Method and system for predicting oxidative stress reaction of livestock and poultry

Abstract

The invention provides a livestock oxidative stress reaction prediction method and system, and the method comprises the steps: obtaining a real-time physiological monitoring data set with a timestamp mark of a livestock group, carrying out the feature extraction of the real-time physiological monitoring data set, and obtaining a target physiological index feature reflecting an oxidative stress state, the target physiological indexes comprise metabolic activity characteristics, oxidation product accumulation characteristics and anti-oxidation response characteristics. Calling a pre-training model to carry out time sequence correlation analysis on the target physiological index features, and generating an oxidative stress state prediction sequence; and determining a target individual with oxidative stress reaction and reaction development stage information according to the prediction sequence, and finally generating an oxidative stress reaction prediction result for the target individual. Therefore, the oxidative stress reaction of the livestock and poultry can be accurately predicted in real time.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and in particular to a method and system for predicting oxidative stress responses in livestock and poultry. Background Technology

[0002] In the livestock and poultry farming industry, oxidative stress is one of the key factors affecting the health and production performance of livestock and poultry. Oxidative stress refers to the process in which the body produces excessive amounts of highly reactive molecules such as reactive oxygen free radicals when exposed to various harmful stimuli, leading to an imbalance between the oxidation and antioxidant systems and resulting in tissue damage.

[0003] Currently, traditional methods for detecting oxidative stress in livestock and poultry mainly rely on periodically collecting blood and tissue samples, followed by complex biochemical analyses in the laboratory to determine relevant oxidative stress indicators, such as malondialdehyde (MDA) and superoxide dismutase (SOD). However, these methods have significant limitations. On the one hand, the sampling process can cause stress to livestock and poultry, potentially affecting the accuracy of the test results, and the limited sampling frequency makes it difficult to monitor the oxidative stress status of livestock and poultry in real time and dynamically. On the other hand, laboratory analysis is costly and time-consuming, failing to provide timely and effective decision-making support for farmers, making it difficult to take effective intervention measures in the early stages of oxidative stress, thereby affecting the growth, reproduction, and economic benefits of livestock and poultry farming. Summary of the Invention

[0004] In view of the aforementioned problems, and in conjunction with the first aspect of the present invention, an embodiment of the present invention provides a method for predicting oxidative stress response in livestock and poultry, comprising: A real-time physiological monitoring dataset for livestock and poultry populations is constructed based on multiple time-stamped physiological parameter units collected continuously. The metabolic activity characteristics, oxidation product accumulation characteristics, and antioxidant response characteristics of the real-time physiological monitoring data set are extracted to obtain target physiological indicators that reflect the oxidative stress state of livestock and poultry. The pre-trained oxidative stress prediction model is invoked to perform time-series correlation analysis on the target physiological indicators to generate a predicted sequence of oxidative stress status for the livestock and poultry population. Based on the oxidative stress state prediction sequence, the target individuals in the livestock and poultry population that exhibit oxidative stress response and the developmental stage information of the target individuals in producing the oxidative stress response are determined. Based on the target individual and the corresponding developmental stage information, an oxidative stress response prediction result is generated for the target individual.

[0005] In some possible implementations, the extraction of metabolic activity characteristics, oxidative product accumulation characteristics, and antioxidant response characteristics from the real-time physiological monitoring data set yields target physiological indicators reflecting the oxidative stress state of livestock and poultry, including: The real-time physiological monitoring data set is divided into time windows to obtain multiple monitoring data segment units with continuous time series relationships; For each monitoring data segment unit, metabolic parameters are screened, and respiratory rate fluctuation characteristics and body temperature change rate characteristics related to energy metabolism are extracted as metabolic activity characteristics. Each monitoring data segment unit is processed for oxidation product detection, and specific biomolecule concentration change features that reflect the degree of free radical accumulation are extracted, and the specific biomolecule concentration change features are used as the oxidation product accumulation features. Each monitoring data segment unit is subjected to defense system analysis and processing to extract the characteristics of antioxidant enzyme activity fluctuations and the characteristics of non-enzymatic antioxidant substance content changes as the antioxidant response characteristics. The metabolic activity characteristics, the oxidation product accumulation characteristics, and the antioxidant response characteristics are analyzed and processed simultaneously to generate the target physiological indicator characteristics.

[0006] In some possible implementations, the metabolic parameter screening process for each of the monitored data segments, extracting respiratory rate fluctuation features and body temperature change rate features related to energy metabolism as the metabolic activity features, includes: Extract the continuous sampled value sequence of respiratory rate parameter from each of the monitoring data segment units, and calculate the absolute value of the frequency difference between adjacent sampling points in the continuous sampled value sequence as the frequency fluctuation amplitude parameter; The frequency fluctuation amplitude parameters of each frequency fluctuation parameter within a preset time window are statistically distributed, and a respiratory frequency fluctuation pattern vector is generated based on the frequency distribution. Extract the continuous sampling value sequence of body temperature parameters from each of the monitoring data segment units, and calculate the amount of body temperature change per unit time in the continuous sampling value sequence as the body temperature change rate parameter; The synchronous correlation between the body temperature change rate parameter and the respiratory rate fluctuation pattern vector was analyzed to generate a metabolic activity synergy index. The respiratory rate fluctuation pattern vector, the body temperature change rate parameter, and the metabolic activity synergy index are integrated into metabolic activity characteristics.

[0007] In some possible implementations, the oxidation product detection processing of each of the monitoring data segments to extract specific biomolecule concentration change features reflecting the degree of free radical accumulation includes: Identify the concentration parameters of biomolecules related to free radical metabolism in each of the monitoring data segment units, and extract the concentration value sequence of continuous sampling; Calculate the mean and standard deviation of the concentration value sequence to generate a concentration distribution stability index; Analyze the trend of the concentration value sequence over time to generate parameters for the direction of concentration change; Extract the number of peak points exceeding the baseline level and the duration of the peak points from the concentration value sequence to generate an abnormal accumulation intensity index; The concentration distribution stability index, the concentration change direction parameter, and the abnormal accumulation intensity index are integrated into an oxidation product accumulation characteristic.

[0008] In some possible implementations, the defense system analysis processing of each monitoring data segment unit, extracting antioxidant enzyme activity fluctuation characteristics and non-enzymatic antioxidant substance content change characteristics as the antioxidant response characteristics, includes: Extract the antioxidant enzyme activity parameters from each monitoring data segment unit to generate a first continuous sampling value sequence, and calculate the difference between the maximum and minimum activity values ​​in the first continuous sampling value sequence as the activity fluctuation range parameter. The percentage of activity values ​​higher than the baseline value in the first continuous sampling value sequence is counted to generate an activity enhancement index. Extract the content parameters of non-enzymatic antioxidants from each monitoring data segment unit to generate a second continuous sampling value sequence, and calculate the linear regression slope of the content values ​​in the second continuous sampling value sequence as the content change rate parameter; The correlation between the rate of change of non-enzymatic antioxidant substances and the fluctuation range of antioxidant enzyme activity was analyzed to generate a synergistic response index for the defense system. The activity fluctuation range parameter, the activity enhancement degree index, the content change rate parameter, and the defense system synergistic response index are integrated into an antioxidant response feature.

[0009] In some possible implementations, the step of invoking a pre-trained oxidative stress prediction model to perform time-series correlation analysis on the target physiological indicator features to generate a predicted sequence of oxidative stress status for the livestock and poultry population includes: The target physiological index features are input into the temporal coding layer of the oxidative stress prediction model, and the time series features are represented by temporal coding through a long short-term memory network to generate a temporal coding representation vector. The oxidative stress prediction model uses a feature interaction layer to perform feature interaction on the temporal encoded representation vector to generate a fusion feature vector that characterizes the association information of metabolic oxidative defense. The state classification layer of the oxidative stress prediction model is used to perform stress state probability prediction processing on the fused feature vector to generate the probability distribution of oxidative stress state at each time point. The probability distribution of the oxidative stress state is smoothed over time by the trend analysis layer of the oxidative stress prediction model to generate a state probability change curve within a continuous time window. The state probability change curve is converted into an oxidative stress state prediction sequence that includes normal state, potential stress state, and overt stress state.

[0010] In some possible implementations, the step of performing feature interaction on the temporal encoded representation vector through the feature interaction layer of the oxidative stress prediction model to generate a fused feature vector for characterizing metabolic oxidative defense association information includes: The temporal coding representation vector is decomposed into metabolic feature vectors, oxidation feature vectors, and defense feature vectors; Perform a dot product operation on the metabolic feature vector and the oxidation feature vector to generate a correlation strength parameter between metabolism and oxidation. The difference operation is performed on the oxidation feature vector and the defense feature vector to generate the antagonistic strength parameter between oxidation and defense. The metabolic feature vector and the defense feature vector are summed to generate a synergistic strength parameter between metabolism and defense. The correlation strength parameter, the antagonistic strength parameter, and the cooperative strength parameter are concatenated with the original sub-vector to generate a fusion feature vector that characterizes the correlation information of metabolic oxidative defense.

[0011] In some possible implementations, determining the target individuals in the livestock population exhibiting oxidative stress responses and the developmental stage information of the oxidative stress responses in the target individuals based on the oxidative stress state prediction sequence includes: The individual identifier and state label corresponding to each time point in the oxidative stress state prediction sequence are analyzed, and individuals whose state label is a potential stress state or an overt stress state are selected as candidate individuals. Time series backtracking is performed on the state prediction sequence of each candidate individual to identify the starting time point of the transition from the normal state to the potential stress state; The duration of the candidate individuals under the potential stress state is statistically analyzed, and combined with the occurrence time of the overt stress state, a stress response development timeline is generated. The characteristic change patterns of the candidate individuals at each stage of the stress response development timeline are analyzed, and a preset stage division rule library is matched to determine the time intervals of the initial triggering stage, the continuous development stage, and the severe deterioration stage corresponding to the oxidative stress response of the candidate individuals. Individuals among the candidate individuals whose state label is the dominant stress state are identified as target individuals, and the time interval is associated with the target individuals as the development stage information of the target individuals in generating the oxidative stress response.

