is configured to align time-series monitoring data in the first information with effective component content data and nutritional component data, to obtain a dynamic data set with time causal properties by correlating time point changes of animal indicators with metabolic absorption periods of feeding components; A second fusion unit is configured to perform cross-modal correlation mining according to the dynamic data set, to obtain deep feature correlation pairs by introducing synergy or antagonism logic of Chinese herbal medicine-feed components in biological metabolic pathways as constraint conditions; A third fusion unit is configured to perform feature space construction according to the deep feature correlation pairs, to obtain an integrated feature set by mapping correlation pairs into vectors in space and adjusting space topology according to biological semantic distance.

7. The system for evaluating Chinese herbal medicine-feed compatibility interaction effects according to claim 5, wherein, The construction module comprises: A first construction unit is configured to perform network node influence evaluation according to the integrated feature set, to obtain a key role node set by analyzing contribution degrees and independencies of each Chinese herbal medicine component and nutrient to animal physiological indicator trend changes; A second construction unit is configured to define inter-node interaction relationship according to the key role node set, to obtain directional interaction relationship strength between nodes by calculating functional synergy, effect antagonism or conditional dependency of any two nodes in influencing the same physiological indicator; A third construction unit is configured to generate a dynamic network structure according to the directional interaction relationship strength, to obtain a compatibility effect network by mapping interaction relationships into network edges and constructing a weighted directed graph containing multiple feedback loops according to interaction strength and direction.

8. The system for evaluating Chinese herbal medicine-feed compatibility interaction effects according to claim 5, wherein, The prediction module comprises: A first prediction unit is configured to perform node steady-state deviation quantification according to the reconstruction result, to obtain instability degree measurement of network function output by calculating state differences and variances of corresponding nodes of key animal physiological indicators before and after disturbance; A second prediction unit is configured to perform system recovery trajectory analysis according to the instability degree measurement, to obtain self-organizing recovery capability evaluation of the network by simulating paths and rates of node states converging to the initial steady state along the network structure after disturbance removal; A third prediction unit is configured to perform robustness integrated quantification according to the instability degree measurement and the self-organizing recovery capability evaluation, to generate a robustness score result by fusing stability of function output and agility of system recovery in two dimensions.

Images