Cow breeding information intelligent management system

Abstract

The invention discloses a dairy cow breeding information intelligent management system, and belongs to the technical field of dairy cow breeding. An intelligent management system for dairy cow breeding information comprises an information acquisition module, an analysis management module, a statistical arrangement module and a browsing query module. According to the method, the problem that problems in the breeding process cannot be found out and corresponding solutions cannot be given according to the breeding information of the dairy cows due to the fact that the breeding information of the dairy cows cannot be collected and managed in the prior art is solved, scientific breeding is achieved by managing the breeding information of the dairy cows, the breeding cost is effectively reduced, and the breeding efficiency of the dairy cows is improved. The milk production of the dairy cows is improved, the economic income of a dairy cow farm is increased, so that the quality and yield of the milk are improved, a coping scheme can be matched according to the growth condition of the dairy cows, abnormal conditions occurring in the breeding process of the dairy cows are rapidly solved, intelligent management of dairy cow breeding information is achieved, the quality of dairy cow breeding is improved, and resource use is optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of dairy cow breeding, and in particular to an intelligent management system for dairy cow breeding information. Background Art

[0002] The quality of a dairy cow is measured by its milk production, but in addition to the relationship between breeds, milk production is also affected by factors such as the cow's dairy cycle, breeding, health status, and weight.

[0003] The Chinese patent with the publication number CN114896441A discloses an intelligent dairy farming method and system. The method uses an image acquisition device to collect images of the activity status of dairy cows to establish a dairy cow image library including codes and activity status images; uses a dairy cow farming experience database to build a dairy cow evaluation system model; dairy cow farming information is sequentially input into the dairy cow evaluation system model to obtain dairy cow evaluation results; data clustering is performed based on the dairy cow evaluation results, and dairy cows are classified by state according to the data clustering analysis results to establish dairy cow classification clusters, and farming feature marks are respectively performed on the clusters, and the marks are bound to the dairy cow codes; dairy cow coding information is used to identify dairy cow farming feature marks, and farming information is determined to conduct informatized farming supervision on dairy cows. The technical effect of using informatization guidance for targeted feeding, improving farming efficiency and ensuring milk production is achieved.

[0004] The above patent cannot collect and manage the breeding information of dairy cows in actual use, resulting in the inability to find out the problems in the breeding process and provide corresponding solutions based on the breeding information of dairy cows; therefore, it does not meet the existing needs. We have proposed an intelligent management system for dairy cow breeding information. Summary of the invention

[0005] The purpose of the present invention is to provide an intelligent management system for dairy cow breeding information, which can effectively reduce feeding costs, increase milk production of dairy cows, increase the economic income of dairy farms, and thus improve the quality and output of milk by realizing scientific breeding. It can match response plans according to the growth conditions of dairy cows, quickly solve abnormal situations that occur during the breeding process of dairy cows, realize intelligent management of dairy cow breeding information, improve the quality of dairy cow breeding, optimize resource utilization, and solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: an intelligent management system for dairy cattle breeding information, comprising: An information collection module is used to preset a collection frequency and obtain dairy cattle breeding data according to the preset collection frequency; An analysis management module is used to construct a target loss learning function of a data classification model, classify the real-time collected data based on the target loss learning function, and analyze the growth information and behavior status data of each cow, determine direct abnormal data from the breeding information of the cow according to the analysis results, perform a first labeling process, and determine the associated data of each first cluster set as a second cluster set, determine indirect abnormal data from the second cluster set, perform a second labeling process, and determine the final result of labeling the abnormal data based on the first labeling process result and the second labeling process result; The statistical arrangement module is used to summarize, arrange and count various dairy cow breeding information and generate corresponding query passwords; The browsing query module is used to query and display dairy cow breeding information and quickly locate the location of dairy cows.