[0012] In some possible implementations, the analysis of the characteristic change patterns of the candidate individual at each stage of the stress response development timeline, matching them with a preset stage division rule base, and determining the time intervals for the candidate individual to generate the oxidative stress response corresponding to the initial triggering stage, the continuous development stage, and the severe deterioration stage, includes: Extract the target physiological index features of the candidate individuals at each time point in the stress response development timeline, and generate a feature change trajectory map; The time point at which the metabolic activity features first show abnormal fluctuations in the feature change trajectory map is identified as the first starting point of the initial triggering phase. The point in time when the accumulation of oxidation products in the characteristic change trajectory diagram continues to rise and the antioxidant response characteristic begins to decline is identified as the second starting point of the continuous development stage. The time point in the trajectory of the characteristic change where the metabolic activity characteristics show irreversible decline and the accumulation of oxidative products reaches its peak is identified as the third starting point of the severe deterioration stage. Based on the first starting point, the second starting point, and the third starting point, and combined with the typical characteristic duration parameters of each stage, the time intervals for the initial triggering stage, the continuous development stage, and the severe deterioration stage corresponding to the oxidative stress response of the candidate individual are determined.

[0013] In conjunction with a second aspect of the present invention, an embodiment of the present invention provides a livestock and poultry oxidative stress response prediction system, the device comprising: a memory storing a computer program thereon; and a processor for executing the computer program stored in the memory to implement the livestock and poultry oxidative stress response prediction method according to any one of the first aspects.

[0014] The above technical solution acquires a set of real-time physiological monitoring data continuously collected from livestock and poultry populations and marked with timestamps. It can comprehensively and dynamically capture information on changes in the physiological state of livestock and poultry. By performing feature extraction processing on the real-time physiological monitoring data, it obtains target physiological indicators covering multiple aspects such as metabolic activities, accumulation of oxidative products, and response of the antioxidant defense system. It reflects the oxidative stress state of livestock and poultry from multiple dimensions, improving the accuracy and comprehensiveness of prediction.

[0015] Furthermore, a pre-trained oxidative stress prediction model is invoked to perform time-series correlation analysis on the target physiological indicators, fully considering the changing trends and interrelationships of physiological indicators over time, enabling the generation of more accurate oxidative stress state prediction sequences. Based on the prediction sequences, target individuals exhibiting oxidative stress responses and their developmental stages are identified, achieving precise localization and graded early warning of oxidative stress responses. Finally, based on the target individuals and their corresponding developmental stages, oxidative stress response prediction results are generated for each target individual. This allows for real-time and accurate prediction of oxidative stress responses in livestock and poultry, effectively preventing and mitigating the harm of oxidative stress to livestock and poultry, and improving their health and farming efficiency.

[0016] Other features and advantages of this disclosure will be described in detail in the following detailed description section. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the present disclosure and form part of the specification. They are used together with the following detailed description to explain the present disclosure, but do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram of the execution flow of the livestock and poultry oxidative stress response prediction method provided in the embodiments of the present invention.

[0018] Figure 2 This is one implementation provided by an embodiment of the present invention. Figure 1 A flowchart illustrating step S12.

[0019] Figure 3 This is one implementation provided by an embodiment of the present invention. Figure 1 A flowchart illustrating step S13.

[0020] Figure 4 This is one implementation provided by an embodiment of the present invention. Figure 1 A flowchart illustrating step S14.

[0021] Figure 5 This is a schematic diagram of exemplary hardware and software components of the livestock and poultry oxidative stress response prediction device provided in an embodiment of the present invention. Detailed Implementation

[0022] 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 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.

[0023] The specific embodiments of this disclosure will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit this disclosure.

[0024] This invention provides a method for predicting oxidative stress in livestock and poultry. (See also...) Figure 1 As shown, the method includes: In step S11, a real-time physiological monitoring data set of livestock and poultry population is constructed based on multiple physiological parameter units with timestamps collected continuously. Among them, the physiological parameter unit is a variety of parameters used to reflect the physiological state of livestock and poultry, such as body temperature, heart rate, respiratory rate, and the content of certain components in the blood (such as blood glucose, blood lipids, antioxidant enzyme activity, etc.). These parameters reflect the physiological functions and health status of livestock and poultry from different perspectives.

[0025] Among them, the real-time physiological monitoring data set is a data whole formed by organizing multiple physiological parameter units with timestamps that are continuously collected according to certain rules, which is used to reflect the changes in the physiological state of livestock and poultry groups in a specific time period in real time.

[0026] In this embodiment, physiological parameter acquisition devices (such as body temperature sensors, heart rate monitors, and blood analyzers) continuously collect physiological parameters of livestock and poultry at set time intervals, and add a timestamp to each collected physiological parameter unit to record the specific time of collection. Then, these timestamped physiological parameter units are organized and stored in chronological order to construct a real-time physiological monitoring data set for the livestock and poultry population. This allows for real-time tracking of changes in the physiological state of the livestock and poultry population.

[0027] For example, to accurately monitor and manage the health of pig herds and predict potential oxidative stress responses, real-time physiological monitoring data can be acquired. Smart wearable devices are installed on each pig's ear, integrating multiple sensors to continuously collect various physiological parameters in real time. These parameters include respiratory rate, body temperature, concentrations of specific biomolecules in the blood, activity of antioxidant enzymes, and levels of non-enzymatic antioxidants.

[0028] Furthermore, each collected physiological parameter is automatically tagged with a precise timestamp, recording the exact moment the parameter was collected. For example, at a certain moment, the smart wearable device collects data on pig A's respiratory rate, body temperature, concentration of specific biomolecules, antioxidant enzyme activity, and non-enzymatic antioxidant content. These data collectively constitute a physiological parameter unit, each bearing a timestamp from that moment. As time progresses, the device continuously collects data, gradually forming a real-time physiological monitoring dataset containing a large number of timestamped physiological parameter units. The data in this real-time physiological monitoring dataset is continuous, reflecting the physiological changes of the pig herd at different points in time.

[0029] In step S12, metabolic activity characteristics, oxidation product accumulation characteristics and antioxidant response characteristics are extracted from the real-time physiological monitoring data set to obtain target physiological indicators that reflect the oxidative stress state of livestock and poultry. Metabolic activity characteristics refer to the various features exhibited by livestock and poultry during the process of substance metabolism, including metabolic rate, types and amounts of metabolic products, etc. For example, the rates and products of energy metabolism, protein metabolism, and fat metabolism in livestock and poultry will differ under normal metabolic and stress conditions.

[0030] Among these characteristics, the accumulation of oxidation products is a key feature. In livestock and poultry, oxidation reactions produce a series of oxidation products, such as reactive oxygen species (ROS) and malondialdehyde (MDA). The accumulation characteristics of oxidation products reflect the degree and trend of these oxidation products in livestock and poultry. Excessive accumulation of oxidation products can damage cells and tissues.

[0031] Among these, antioxidant response characteristics refer to the presence of an antioxidant defense system in livestock and poultry, including antioxidant enzymes (such as superoxide dismutase SOD and glutathione peroxidase GSH-Px) and non-enzymatic antioxidants (such as vitamin C and vitamin E). Antioxidant response characteristics reflect the responsiveness and regulatory mechanisms of the antioxidant system in livestock and poultry to oxidative stress, such as changes in antioxidant enzyme activity.

[0032] In this embodiment, real-time physiological monitoring data is analyzed to extract features from three aspects: metabolic activity, accumulation of oxidative products, and antioxidant response. For metabolic activity features, the changes in physiological parameters related to energy metabolism and material metabolism (such as blood glucose levels and respiratory quotient) are analyzed to extract features. For oxidative product accumulation features, changes in the concentration of oxidative products (such as MDA content) in blood or tissues are detected. For antioxidant response features, changes in the activity of antioxidant enzymes (such as SOD activity) are measured. Then, statistical methods or machine learning algorithms are used to comprehensively analyze and screen these features to obtain target physiological indicators that accurately reflect the oxidative stress state of livestock and poultry.

[0033] In step S13, a pre-trained oxidative stress prediction model is invoked to perform time-series correlation analysis on the target physiological indicator features to generate a predicted sequence of oxidative stress status of the livestock and poultry population. Among them, the pre-trained oxidative stress prediction model establishes a mapping relationship between target physiological indicators and oxidative stress state by learning a large amount of known oxidative stress state livestock physiological data and their corresponding oxidative stress results, and can predict the oxidative stress state of new livestock physiological data.

[0034] Among them, time-series correlation analysis considers the time order and correlation of data, and performs correlation analysis on the data of target physiological indicators at different time points to discover the patterns and trends of data changes over time, thereby more accurately predicting the development and changes of oxidative stress in livestock and poultry.

[0035] Among them, the oxidative stress state prediction sequence is a series of prediction results generated after performing time-series correlation analysis on the characteristics of target physiological indicators based on the oxidative stress prediction model. These results are arranged in chronological order, reflecting the changing trend of oxidative stress state of livestock and poultry populations in the future.

[0036] In this embodiment, the pre-trained oxidative stress prediction model is trained on a large amount of livestock physiological data with known oxidative stress states. This model learns the complex nonlinear relationship between target physiological indicators and oxidative stress states. When new target physiological indicators are input, the model performs temporal correlation analysis on these features based on its internally learned patterns. Temporal correlation analysis considers the interrelationships and trends of data at different time points. By analyzing the changes in target physiological indicators over time, it predicts the oxidative stress state of livestock populations at future time points, generating an oxidative stress state prediction sequence.

[0037] In step S14, based on the oxidative stress state prediction sequence, the target individuals in the livestock and poultry population exhibiting oxidative stress responses and the developmental stage information of the target individuals in generating the oxidative stress responses are determined. Among these, the developmental stage information refers to the different stages at which livestock and poultry experience oxidative stress, such as early, middle, and late stages. Oxidative stress responses at different stages have different characteristics and degrees of impact.

[0038] In this embodiment, different oxidative stress thresholds are set based on an oxidative stress state prediction sequence. When the oxidative stress state value at a certain time point in the prediction sequence exceeds the corresponding threshold, it is determined that the livestock individual has an oxidative stress response. Simultaneously, based on the magnitude and trend of the oxidative stress state value, combined with predefined oxidative stress development stage standards, the development stage information of the target individual's oxidative stress response is determined. For example, if the oxidative stress state value is at a low level and rises slowly, it may be in the early stage; if the state value is high and continues to rise, it may be in the middle or late stage.

[0039] In step S15, based on the target individual and the corresponding developmental stage information, a prediction result of oxidative stress response for the target individual is generated.

[0040] In this embodiment of the disclosure, based on the identified target individual and its corresponding developmental stage information, and combined with relevant knowledge and experience regarding oxidative stress in livestock and poultry, a prediction result of oxidative stress response for the target individual is generated. The prediction result may include the severity of the oxidative stress response and its possible development trend.

[0041] In this embodiment of the disclosure, an oxidative stress early warning instruction carrying an individual identifier can also be generated based on the oxidative stress response prediction result, and the oxidative stress early warning instruction can be sent to the aquaculture management terminal to trigger an intervention operation.

[0042] In one possible implementation, an oxidative stress early warning instruction carrying an individual identifier is generated based on the oxidative stress response prediction result, and the oxidative stress early warning instruction is sent to the aquaculture management terminal to trigger an intervention operation. Specifically, this may include: extracting the unique identifier information of the target individual and the stage type label from the corresponding development stage information.

[0043] Specifically, the unique identifier of each target individual is crucial for distinguishing different pigs. Each pig has a specific identifier, such as an electronic ear tag number. This unique identifier is extracted from the relevant data of the target individual to ensure accurate identification of each pig. Simultaneously, a stage type label is extracted from the previously determined development stage information. This label clarifies the current stage of oxidative stress response in the target individual, such as the initial triggering stage, the ongoing development stage, or the severe deterioration stage. This information will serve as key content for generating early warning commands.

[0044] The system parses the pre-defined intervention strategy rule base and matches the intervention measure type and execution priority parameter associated with the stage type label.

[0045] Specifically, the pre-defined intervention strategy rule base is established based on extensive aquaculture experience and research findings. It includes intervention types and execution priority parameters corresponding to different stages of oxidative stress response. This rule base is parsed, and the extracted stage type tags are matched with the information in the rule base.

[0046] For example, if the stage type label is "initial trigger stage," the rule base might specify the corresponding intervention type as adjusting feed nutrient composition to enhance the antioxidant capacity of livestock and poultry, and this intervention would have a high priority. If the stage type label is "severe deterioration stage," the corresponding intervention type might include immediate drug treatment, with the highest priority. Through this matching, the most appropriate intervention for the target individual at the current stage of oxidative stress response, as well as the order in which these interventions are implemented, can be determined.

[0047] By integrating the unique identifier of the target individual, the stage type label, and the intervention type, an early warning information unit containing the individual identifier, stage identifier, and intervention identifier is generated.

[0048] Specifically, the extracted unique identifiers of the target individuals, stage type tags, and matched intervention types are integrated. A warning information unit containing an individual identifier, stage identifier, and intervention identifier is generated for each target individual. The individual identifier identifies which livestock or poultry is exhibiting oxidative stress; the stage identifier clearly indicates the stage of oxidative stress the livestock or poultry is in; and the intervention identifier details the intervention measures to be taken for that stage. This warning information unit comprehensively and accurately conveys key information about the oxidative stress response of the target individual.

[0049] The early warning information units are timestamped to ensure that the timestamps of each information unit are consistent with the timestamps of the oxidative stress state prediction sequence.

[0050] Specifically, to ensure the timeliness and accuracy of early warning information, the generated early warning information units can be timestamped. Each time point in the oxidative stress state prediction sequence has a corresponding timestamp, which records the oxidative stress state of livestock and poultry at different times. The timestamps in the early warning information units are compared and adjusted with the timestamps of the oxidative stress state prediction sequence to ensure consistency. The purpose of this is to ensure that the early warning information accurately reflects the oxidative stress response of the target individual at a specific time point, avoiding delays or inaccuracies in intervention operations due to time inconsistencies.

[0051] The timestamp-aligned early warning information unit is encapsulated into an oxidative stress early warning command, and the oxidative stress early warning command is sent to the aquaculture management terminal through a communication protocol. The aquaculture management terminal triggers intervention operations such as environmental regulation, nutritional supplementation or medical intervention according to the type of intervention measure.

[0052] Specifically, the timestamped early warning information units are encapsulated to form a complete oxidative stress early warning instruction. This instruction includes detailed information about the target individual, the stage of oxidative stress response, and corresponding intervention measures. The oxidative stress early warning instruction is sent to the livestock management terminal via a pre-set communication protocol. Upon receiving the instruction, the terminal automatically triggers the appropriate intervention based on the type of intervention. For example, if the intervention is environmental control, the terminal adjusts environmental parameters such as temperature, humidity, and ventilation in the livestock shed to improve the livestock's living environment; if it's nutritional supplementation, the terminal controls the feed delivery system to increase the proportion of antioxidants in the feed; if it's medical intervention, the terminal reminds livestock personnel to administer medication or other medical treatment to the target individual. Through this method, oxidative stress in livestock is addressed promptly and effectively, ensuring the healthy growth of livestock.

[0053] The above technical solution acquires a set of real-time physiological monitoring data continuously collected from livestock and poultry populations and marked with timestamps. It can comprehensively and dynamically capture information on changes in the physiological state of livestock and poultry. By performing feature extraction processing on the real-time physiological monitoring data, it obtains target physiological indicators covering multiple aspects such as metabolic activities, accumulation of oxidative products, and response of the antioxidant defense system. It reflects the oxidative stress state of livestock and poultry from multiple dimensions, improving the accuracy and comprehensiveness of prediction.

[0054] Furthermore, a pre-trained oxidative stress prediction model is invoked to perform time-series correlation analysis on the target physiological indicators, fully considering the changing trends and interrelationships of physiological indicators over time, enabling the generation of more accurate oxidative stress state prediction sequences. Based on the prediction sequences, target individuals exhibiting oxidative stress responses and their developmental stages are identified, achieving precise localization and graded early warning of oxidative stress responses. Finally, based on the target individuals and their corresponding developmental stages, oxidative stress response prediction results are generated for each target individual. This allows for real-time and accurate prediction of oxidative stress responses in livestock and poultry, effectively preventing and mitigating the harm of oxidative stress to livestock and poultry, and improving their health and farming efficiency.

[0055] See also some possible implementation methods. Figure 2As shown, in step S12, the extraction of metabolic activity characteristics, oxidation product accumulation characteristics, and antioxidant response characteristics from the real-time physiological monitoring data set yields target physiological indicators reflecting the oxidative stress state of livestock and poultry, including: In step S121, the real-time physiological monitoring data set is divided into time windows to obtain multiple monitoring data segment units with continuous time series relationships.

[0056] The time window segmentation process involves dividing continuous real-time physiological monitoring data into multiple monitoring data segments with continuous time series relationships. For example, using one hour as a time window, the data collected throughout the day is divided into 24 monitoring data segments, each containing all physiological parameter data collected within that hour. This approach facilitates targeted analysis and processing of data from different time periods, capturing the changing characteristics of the data at different time scales.

[0057] The monitoring data segment unit is a collection of physiological monitoring data within a specific time period that has a continuous time series relationship, obtained after time window division. Each monitoring data segment unit contains information on multiple physiological parameters of livestock and poultry within that time period.

[0058] In this embodiment, an appropriate time window length is determined based on the temporal patterns of changes in livestock and poultry physiological states and research needs. For example, if the focus is on the stress response of livestock and poultry within a short period, a shorter time window, such as 10 minutes, can be selected; if the study is on the physiological change trends of livestock and poultry throughout the day, a time window of 1 hour or longer can be selected. Then, according to the determined time window length, the continuous real-time physiological monitoring data is sequentially divided into multiple monitoring data segment units with continuous time series relationships. This decomposes long-term series data into multiple relatively independent yet time-related units.

[0059] In this embodiment of the disclosure, the real-time physiological monitoring data set can be divided according to certain time rules. Taking the monitoring data of livestock and poultry farms as an example, a suitable time window length is determined based on the physiological activity characteristics of livestock and poultry and the possible changing cycle of oxidative stress response. The entire real-time physiological monitoring data set is then divided according to this fixed time window length.

[0060] For example, starting from the moment data collection begins, the first time window encompasses data from the start time to the time plus the window length, with all physiological parameter units within this time period constituting a monitoring data segment unit. The second time window follows immediately after the first, starting from the end time of the first window and extending to the time plus the window length, forming the second monitoring data segment unit. This process continues, dividing the entire real-time physiological monitoring dataset into multiple monitoring data segment units with a continuous time series relationship. These monitoring data segment units are arranged chronologically, and each unit contains physiological parameter information of the livestock herd within a specific time period.

[0061] In step S122, metabolic parameters are screened for each monitoring data segment unit, and respiratory rate fluctuation characteristics and body temperature change rate characteristics related to energy metabolism are extracted as metabolic activity characteristics.

[0062] Among them, respiratory rate fluctuation characteristics refer to the changes in the respiratory rate of livestock and poultry over a certain period of time, including the average respiratory rate, standard deviation, and amplitude of change. Respiratory rate is one of the important indicators reflecting the energy metabolism and physiological state of livestock and poultry. Under oxidative stress, the respiratory rate of livestock and poultry may change significantly. Therefore, extracting its fluctuation characteristics helps to understand the metabolic activities of livestock and poultry.

[0063] Among them, the body temperature change rate characteristic is used to represent how quickly the body temperature of livestock and poultry changes over time, and can be obtained by calculating the amount of change in body temperature per unit time. Body temperature is an important indicator of the physiological state of livestock and poultry. Oxidative stress may affect the body temperature regulation mechanism of livestock and poultry, leading to changes in the body temperature change rate. Therefore, this characteristic can reflect the metabolic and stress state of livestock and poultry.

[0064] In this embodiment of the disclosure, for each monitoring data segment, respiratory rate and body temperature data closely related to energy metabolism are selected from the physiological parameters contained therein. First, the average value and standard deviation of respiratory rate within this time period are calculated to describe the fluctuation of respiratory rate and obtain respiratory rate fluctuation characteristics. Simultaneously, the rate of body temperature change is obtained by calculating the difference in body temperature between adjacent time points, and its trend within this time period is analyzed to extract the rate of body temperature change characteristics. Energy metabolism is the foundation of livestock and poultry life activities. Changes in respiratory rate and body temperature can directly or indirectly reflect the energy metabolism status of livestock and poultry. Under oxidative stress, energy metabolism will be disordered; therefore, these two characteristics can be used as metabolic activity characteristics to reflect the oxidative stress status of livestock and poultry.