[0007] Preferably, the information collection module includes: A real-time collection module is used to preset a collection frequency and obtain dairy cow breeding data according to the preset collection frequency. The dairy cow breeding data includes environmental influencing factors, dairy cow growth information, dairy cow business data and dairy cow behavior status data; The data cleaning module is used to judge the collected data, determine whether the collected data is missing or repeated, and process the missing and repeated data; Data cleaning module, including: Fill in missing values: Determine whether the real-time collected data is missing. If so, use the preset rules to fill in the missing values. ‌Delete Duplicate Data‌: Determine whether there are duplicates in the real-time collected data. If there are duplicates, choose to keep the first or last row of data in each group of duplicate rows and delete the remaining rows; ‌Handling duplicate values‌: Check and remove invalid duplicate data through case processing, or mark the first duplicate value that appears; ‌Identify and handle outliers‌: Use the MAD method to identify outliers and replace them with appropriate values; ‌Data normalization‌: Eliminate dimensional differences between data through min-max normalization method‌.

[0008] Preferably, the analysis management module includes: A storage module is used to store the cleaned data according to environmental factors, growth information of dairy cows, business data of dairy cows and behavior status data of dairy cows, and set corresponding labels; An analysis module is used to analyze the health status of dairy cows based on environmental factors, growth information of dairy cows, business data of dairy cows, and behavioral status data of dairy cows; The management module is used to determine whether there are any abnormalities in the dairy cow breeding process according to the health status of the dairy cows, and take corresponding management measures based on the results.

[0009] Preferably, the analysis module specifically includes: A database establishment module is used to establish a dairy cattle breeding information management database based on the standard growth information and behavior status data of dairy cows at different growth stages; A retrieval module is used to retrieve the breeding information of the dairy cow at this stage according to the growth stage of the dairy cow, obtain the standard breeding information of the dairy cow, and retrieve the breeding information of the dairy cow; The comparison module is used to compare the dairy cow breeding information collected in real time with the retrieved dairy cow standard breeding information to determine the comparison result.

[0010] Preferably, the management module includes: The judgment module is used to judge the health status of dairy cows and whether there are any abnormalities in their activity status based on the analysis results; A determination module is used to determine the problems existing in the dairy cow breeding process according to the determined results, and match corresponding solutions according to the existing problems; Determine the workflow of the module, including: Obtain the influencing factors corresponding to the abnormal information in the historical dairy farming process, formulate corresponding solutions based on the abnormal information of dairy farming, and establish a solution database; Compare the marked abnormal data with the influencing factors corresponding to the abnormal information in the dairy farming process; According to the comparison results, the corresponding solution is matched in the solution data.

[0011] Preferably, the management process of the management module specifically includes: The breeding information of the dairy cow is judged according to the analysis result of the analysis module to determine whether there is any abnormality in the breeding information of the dairy cow; If there is an abnormality, the abnormal data is marked, and the specific information in the dairy cattle breeding information represented by the marked abnormal data is determined; According to the determined results, the problems existing in the dairy cow breeding process are judged, and corresponding solutions are matched according to the existing problems.

[0012] Preferably, the judging module comprises: The first judgment module is used to judge the health status of dairy cow breeding according to the growth information and business data of dairy cows; The second judgment module is used to judge the activity status of dairy cow farming according to the behavior status data of the dairy cow.

[0013] Preferably, the statistical arrangement module includes: The data analysis report generation module is used to receive a request for report generation, retrieve the analysis results obtained by the analysis management module according to the received report generation request, and generate corresponding report information; The statistical result display module is used to visually display the report information generated by the report generation module.

[0014] Preferably, the target loss learning function for constructing the data classification model includes: Get the construction data set; divide each construction data in the construction data set into sub-data, obtain the data category of each sub-data, and then construct the initial loss learning function :

[0015] Among them, is the constructed initial loss learning function; is the bias parameter after initialization; is a natural constant; To build data Middle The category of each sub-data; Structuring data for prediction Middle The probability that the category of each sub-data is accurate; is the category of the labeled sub-data; Adjust the initialized bias parameters during the construction process;

[0016] in, is the adjusted bias parameter; is the loss learning rate of the loss learning function, with an initial value of 0.002; Learn a function for the initial loss After initialization, the bias parameters Conduct partial guidance; According to the adjusted bias parameters, the initial loss learning function is adjusted to obtain the target loss learning function ;

[0017] in, For The set of categories predicted by each sub-data; is the category of the unlabeled sub-data.