[0065] In step S123, oxidation product detection processing is performed on each of the monitoring data segment units to extract specific biomolecule concentration change features that reflect the degree of free radical accumulation, and the specific biomolecule concentration change features are used as the oxidation product accumulation features.

[0066] Free radicals are highly reactive molecules or atoms that are generated during oxidation reactions in livestock and poultry. Under normal circumstances, the antioxidant system in livestock and poultry can eliminate excess free radicals. However, under oxidative stress, the production of free radicals exceeds the scavenging capacity of the antioxidant system, leading to their accumulation in the body. The degree of free radical accumulation reflects the severity of oxidative stress in livestock and poultry.

[0067] In the detection of the degree of free radical accumulation, specific biomolecules are usually selected as the detection targets, such as malondialdehyde (MDA) and 8-hydroxydeoxyguanosine (8-OHdG). Changes in the concentration of these biomolecules can indirectly reflect the accumulation of free radicals in livestock and poultry.

[0068] In this embodiment of the disclosure, for each monitoring data segment unit, specific detection methods (such as chemical analysis, immunoassay, etc.) are used to detect specific biomolecules (such as MDA, 8-OHdG, etc.) that reflect the degree of free radical accumulation. By measuring the concentration of these biomolecules at different time points and calculating their concentration changes, the concentration change characteristics of specific biomolecules are obtained. Since the degree of free radical accumulation is closely related to the state of oxidative stress, the concentration change characteristics of specific biomolecules can indirectly reflect the accumulation of free radicals in livestock and poultry, and thus serve as a characteristic of oxidative product accumulation to assess the degree of oxidative stress.

[0069] In step S124, each monitoring data segment unit is subjected to defense system analysis processing, and the antioxidant enzyme activity fluctuation characteristics and non-enzymatic antioxidant substance content change characteristics are extracted as the antioxidant response characteristics.

[0070] Antioxidant enzymes are an important component of the antioxidant system in livestock and poultry, such as superoxide dismutase (SOD) and glutathione peroxidase (GSH-Px). The fluctuation characteristics of antioxidant enzyme activity refer to the changes in the activity of these enzymes at different time points, including the average activity and the trend of change. Under oxidative stress, the activity of antioxidant enzymes may change in response to increased oxidative stress in the body.

[0071] Non-enzymatic antioxidants, such as vitamin C, vitamin E, and glutathione, can also participate in the antioxidant process in livestock and poultry. The characteristics of changes in the content of non-enzymatic antioxidants reflect how the content of these substances in livestock and poultry changes over time, and these changes can reflect the responsiveness of the antioxidant system in livestock and poultry.

[0072] In this embodiment, for each monitoring data segment, the activity of antioxidant enzymes (such as SOD, GSH-Px, etc.) is detected using biochemical methods, and the changes in their activity at different time points are analyzed to extract the fluctuation characteristics of antioxidant enzyme activity. Simultaneously, the content of non-enzymatic antioxidants (such as vitamin C, vitamin E, etc.) is detected, and their content changes are calculated to obtain the characteristics of non-enzymatic antioxidant content changes. Under oxidative stress, the antioxidant system in livestock and poultry responds, and the activity of antioxidant enzymes and the content of non-enzymatic antioxidants change. These characteristics can reflect the state of the antioxidant system and the ability of livestock and poultry to cope with oxidative stress.

[0073] In step S125, the metabolic activity characteristics, the oxidation product accumulation characteristics, and the antioxidant response characteristics are analyzed and processed synchronously to generate the target physiological indicator characteristics.

[0074] The synchronicity analysis involves a comprehensive analysis of metabolic activity characteristics, oxidative product accumulation characteristics, and antioxidant response characteristics to investigate whether their changes at different time points are synchronous or correlated. Through synchronicity analysis, the interrelationships between metabolic, oxidative, and antioxidant processes in livestock and poultry can be understood, thereby generating more comprehensive and accurate target physiological indicators that reflect the oxidative stress state of livestock and poultry.

[0075] In this embodiment, statistical methods or correlation analysis algorithms are used to perform synchronicity analysis on the extracted metabolic activity characteristics, oxidation product accumulation characteristics, and antioxidant response characteristics. For example, the correlation coefficients between different characteristics are calculated to analyze whether their changes at different time points are consistent or correlated. Through synchronicity analysis, the interaction between metabolic, oxidative, and antioxidant processes in livestock and poultry can be understood, and characteristics with synchronicity or correlation can be integrated to generate a set of target physiological indicator characteristics. This can more comprehensively and accurately reflect the oxidative stress state of livestock and poultry, improving the accuracy and reliability of subsequent oxidative stress prediction.

[0076] In one implementation, firstly, metabolic activity characteristics, oxidative product accumulation characteristics, and antioxidant response characteristics are aligned along the time dimension. Since these characteristics are extracted from the same monitoring data segment unit, they possess the same time-series information. Timestamp matching ensures that the data for these three characteristics correspond at the same point in time.

[0077] Then, analyze the trends and interrelationships of these three characteristics at different time points. For example, observe whether the fluctuations in respiratory rate and body temperature in the metabolic activity characteristics are synchronized with changes in the concentration of specific biomolecules in the oxidation product accumulation characteristics and changes in the activity of antioxidant enzymes and the content of non-enzymatic antioxidants in the antioxidant response characteristics at certain time points. If, within a certain time period, the metabolic activity characteristics show increased metabolic activity in livestock and poultry, while the oxidation product accumulation characteristics show increased free radical accumulation, and the antioxidant response characteristics show a corresponding enhancement of the antioxidant defense system function, then it indicates a synchronous response relationship among these three aspects within that time period.

[0078] By comprehensively analyzing and quantifying the aforementioned synchronous changes over the entire time series, a set of target physiological indicator features is generated. This set includes not only individual features of metabolic activity, accumulation of oxidative products, and response of the antioxidant defense system, but also synchronous correlations between them. This set of target physiological indicator features can more comprehensively and accurately reflect the oxidative stress state of livestock and poultry.

[0079] In some possible implementations, in step S122, the metabolic parameter screening process for each of the monitoring data segments, extracting respiratory rate fluctuation features and body temperature change rate features related to energy metabolism as the metabolic activity features, includes: Extract the continuous sampled value sequence of respiratory rate parameter from each of the monitoring data segment units, and calculate the absolute value of the frequency difference between adjacent sampling points in the continuous sampled value sequence as the frequency fluctuation amplitude parameter.

[0080] In this embodiment of the disclosure, the respiratory rate parameter collected by the smart wearable device in each monitoring data segment unit is a series of continuous sampled values. These sampled values ​​are arranged in chronological order to form a continuous sampled value sequence of the respiratory rate parameter. To measure the fluctuation of the respiratory rate, the frequency difference between adjacent sampling points can be calculated. For every two adjacent sampling points in the sequence, the respiratory rate value of the latter sampling point is subtracted from the respiratory rate value of the former sampling point, and the absolute value is taken. The result is the frequency fluctuation amplitude parameter between these two adjacent sampling points.

[0081] Furthermore, the above calculations are performed on each group of adjacent sampling points in the entire continuous sampling value sequence, ultimately yielding a set of frequency fluctuation amplitude parameters. This set of parameters reflects the degree of fluctuation in the respiratory rate of livestock and poultry within the monitoring data segment. The greater the frequency fluctuation amplitude, the more drastic the change in the respiratory rate of livestock and poultry, which may indicate abnormalities in their metabolic activities.

[0082] The frequency fluctuation amplitude parameters of each frequency fluctuation parameter within a preset time window are statistically analyzed and a respiratory frequency fluctuation pattern vector is generated based on the frequency fluctuation distribution.

[0083] In this embodiment of the disclosure, the calculated frequency fluctuation amplitude parameters can be further analyzed. A preset time window is set, which can be the same as the time window for dividing the monitoring data segment units, or it can be adjusted according to the actual situation.

[0084] Within a preset time window, the frequency fluctuation amplitude parameter is counted to determine its occurrence frequency. Different frequency fluctuation amplitude parameters are used as dimensions of a vector, and the frequency of occurrence for each dimension is used as its value, thus generating a respiratory frequency fluctuation pattern vector. This vector reflects the pattern of respiratory frequency fluctuations in livestock and poultry within the preset time window. Different fluctuation patterns may be related to different physiological states of livestock and poultry, particularly oxidative stress.

[0085] Extract the continuous sampled value sequence of body temperature parameters from each of the monitoring data segment units, and calculate the change in body temperature per unit time in the continuous sampled value sequence as the body temperature change rate parameter.

[0086] In this embodiment of the disclosure, in addition to respiratory rate, body temperature is also an important indicator reflecting the metabolic activity of livestock and poultry. A continuous sequence of body temperature parameters can also be obtained in each monitoring data segment unit. To understand the changes in the body temperature of livestock and poultry, the change in body temperature per unit time can be calculated. For every two adjacent sampling points in the sequence, the change in body temperature is obtained by subtracting the body temperature value of the previous sampling point from the body temperature value of the latter sampling point.

[0087] Then, the change in body temperature is divided by the time interval between the two sampling points to obtain the change in body temperature per unit time, i.e., the rate of change in body temperature parameter. This calculation is performed on each pair of adjacent sampling points in the entire continuous sampling sequence to obtain a set of rate of change in body temperature parameters. This set of parameters reflects the rate of change in body temperature of livestock and poultry within the monitoring data segment. An abnormal rate of change in body temperature may indicate problems with the metabolic activities of livestock and poultry, and may be related to oxidative stress.

[0088] The synchronicity correlation between the body temperature change rate parameter and the respiratory rate fluctuation pattern vector was analyzed to generate a metabolic activity synergy index.

[0089] In this embodiment of the disclosure, the synchronicity between the body temperature change rate parameter and the respiratory rate fluctuation pattern vector is analyzed. It can be determined whether there is synchronous change between the two by comparing the changing trend of the body temperature change rate parameter and the changes in the values ​​of each dimension in the respiratory rate fluctuation pattern vector.

[0090] If, within certain time periods, the rate of change of body temperature increases while certain dimensions of the respiratory rate fluctuation pattern vector also increase accordingly, it indicates a synchronous trend in these two parameters during those time periods. By comprehensively analyzing and quantifying the aforementioned synchronous changes across the entire monitoring data segment, a metabolic activity synergy index is generated. This index reflects the degree of synergistic change between the two metabolic indicators of respiratory rate and body temperature in livestock and poultry. Abnormalities in synergy may suggest that the metabolic activities of livestock and poultry have been disturbed, possibly related to oxidative stress.

[0091] The respiratory rate fluctuation pattern vector, the body temperature change rate parameter, and the metabolic activity synergy index are integrated into metabolic activity characteristics.

[0092] In this embodiment, the respiratory rate fluctuation pattern vector, body temperature change rate parameter, and metabolic activity synergy index are integrated to form a comprehensive metabolic activity feature. This feature includes information on respiratory rate fluctuations, body temperature changes, and the synergistic relationship between the two in livestock and poultry, and can more comprehensively reflect the metabolic activity of livestock and poultry.

[0093] In some possible implementations, in step S123, the oxidation product detection processing of each of the monitoring data segment units, extracting specific biomolecule concentration change features reflecting the degree of free radical accumulation, includes: Identify the concentration parameters of biomolecules related to free radical metabolism in each of the monitoring data segments, and extract the concentration value sequence of continuous sampling.

[0094] In this embodiment of the disclosure, the blood data collected by the smart wearable device in each monitoring data segment unit contains concentration information of various biomolecules. The concentration parameters of specific biomolecules related to free radical metabolism can be identified from this data. These specific biomolecules are key substances in the free radical metabolism process, and their concentration changes can reflect the generation and clearance of free radicals. The concentration parameters of these specific biomolecules are arranged in chronological order to form a continuously sampled concentration value sequence. This concentration value sequence records the changes in the concentration of specific biomolecules within the monitoring data segment unit.

[0095] The mean and standard deviation of the concentration value sequence are calculated to generate a concentration distribution stability index.

[0096] In this embodiment of the disclosure, to describe the distribution of a specific biomolecule concentration, the mean and standard deviation of the concentration value sequence can be calculated. The mean reflects the average level of all concentration values ​​in the sequence, while the standard deviation reflects the dispersion of the concentration values ​​relative to the mean. By comprehensively analyzing the mean and standard deviation, a concentration distribution stability index can be generated. If the standard deviation is small, it indicates that the concentration values ​​are relatively concentrated around the mean, and the concentration distribution is relatively stable; conversely, if the standard deviation is large, it indicates that the concentration values ​​are highly dispersed, and the concentration distribution is unstable. This concentration distribution stability index can help determine whether the generation and scavenging of free radicals are in a relatively balanced state. An unstable concentration distribution may indicate abnormalities in the accumulation or scavenging of free radicals, which is related to oxidative stress.

[0097] Analyze the trend of the concentration value sequence over time to generate parameters for the direction of concentration change.

[0098] In this embodiment of the disclosure, the temporal trend of specific biomolecule concentrations can also provide important information. The concentration value sequence can be analyzed to determine whether it shows an upward or downward trend within the entire monitoring data segment. The direction of concentration change can be determined by comparing the concentration values ​​at different time points in the sequence and by comprehensively analyzing the differences between adjacent concentration values.

[0099] If, for most of the time points, the concentration value at a later time point is greater than that at a previous time point, it indicates an upward trend in concentration; conversely, it indicates a downward trend. Based on the analysis results, a parameter for the direction of concentration change is generated. An upward trend in concentration change may indicate the continuous accumulation of free radicals, while a downward trend may indicate an enhanced ability to scavenge free radicals. This parameter can intuitively reflect the dynamic changes in free radical accumulation or scavenging, and is of great significance for judging the development of oxidative stress reactions.

[0100] The number of peak points exceeding the baseline level and the duration of the peak points are extracted from the concentration value sequence to generate an abnormal accumulation intensity index.

[0101] In this embodiment of the disclosure, some peak points exceeding the baseline level may appear in the concentration value sequence of a specific biomolecule. These peak points may indicate a sudden increase in the production of free radicals at certain times, leading to a sharp increase in the concentration of the specific biomolecule. These peak points exceeding the baseline level can be extracted from the concentration value sequence, and their number and duration of each peak point can be counted. The baseline level can be the average concentration of the biomolecule under normal physiological conditions. By comprehensively analyzing the number of peak points and the duration of the peaks, an abnormal accumulation intensity index is generated. The more peak points and the longer the duration of the peaks, the more severe the abnormal accumulation of free radicals and the stronger the correlation with oxidative stress.

[0102] The concentration distribution stability index, the concentration change direction parameter, and the abnormal accumulation intensity index are integrated into an oxidation product accumulation characteristic.

[0103] In this embodiment, the concentration distribution stability index, concentration change direction parameter, and abnormal accumulation intensity index are integrated to form a comprehensive characteristic of oxidation product accumulation. This characteristic includes information on the distribution stability, trend, and abnormal accumulation of specific biomolecule concentrations, and can comprehensively reflect the degree of free radical accumulation in livestock and poultry.

[0104] In subsequent predictions of oxidative stress responses, the accumulation characteristics of these oxidative products will serve as an important input to determine whether livestock and poultry are under oxidative stress and the degree of oxidative stress.

[0105] In some possible implementations, in step S124, the defense system analysis processing of each monitoring data segment unit, extracting antioxidant enzyme activity fluctuation characteristics and non-enzymatic antioxidant substance content change characteristics as the antioxidant response characteristics, includes: Extract the antioxidant enzyme activity parameters from each monitoring data segment unit to generate a first continuous sampling value sequence, and calculate the difference between the maximum and minimum activity values ​​in the first continuous sampling value sequence as the activity fluctuation range parameter. In this embodiment of the disclosure, the blood data collected by the smart wearable device in each monitoring data segment unit includes the activity parameters of antioxidant enzymes. These activity parameters are arranged in chronological order to form a continuous sampling value sequence of antioxidant enzyme activity parameters. To measure the fluctuation of antioxidant enzyme activity, the difference between the maximum and minimum activity values ​​in this sequence can be calculated. This difference is the activity fluctuation range parameter. The activity fluctuation range parameter reflects the magnitude of change in antioxidant enzyme activity within the monitoring data segment unit. A larger fluctuation range may indicate that the antioxidant defense system has been interfered with by external factors, possibly related to oxidative stress.

[0106] The percentage of activity values ​​higher than the baseline value in the first continuous sampling value sequence is counted to generate an activity enhancement index. In this embodiment of the disclosure, to further understand the changes in antioxidant enzyme activity, a baseline value can be set. This baseline value can be the average activity of antioxidant enzymes under normal physiological conditions. The number of activity values ​​higher than the baseline value in the continuous sampling sequence of antioxidant enzyme activity parameters is counted, and the proportion of these activity values ​​higher than the baseline value in the entire sequence is calculated. This proportion is the activity enhancement index. The activity enhancement index reflects the degree of enhancement of antioxidant enzyme activity within the monitored data segment. A higher proportion indicates that the activity of antioxidant enzymes is higher than normal at more time points, possibly indicating that the body has enhanced the function of its antioxidant defense system to cope with oxidative stress.

[0107] Extract the content parameters of non-enzymatic antioxidants from each monitoring data segment unit to generate a second continuous sampling value sequence, and calculate the linear regression slope of the content values ​​in the second continuous sampling value sequence as the content change rate parameter; In this embodiment of the disclosure, the blood data collected by the smart wearable device in each monitoring data segment unit includes the content parameters of non-enzymatic antioxidants. These content parameters are arranged in chronological order to form a continuous sampling value sequence of non-enzymatic antioxidant content parameters. To understand the changes in the content of non-enzymatic antioxidants, linear regression analysis can be performed on this sequence.

[0108] Linear regression analysis yields a fitted straight line, the slope of which reflects the rate of change of non-enzymatic antioxidant content over time. This slope is the content change rate parameter. The content change rate parameter directly reflects the increasing or decreasing trend of non-enzymatic antioxidant content; a positive slope indicates an increase, while a negative slope indicates a decrease. Abnormal content change rates may indicate a disruption of the balance of the antioxidant defense system, potentially linked to oxidative stress.

[0109] The correlation between the rate of change of non-enzymatic antioxidant substances and the fluctuation range of antioxidant enzyme activity was analyzed to generate a synergistic response index for the defense system. In this embodiment, the correlation between the rate of change parameter of non-enzymatic antioxidant substance content and the fluctuation range parameter of antioxidant enzyme activity is analyzed. The presence of a synergistic relationship between the two can be determined by comparing their trends and numerical changes. If, within certain time periods, the rate of change parameter of non-enzymatic antioxidant substance content increases while the fluctuation range parameter of antioxidant enzyme activity also changes accordingly, it indicates a synergistic trend between the two during those time periods.

[0110] By comprehensively analyzing and quantifying the aforementioned synergistic changes within the entire monitoring data segment, a synergistic response index for the defense system is generated. This index reflects the degree of synergistic effect of antioxidant enzymes and non-enzymatic antioxidants in responding to oxidative stress. Abnormalities in synergy may indicate that the function of the antioxidant defense system is affected and is related to oxidative stress responses.

[0111] The activity fluctuation range parameter, the activity enhancement degree index, the content change rate parameter, and the defense system synergistic response index are integrated into an antioxidant response feature.

[0112] In this embodiment, the activity fluctuation range parameter, activity enhancement degree index, content change rate parameter, and defense system synergistic response index are integrated to form a comprehensive antioxidant response characteristic. This characteristic includes information on the fluctuation of antioxidant enzyme activity, the degree of activity enhancement, the change rate of non-enzymatic antioxidant substance content, and the synergistic effect between the two, which can comprehensively reflect the response of livestock and poultry's antioxidant defense system to oxidative stress.

[0113] In subsequent predictions of oxidative stress response, this antioxidant response characteristic will serve as an important input to determine whether livestock and poultry are under oxidative stress and the degree of oxidative stress.

[0114] See also some possible implementation methods. Figure 3 As shown, in step S13, the step of calling the pre-trained oxidative stress prediction model to perform time-series correlation analysis on the target physiological indicator features and generating a predicted sequence of oxidative stress status for the livestock and poultry population includes: In step S131, the target physiological index features are input into the temporal coding layer of the oxidative stress prediction model, and the time series features are temporally encoded and represented by a long short-term memory network to generate a temporal coding representation vector. In this embodiment of the disclosure, the first key module of the oxidative stress prediction model is the temporal encoding layer, which uses a Long Short-Term Memory (LSTM) network to process the time-series information in the target physiological indicator features. LSTM is a special type of recurrent neural network that can effectively process data with time dependencies, capturing the long-term dependencies and trends of data over time.