[0018] Preferably, the management module includes: The first processing module is used for: Extracting first data to be labeled from the dairy cow breeding information according to pre-set keywords; According to a clustering algorithm with a distance threshold and a density threshold set, the first data to be labeled are clustered to obtain each first cluster set after clustering, and a first labeling process is performed; wherein the distance threshold is the maximum distance between each first data to be labeled in each type of the first cluster set after clustering; and the density threshold is the maximum value of the total number of each first data to be labeled in each type of the first cluster set; The second processing module is used for: Determine the associated data of each first cluster set as the second cluster set; Determine the hyperparameters of the single-class support vector machine algorithm and build a prediction model; Inputting the second cluster set into the prediction model, and outputting the prediction value corresponding to each second cluster set; Screening out a second cluster set whose predicted value is not within a preset prediction range, and performing a second marking process; The determination module is used to determine a final result of marking the abnormal data based on the first marking processing result and the second marking processing result.

[0019] Compared with the prior art, the present invention has the following beneficial effects: The present invention manages the breeding information of dairy cows to achieve scientific breeding, effectively reduce feeding costs, increase the milk production of dairy cows, and increase the economic income of dairy farms. At the same time, it ensures that the cows have sufficient nutrients to supplement the energy consumed by the cows in milk production, thereby improving the quality and output of milk. It can match response plans according to the growth conditions of dairy cows, quickly solve abnormal situations that occur in the breeding process of dairy cows, realize intelligent management of dairy cow breeding information, improve the quality of dairy cow breeding, and optimize resource utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 The following is a work flow chart of the intelligent management system for dairy cattle breeding information of the present invention; Figure 2 The module diagram of the intelligent management system for dairy cattle breeding information of the present invention is shown in FIG. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] In order to solve the problem that the existing patents cannot collect and manage the breeding information of dairy cows in actual use, resulting in the inability to find out the problems in the breeding process and provide corresponding solutions based on the breeding information of dairy cows, please refer to Figure 1-Figure 2 , this embodiment provides the following technical solutions: An intelligent management system for dairy cattle breeding information, comprising: The information collection module is used to preset the collection frequency, obtain dairy cattle breeding data according to the preset collection frequency, and clean the real-time collected data; The analysis and management module is used to classify the real-time collected data and analyze the growth information and behavior status data of each dairy cow based on the classification results; The statistical arrangement module is used to summarize, organize and count various dairy cow breeding information, and generate corresponding query passwords based on the statistical arrangement results. By statistics on the basic information of dairy cow breeding, it provides users with more intuitive and convenient information management.

[0023] The browsing query module is used to query dairy cow breeding information through the browser, display dairy cow breeding information and display solutions for dairy cows with abnormal breeding status. At the same time, the location of the dairy cow can be quickly located according to the dairy cow number, realizing visual query of dairy cow breeding information.

[0024] Information collection module, including: A real-time collection module is used to preset a collection frequency and obtain dairy cow breeding data according to the preset collection frequency. The dairy cow breeding data includes environmental influencing factors, dairy cow growth information, dairy cow business data and dairy cow behavior status data; The data cleaning module is used to judge the collected data, determine whether the collected data is missing or repeated, and process the missing and repeated data.