[0115] The target physiological indicators are input into the Long Short-Term Memory (LSTM) network in chronological order. The LTM network contains multiple memory units that store and update historical information. For each time step's input, the LTM network uses a series of gating mechanisms, based on the current input and the hidden state of the previous time step, to determine which information can be retained, updated, and output. Specifically, the LTM network includes input gates, forget gates, and output gates. The input gate determines how much information from the current input can be added to the memory unit; the forget gate determines which historical information in the memory unit can be forgotten; and the output gate determines how much information is output from the memory unit as the hidden state of the current time step.

[0116] After processing the target physiological indicator features across the entire time series using a Long Short-Term Memory (LSTM) network, a corresponding hidden state is generated at each time step. Arranging these hidden states in chronological order forms a temporal encoding vector. This vector contains the encoded information of the target physiological indicator features over time, enabling it to more effectively represent the temporal dependencies and trends within these features.

[0117] In step S132, the temporal encoded representation vector is subjected to feature interaction through the feature interaction layer of the oxidative stress prediction model to generate a fusion feature vector for characterizing metabolic oxidative defense association information. In this embodiment of the disclosure, after obtaining the temporal encoded representation vector, it can be input into the feature interaction layer of the oxidative stress prediction model for feature interaction processing to generate a fused feature vector containing metabolic-oxidative-defense correlation information.

[0118] In step S133, the state classification layer of the oxidative stress prediction model is used to perform stress state probability prediction processing on the fused feature vector to generate the probability distribution of oxidative stress state at each time point. In this embodiment of the disclosure, after obtaining the fused feature vector containing metabolic-oxidative-defense correlation information, it is input into the state classification layer of the oxidative stress prediction model for stress state probability prediction. The state classification layer is a neural network-based classifier that can predict the probability of livestock and poultry being in different oxidative stress states at each time point based on the input fused feature vector.

[0119] The state classification layer contains multiple neurons and activation functions. A neuron is the basic computational unit of a neural network; it receives input signals, performs a weighted summation, and then converts the summation into an output signal through an activation function. The role of the activation function is to introduce nonlinearity, enabling the neural network to learn complex nonlinear relationships. Common activation functions include the sigmoid function and the ReLU function.

[0120] In this process, the fused feature vector serves as the input to the state classification layer. After weighted summation and activation function processing by neurons, it outputs a set of probability values. These probability values ​​represent the probability that livestock are in a normal state, a potential stress state, or a manifest stress state at that time point. For example, the output probability distribution might be [P1, P2, P3], where P1 represents the probability that livestock are in a normal state, P2 represents the probability that livestock are in a potential stress state, and P3 represents the probability that livestock are in a manifest stress state, and P1 + P2 + P3 = 1. By performing the above processing on the fused feature vector at each time point, the probability distribution of oxidative stress state corresponding to each time point can be generated.

[0121] In step S134, the probability distribution of the oxidative stress state is smoothed over time by the trend analysis layer of the oxidative stress prediction model to generate a state probability change curve within a continuous time window. In this embodiment of the disclosure, in order to more clearly observe the changing trend of oxidative stress state in livestock and poultry over time, the generated probability distribution of oxidative stress state can be smoothed over time. The trend analysis layer of the oxidative stress prediction model is responsible for accomplishing this task.

[0122] The trend analysis layer employs a smoothing algorithm, such as moving average or exponential smoothing, to process the probability distribution of oxidative stress states. Moving average is a simple smoothing method that smooths the data by calculating the average value over a certain time window. For example, for the probability distribution of oxidative stress states at each time point, the average of the probability distributions at several time points before and after it is taken as the smoothed probability distribution for that time point. Exponential smoothing is a more complex smoothing method that assigns different weights to data at different time points, with data closer to the current time point receiving greater weight, thus better reflecting the latest trends in data changes.

[0123] After smoothing the probability distribution of oxidative stress states over the entire time series, a smoothed probability distribution of oxidative stress states for each time point was obtained. Connecting the probabilities of normal state, potential stress state, and dominant stress state at each time point forms three continuous state probability change curves. These curves can intuitively show the probability changes of livestock and poultry in different oxidative stress states within a continuous time window, which helps to discover the development trend and potential changes of oxidative stress states in livestock and poultry.

[0124] In step S135, the state probability change curve is converted into an oxidative stress state prediction sequence that includes normal state, potential stress state, and overt stress state.

[0125] In this embodiment of the disclosure, after obtaining the state probability change curve, it can be converted into an oxidative stress state prediction sequence that includes normal state, potential stress state, and dominant stress state. This is achieved by setting a certain probability threshold.

[0126] For each time point, the probability distribution of oxidative stress states is compared with the probability of normal state, the probability of potential stress state, and the probability of dominant stress state, and then compared with preset probability thresholds. If the probability of normal state is greater than the preset normal state threshold, the oxidative stress state corresponding to that time point is marked as normal state; if the probability of potential stress state is greater than the preset potential stress state threshold and is also greater than the probability of normal state and dominant stress state, the oxidative stress state corresponding to that time point is marked as potential stress state; if the probability of dominant stress state is greater than the preset dominant stress state threshold and is also greater than the probability of normal state and potential stress state, the oxidative stress state corresponding to that time point is marked as dominant stress state.

[0127] Arranging the oxidative stress state markers corresponding to each time point in chronological order forms an oxidative stress state prediction sequence. This oxidative stress state prediction sequence clearly shows the oxidative stress state of livestock and poultry at different time points.

[0128] In some possible implementations, in step S132, the step of performing feature interaction on the temporal encoded representation vector through the feature interaction layer of the oxidative stress prediction model to generate a fusion feature vector for characterizing metabolic oxidative defense association information includes: The temporal coding representation vector is decomposed into metabolic feature vectors, oxidation feature vectors, and defense feature vectors.

[0129] In this embodiment of the disclosure, the temporal coding representation vector is a vector that integrates metabolic activity features, oxidation product accumulation features, and antioxidant response features. To further explore the correlation information between these features, the temporal coding representation vector can be decomposed into three sub-vectors: a metabolic feature sub-vector, an oxidation feature sub-vector, and a defense feature sub-vector.

[0130] The metabolic feature subvector contains feature information related to metabolic activities in the time-series encoded representation vector, such as respiratory rate fluctuation features, body temperature change rate features, and metabolic activity synergy indicators; the oxidation feature subvector contains feature information related to the accumulation of oxidation products, such as specific biomolecule concentration change features, concentration distribution stability indicators, concentration change direction parameters, and abnormal accumulation intensity indicators; the defense feature subvector contains feature information related to the response of the antioxidant defense system, such as antioxidant enzyme activity fluctuation features, activity enhancement degree indicators, non-enzymatic antioxidant substance content change features, and defense system synergy response indicators.

[0131] The metabolic feature vector and the oxidation feature vector are subjected to a dot product operation to generate a correlation strength parameter between metabolism and oxidation.

[0132] In this embodiment of the disclosure, to measure the correlation between metabolic activity and the accumulation of oxidative products, a dot product operation can be performed on the metabolic feature vector and the oxidation feature vector. The dot product operation is a common vector operation that can measure the similarity and correlation between two vectors. Multiplying the elements of the corresponding dimensions of the metabolic and oxidation feature vectors, and then summing all the multiplication results, yields the metabolic-oxidation correlation strength parameter.

[0133] It can be noted that the correlation strength parameter reflects the degree of mutual influence between metabolic activity and the accumulation of oxidation products. If the metabolism-oxidation correlation strength parameter is large, it indicates that there is a strong correlation between changes in metabolic activity and the accumulation of oxidation products, which may suggest that abnormalities in metabolic activity can lead to a large accumulation of oxidation products.

[0134] The difference between the oxidation feature vector and the defense feature vector is calculated to generate the antagonistic strength parameter between oxidation and defense.

[0135] In this embodiment of the disclosure, in order to understand the antagonistic relationship between the accumulation of oxidation products and the response of the antioxidant defense system, a difference operation can be performed on the oxidation feature vector and the defense feature vector. The elements of the corresponding dimensions of the oxidation feature vector and the defense feature vector are subtracted, and then the resulting difference vector is processed in a certain way (such as taking the absolute value or performing normalization operations to ensure the rationality and comparability of the results). The result is the oxidation-defense antagonistic strength parameter.

[0136] It can be explained that the antagonistic strength parameter reflects the power balance between the accumulation of oxidative products and the antioxidant defense system. If the oxidative-defense antagonistic strength parameter is large, it means that the accumulation of oxidative products exceeds the defense capacity of the antioxidant defense system, which may mean that livestock and poultry are in a state of oxidative stress and the situation is relatively serious.

[0137] The metabolic feature vector and the defense feature vector are summed to generate a synergistic strength parameter between metabolism and defense.

[0138] In this embodiment of the disclosure, to explore the synergistic relationship between metabolic activity and the antioxidant defense system, the metabolic feature vector and the defense feature vector can be summed. Adding the elements of the corresponding dimensions of the metabolic and defense feature vectors yields the metabolic-defense synergistic strength parameter.

[0139] It can be noted that the synergy strength parameter reflects the degree of synergy between metabolic activities and the antioxidant defense system. If the metabolic-defense synergy strength parameter is large, it indicates that metabolic activities and the antioxidant defense system can cooperate with each other to maintain the physiological balance of livestock and poultry. Conversely, if the parameter is small, it may indicate that there is a problem with the synergy between metabolic activities and the antioxidant defense system, and the physiological state of livestock and poultry may be affected.

[0140] The correlation strength parameter, the antagonistic strength parameter, and the cooperative strength parameter are concatenated with the original sub-vector to generate a fusion feature vector that characterizes the correlation information of metabolic oxidative defense.

[0141] In this embodiment, the calculated correlation strength parameters, antagonistic strength parameters, and synergistic strength parameters are concatenated with the original metabolic feature vectors, oxidative feature vectors, and defense feature vectors. The concatenation process arranges these vectors in a specific order to form a longer vector. This fused feature vector not only contains individual feature information from the three aspects of metabolism, oxidation, and defense, but also includes the correlation information between them, enabling a more comprehensive and in-depth reflection of the oxidative stress state of livestock and poultry.

[0142] See also some possible implementation methods. Figure 4 As shown, in step S14, determining the target individuals in the livestock and poultry population exhibiting oxidative stress response and the developmental stage information of the oxidative stress response in the target individuals based on the oxidative stress state prediction sequence includes: In step S141, the individual identifier and state label corresponding to each time point in the oxidative stress state prediction sequence are parsed, and individuals whose state label is a potential stress state or an overt stress state are selected as candidate individuals.

[0143] In this embodiment of the disclosure, the prediction result at each time point in the oxidative stress state prediction sequence is associated with a specific livestock or poultry individual and carries the individual's identification information and oxidative stress state tag. The oxidative stress state prediction sequence can be parsed to extract the individual identifier and state tag corresponding to each time point.

[0144] Next, livestock and poultry individuals are screened based on their status labels. Individuals labeled as potentially under stress or under obvious stress are identified as candidates for oxidative stress. This screening process narrows the scope of subsequent analysis, focusing on individuals that may warrant attention and intervention.

[0145] In step S142, time series backtracking is performed on the state prediction sequence of each candidate individual to identify the starting time point of the transition from the normal state to the potential stress state.

[0146] In this embodiment of the disclosure, for each candidate individual, its state prediction sequence records the individual's oxidative stress state at different time points. To understand the initiation of the oxidative stress response, time series backtracking processing can be performed on its state prediction sequence.

[0147] Starting from the last time point in the state prediction sequence of a candidate individual, trace back to find the first time point where the individual transitions from a normal state to a potential stress state. This time point is considered the starting point of the individual's oxidative stress response, marking the beginning of abnormal physiological states and the initiation of the oxidative stress response. By identifying this starting time point, the initial stage of the oxidative stress response can be determined.

[0148] In step S143, the duration of the candidate individual under the potential stress state is counted, and combined with the time point of occurrence of the overt stress state, a stress response development timeline is generated.

[0149] In this embodiment of the disclosure, after determining the starting time point of the transition from a normal state to a potential stress state for a candidate individual, the duration of that individual in the potential stress state can be statistically analyzed. By analyzing the state prediction sequence, the time interval from the starting time point to the individual's first entry into the overt stress state is calculated; this time interval is the duration of that individual in the potential stress state.

[0150] By combining the timing of the onset of the overt stress state, the entire process of the candidate individual from the normal state to the potential stress state and then to the overt stress state is arranged in chronological order to generate a stress response development timeline. This timeline clearly shows the development process of the individual's oxidative stress response, including the initial stage, the potential stress stage, and the overt stress stage.

[0151] In step S144, the characteristic change patterns of the candidate individual at each stage of the stress response development timeline are analyzed, and a preset stage division rule library is matched to determine the time intervals of the initial triggering stage, the continuous development stage, and the severe deterioration stage corresponding to the oxidative stress response of the candidate individual.

[0152] In this embodiment of the disclosure, in order to understand the development process of oxidative stress response in candidate individuals in more detail, the characteristic change patterns of each stage in the stress response development timeline can be analyzed and matched with a preset stage division rule base to determine the time intervals of the initial triggering stage, the continuous development stage, and the severe deterioration stage of the oxidative stress response.

[0153] In step S145, individuals among the candidate individuals whose state label is the dominant stress state are identified as target individuals, and the time interval is associated with the target individuals as the development stage information of the target individuals in generating the oxidative stress response.

[0154] In this embodiment of the disclosure, after analyzing the development stages of oxidative stress in candidate individuals, individuals with a state label indicating a dominant stress state are selected from the candidate individuals and identified as target individuals. These target individuals are those with the most severe oxidative stress in the current livestock and poultry herd, and can be given special attention and intervention measures.

[0155] The time intervals of the initial triggering stage, the continuous development stage, and the severe deterioration stage are used as information on the developmental stages of oxidative stress response in these target individuals. This information records the developmental process of oxidative stress response in the target individuals in detail, which is of great guiding significance for farmers to understand the health status of target individuals and formulate targeted intervention measures. By clarifying the target individuals and their oxidative stress response development stages, oxidative stress problems in livestock and poultry herds can be managed and controlled more accurately, thereby improving breeding efficiency and the health level of livestock and poultry.

[0156] In some possible implementations, in step S144, analyzing the characteristic change patterns of the candidate individual at each stage of the stress response development timeline, matching them with a preset stage division rule base, and determining the time intervals for the candidate individual to generate the oxidative stress response corresponding to the initial triggering stage, the continuous development stage, and the severe deterioration stage includes: The target physiological indicators of the candidate individuals at each time point in the stress response development timeline are extracted to generate a feature change trajectory map.

[0157] In this embodiment, each time point in the stress response development timeline of a candidate individual corresponds to a set of target physiological indicators. These target physiological indicators are arranged chronologically to generate a feature change trajectory diagram. This trajectory diagram can visually display the changes in target physiological indicators such as metabolic activity characteristics, oxidative product accumulation characteristics, and antioxidant response characteristics of the candidate individual during the development of oxidative stress. By observing the feature change trajectory diagram, the changing trends and characteristics of the features at different stages can be discovered.

[0158] The time point at which the metabolic activity features first show abnormal fluctuations in the feature change trajectory map is identified as the first starting point of the initial triggering phase.

[0159] In this embodiment of the disclosure, changes in metabolic activity characteristics are observed in the characteristic change trajectory graph. These metabolic activity characteristics include respiratory rate fluctuations, body temperature change rates, and metabolic activity synergy indicators. When one or more of these characteristics first exhibit a significant abnormal fluctuation, that time point is recorded as the starting point of the initial triggering phase. This time point marks the initial triggering of the oxidative stress response, which may be caused by external environmental stimuli or changes in the animal's own physiological state.

[0160] The point in time when the accumulation of oxidation products in the characteristic change trajectory diagram continues to rise and the antioxidant response characteristic begins to decline is identified as the second starting point of the continuous development stage.

[0161] In this embodiment of the disclosure, continuing with the characteristic change trajectory map, when it is found that the concentration of specific biomolecules in the oxidation product accumulation characteristics continues to rise, while the activity of antioxidant enzymes and the content of non-enzymatic antioxidants in the antioxidant response characteristics begin to decline, this time point is recorded as the starting point of the continuous development stage. In this stage, the accumulation rate of oxidation products exceeds the clearance capacity of the antioxidant defense system, the oxidative stress response begins to develop continuously, and the physiological state of livestock and poultry gradually deteriorates.

[0162] The point in time when the metabolic activity characteristics in the trajectory of the characteristic change show irreversible decline and the accumulation of oxidative products reach its peak is identified as the third starting point of the severe deterioration stage.

[0163] In this embodiment of the disclosure, the time point in the characteristic change trajectory graph when irreversible decline in metabolic activity characteristics occurs, such as severe abnormalities in respiratory rate and body temperature that cannot be restored to normal, and the accumulation of oxidative products reaches its peak, is recorded as the starting point of the severe deterioration stage. At this stage, the physiological functions of livestock and poultry are severely impaired, and the oxidative stress response has developed to a very serious degree. If intervention measures are not taken in time, it may lead to a rapid deterioration of the health status of livestock and poultry, or even death.

[0164] Based on the first starting point, the second starting point, and the third starting point, and combined with the typical characteristic duration parameters of each stage, the time intervals for the initial triggering stage, the continuous development stage, and the severe deterioration stage corresponding to the oxidative stress response of the candidate individual are determined.

[0165] In this embodiment, after determining the starting point of the initial triggering stage, the starting point of the continuous development stage, and the starting point of the severe deterioration stage, the time interval of each stage is determined by combining preset typical characteristic duration parameters for each stage. For example, the typical characteristic duration parameter of the initial triggering stage is t1, and the time interval of the initial triggering stage extending from the starting point of the initial triggering stage to t1 is the time interval of the initial triggering stage; the typical characteristic duration parameter of the continuous development stage is t2, and the time interval of the continuous development stage extending from the starting point of the continuous development stage to t2 is the time interval of the continuous development stage; the severe deterioration stage begins from the starting point of the severe deterioration stage and continues until the stress response ends or the health status of the livestock and poultry improves. Through the above method, the time intervals of the initial triggering stage, the continuous development stage, and the severe deterioration stage of the oxidative stress response in candidate individuals are accurately determined.

[0166] This invention provides a livestock and poultry oxidative stress response prediction system. The device includes: a memory storing a computer program thereon; and a processor for executing the computer program stored in the memory to implement any of the livestock and poultry oxidative stress response prediction methods described in the foregoing embodiments.

[0167] Figure 5 The livestock and poultry oxidative stress response prediction device 100 shown includes a processor 1001 and a memory 1003. The processor 1001 and the memory 1003 are connected, for example, via a bus 1002. Optionally, the livestock and poultry oxidative stress response prediction device 100 may further include a communication component 1004, which can be used for data interaction between the device 100 and other devices, such as data transmission and / or data reception. It should be noted that in actual scheduling, the communication component 1004 is not limited to one, and the structure of this livestock and poultry oxidative stress response prediction device 100 does not constitute a limitation on the embodiments of this application.

[0168] Processor 1001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 1001 may also be a combination that implements computing functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0169] Bus 1002 may include a pathway for transmitting information between the aforementioned components. Bus 1002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. Bus 1002 can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 5 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0170] The memory 1003 may be ROM (Read Only Memory) or other types of static storage devices capable of storing static information and instructions, RAM (Random Access Memory) or other types of dynamic storage devices capable of storing information and instructions, or EEPROM (Electrically Erasable Programmable Read Only Memory), CD-ROM (Compact Disc Read Only Memory) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media, other magnetic storage devices, or any other medium capable of carrying or storing program code and capable of being read by a computer, without limitation herein.

[0171] The memory 1003 is used to store program code for executing embodiments of the present disclosure, and its execution is controlled by the processor 1001. The processor 1001 is used to execute the program code stored in the memory 1003 to implement the steps shown in the aforementioned embodiments of the livestock and poultry oxidative stress response prediction method.

[0172] The preferred embodiments of the present disclosure have been described in detail above with reference to the accompanying drawings. However, the present disclosure is not limited to the specific details of the above embodiments. Within the scope of the technical concept of the present disclosure, various changes, modifications, substitutions and variations can be made to these embodiments, and all such changes, modifications, substitutions and variations fall within the protection scope of the present disclosure.

[0173] Furthermore, it should be noted that the various specific technical features described in the above embodiments can be combined in any suitable manner without contradiction, and such combinations should also be considered as part of this disclosure. To avoid unnecessary repetition, this disclosure will not further describe the various possible combinations. The technical scope of this application is not limited to the contents of the specification, but must be determined according to the scope of the claims.

Claims

1. A method for predicting oxidative stress response in livestock and poultry, characterized in that, The method includes: A real-time physiological monitoring dataset for livestock and poultry populations is constructed based on multiple time-stamped physiological parameter units collected continuously. The metabolic activity characteristics, oxidation product accumulation characteristics, and antioxidant response characteristics of the real-time physiological monitoring data set are extracted to obtain target physiological indicators that reflect the oxidative stress state of livestock and poultry. The pre-trained oxidative stress prediction model is invoked to perform time-series correlation analysis on the target physiological indicators to generate a predicted sequence of oxidative stress status for the livestock and poultry population. Based on the oxidative stress state prediction sequence, the target individuals in the livestock and poultry population that exhibit oxidative stress response and the developmental stage information of the target individuals in producing the oxidative stress response are determined. Based on the target individual and the corresponding developmental stage information, an oxidative stress response prediction result is generated for the target individual.

2. The method for predicting oxidative stress response in livestock and poultry according to claim 1, characterized in that, The process involves extracting metabolic activity characteristics, oxidative product accumulation characteristics, and antioxidant response characteristics from the real-time physiological monitoring data set to obtain target physiological indicators reflecting the oxidative stress state of livestock and poultry, including: The real-time physiological monitoring data set is divided into time windows to obtain multiple monitoring data segment units with continuous time series relationships; For each monitoring data segment unit, metabolic parameters are screened, and respiratory rate fluctuation characteristics and body temperature change rate characteristics related to energy metabolism are extracted as metabolic activity characteristics. Each monitoring data segment unit is processed for oxidation product detection, and specific biomolecule concentration change features that reflect the degree of free radical accumulation are extracted, and the specific biomolecule concentration change features are used as the oxidation product accumulation features. Each monitoring data segment unit is subjected to defense system analysis and processing to extract the characteristics of antioxidant enzyme activity fluctuations and the characteristics of non-enzymatic antioxidant substance content changes as the antioxidant response characteristics. The metabolic activity characteristics, the oxidation product accumulation characteristics, and the antioxidant response characteristics are analyzed and processed simultaneously to generate the target physiological indicator characteristics.

3. The method for predicting oxidative stress response in livestock and poultry according to claim 2, characterized in that, The process of screening metabolic parameters for each monitoring data segment unit, and extracting respiratory rate fluctuation characteristics and body temperature change rate characteristics related to energy metabolism as metabolic activity characteristics, includes: Extract the continuous sampled value sequence of respiratory rate parameter from each of the monitoring data segment units, and calculate the absolute value of the frequency difference between adjacent sampling points in the continuous sampled value sequence as the frequency fluctuation amplitude parameter; The frequency fluctuation amplitude parameters of each frequency fluctuation parameter within a preset time window are statistically distributed, and a respiratory frequency fluctuation pattern vector is generated based on the frequency distribution. Extract the continuous sampling value sequence of body temperature parameters from each of the monitoring data segment units, and calculate the amount of body temperature change per unit time in the continuous sampling value sequence as the body temperature change rate parameter; The synchronous correlation between the body temperature change rate parameter and the respiratory rate fluctuation pattern vector was analyzed to generate a metabolic activity synergy index. The respiratory rate fluctuation pattern vector, the body temperature change rate parameter, and the metabolic activity synergy index are integrated into metabolic activity characteristics.

4. The method for predicting oxidative stress response in livestock and poultry according to claim 2, characterized in that, The process of detecting oxidation products in each of the monitoring data segments and extracting specific biomolecule concentration changes that reflect the degree of free radical accumulation includes: Identify the concentration parameters of biomolecules related to free radical metabolism in each of the monitoring data segment units, and extract the concentration value sequence of continuous sampling; Calculate the mean and standard deviation of the concentration value sequence to generate a concentration distribution stability index; Analyze the trend of the concentration value sequence over time to generate parameters for the direction of concentration change; Extract the number of peak points exceeding the baseline level and the duration of the peak points from the concentration value sequence to generate an abnormal accumulation intensity index; The concentration distribution stability index, the concentration change direction parameter, and the abnormal accumulation intensity index are integrated into an oxidation product accumulation characteristic.

5. The method for predicting oxidative stress response in livestock and poultry according to claim 2, characterized in that, The step of performing defense system analysis on each monitoring data segment unit, extracting antioxidant enzyme activity fluctuation characteristics and non-enzymatic antioxidant substance content change characteristics as the antioxidant response characteristics, includes: Extract the antioxidant enzyme activity parameters from each monitoring data segment unit to generate a first continuous sampling value sequence, and calculate the difference between the maximum and minimum activity values ​​in the first continuous sampling value sequence as the activity fluctuation range parameter. The percentage of activity values ​​higher than the baseline value in the first continuous sampling value sequence is counted to generate an activity enhancement index. Extract the content parameters of non-enzymatic antioxidants from each monitoring data segment unit to generate a second continuous sampling value sequence, and calculate the linear regression slope of the content values ​​in the second continuous sampling value sequence as the content change rate parameter; The correlation between the rate of change of non-enzymatic antioxidant substances and the fluctuation range of antioxidant enzyme activity was analyzed to generate a synergistic response index for the defense system. The activity fluctuation range parameter, the activity enhancement degree index, the content change rate parameter, and the defense system synergistic response index are integrated into an antioxidant response feature.

6. The method for predicting oxidative stress response in livestock and poultry according to claim 1, characterized in that, The step of calling a pre-trained oxidative stress prediction model to perform time-series correlation analysis on the target physiological indicator features, generating a predicted sequence of oxidative stress status for the livestock and poultry population, including: The target physiological index features are input into the temporal coding layer of the oxidative stress prediction model, and the time series features are represented by temporal coding through a long short-term memory network to generate a temporal coding representation vector. The oxidative stress prediction model uses a feature interaction layer to perform feature interaction on the temporal encoded representation vector to generate a fusion feature vector that characterizes the association information of metabolic oxidative defense. The state classification layer of the oxidative stress prediction model is used to perform stress state probability prediction processing on the fused feature vector to generate the probability distribution of oxidative stress state at each time point. The probability distribution of the oxidative stress state is smoothed over time by the trend analysis layer of the oxidative stress prediction model to generate a state probability change curve within a continuous time window. The state probability change curve is converted into an oxidative stress state prediction sequence that includes normal state, potential stress state, and overt stress state.

7. The method for predicting oxidative stress response in livestock and poultry according to claim 6, characterized in that, The step of performing feature interaction on the temporal encoded representation vector through the feature interaction layer of the oxidative stress prediction model to generate a fused feature vector for characterizing metabolic oxidative defense association information includes: The temporal coding representation vector is decomposed into metabolic feature vectors, oxidation feature vectors, and defense feature vectors; Perform a dot product operation on the metabolic feature vector and the oxidation feature vector to generate a correlation strength parameter between metabolism and oxidation. The difference operation is performed on the oxidation feature vector and the defense feature vector to generate the antagonistic strength parameter between oxidation and defense. The metabolic feature vector and the defense feature vector are summed to generate a synergistic strength parameter between metabolism and defense. The correlation strength parameter, the antagonistic strength parameter, and the cooperative strength parameter are concatenated with the original sub-vector to generate a fusion feature vector that characterizes the correlation information of metabolic oxidative defense.

8. A method for predicting oxidative stress response in livestock and poultry according to any one of claims 1-7, characterized in that, The step of determining the target individuals in the livestock and poultry population exhibiting oxidative stress responses and the developmental stage of these responses based on the oxidative stress state prediction sequence includes: The individual identifier and state label corresponding to each time point in the oxidative stress state prediction sequence are analyzed, and individuals whose state label is a potential stress state or an overt stress state are selected as candidate individuals. Time series backtracking is performed on the state prediction sequence of each candidate individual to identify the starting time point of the transition from the normal state to the potential stress state; The duration of the candidate individuals under the potential stress state is statistically analyzed, and combined with the occurrence time of the overt stress state, a stress response development timeline is generated. The characteristic change patterns of the candidate individuals at each stage of the stress response development timeline are analyzed, and a preset stage division rule library is matched to determine the time intervals of the initial triggering stage, the continuous development stage, and the severe deterioration stage corresponding to the oxidative stress response of the candidate individuals. Individuals among the candidate individuals whose state label is the dominant stress state are identified as target individuals, and the time interval is associated with the target individuals as the development stage information of the target individuals in generating the oxidative stress response.

9. The method for predicting oxidative stress response in livestock and poultry according to claim 8, characterized in that, The analysis of the characteristic change patterns of the candidate individuals at each stage of the stress response development timeline, matching them with a preset stage division rule base, determines the time intervals for the initial triggering stage, the continuous development stage, and the severe deterioration stage corresponding to the oxidative stress response of the candidate individuals, including: Extract the target physiological index features of the candidate individuals at each time point in the stress response development timeline, and generate a feature change trajectory map; The time point at which the metabolic activity features first show abnormal fluctuations in the feature change trajectory map is identified as the first starting point of the initial triggering phase. The point in time when the accumulation of oxidation products in the characteristic change trajectory diagram continues to rise and the antioxidant response characteristic begins to decline is identified as the second starting point of the continuous development stage. The time point in the trajectory of the characteristic change where the metabolic activity characteristics show irreversible decline and the accumulation of oxidative products reaches its peak is identified as the third starting point of the severe deterioration stage. Based on the first starting point, the second starting point, and the third starting point, and combined with the typical characteristic duration parameters of each stage, the time intervals for the initial triggering stage, the continuous development stage, and the severe deterioration stage corresponding to the oxidative stress response of the candidate individual are determined.

10. A prediction system for oxidative stress response in livestock and poultry, characterized in that, The device includes: A memory on which computer programs are stored; A processor is configured to execute the computer program stored in the memory to implement the livestock and poultry oxidative stress response prediction method according to any one of claims 1-9.

Images