[0025] Data cleaning module, including: Fill in missing values: Determine whether the real-time collected data is missing. If so, use the preset rules to fill in the missing values. ‌Delete Duplicate Data‌: Determine whether there are duplicates in the real-time collected data. If there are duplicates, choose to keep the first or last row of data in each group of duplicate rows and delete the remaining rows to ensure the uniqueness and accuracy of the data; ‌Handling duplicate values‌: Check and remove invalid duplicate data through case processing, or mark the first duplicate value that appears; ‌Identify and handle outliers‌: Use the MAD method to identify outliers and replace them with appropriate values; ‌Data standardization‌: The min-max standardization method is used to eliminate dimensional differences between data to facilitate comparison and analysis‌.

[0026] By cleaning the collected data, errors, missing values, duplicate values, outliers, etc. in the data can be eliminated, thereby improving data quality and credibility, making the cleaned data more standardized and consistent, reducing obstacles in the data analysis process, improving analysis efficiency, and enabling data analysts to draw valuable insights more quickly. The cleaned data is more reliable and accurate, and can better support business and decision-making, thereby enhancing the value and utilization of data.

[0027] Analysis management module, including: A storage module is used to store the cleaned data according to environmental factors, growth information of dairy cows, business data of dairy cows and behavior status data of dairy cows, and set corresponding labels; The analysis module is used to analyze the health status of dairy cows based on environmental influencing factors, dairy cow growth information, dairy cow business data and dairy cow behavior status data; dairy cow growth information includes growth cycle, body temperature characteristics, feed formula and feeding ingredients, fur and eating and movement conditions; dairy cow behavior status data includes movement, rest and movement status; business data includes dairy cow weight, body size data records, milk production and calving records; environmental influencing factors include temperature, humidity and air quality; The management module is used to determine whether there are any abnormalities in the dairy cow breeding process according to the health status of the dairy cows, and take corresponding management measures based on the results.

[0028] Analyzing and improving the health status of dairy cows in combination with environmental factors can promote the growth and development of dairy cows and production performance, improve the immunity and disease resistance of dairy cows, reduce the occurrence of diseases, thereby ensuring the health and production efficiency of dairy cows, improving the welfare level of dairy cows, and thus improving the quality and output of milk. Analyzing the health status of dairy cows in combination with the growth information of dairy cows can optimize the feed utilization efficiency, make corresponding adjustments according to the growth stage of dairy cows, and achieve healthy breeding. It can effectively reduce or eliminate hidden dangers such as drug residues in dairy cow products, ensure the safety of milk and beef, thereby ensuring the health and life safety of consumers, promoting product exports, helping to reduce pollution to the environment, and effectively controlling the occurrence of malignant infectious diseases. This is of great significance for protecting the ecological environment and human health. By managing the breeding information of dairy cows, scientific breeding can be achieved, the feeding cost can be effectively reduced, the milk production of dairy cows can be increased, and the economic income of dairy farms can be increased. At the same time, it ensures that the nutrients consumed by cows are sufficient and supplements the energy consumed by dairy cows when producing milk, thereby improving the quality and output of milk.

[0029] Analysis module, including: A database establishment module is used to establish a dairy cattle breeding information management database based on the standard growth information and behavior status data of dairy cows at different growth stages; A retrieval module is used to retrieve the breeding information of the dairy cow at this stage according to the growth stage of the dairy cow, obtain the standard breeding information of the dairy cow, and retrieve the breeding information of the dairy cow; The comparison module is used to compare the dairy cow breeding information collected in real time with the retrieved dairy cow standard breeding information to determine the comparison result.

[0030] For example, when the real-time data of a dairy cow is in the adolescent stage, the collected dairy cow breeding information is compared with the standard breeding information of the dairy cow at this stage. If the comparison results are consistent, it means that there is no abnormality in the breeding of the dairy cow at this stage. If there is inconsistency in the comparison results, the inconsistent data is marked. If the weight or body size of the dairy cow is inconsistent, the data is marked, and the reasons why the weight or body size does not meet the standards are analyzed based on the data. The reasons why the weight or body size of the dairy cow does not meet the standards are generally caused by parasitic infection, insufficient nutrition supply or genetic reasons. According to the reasons why the weight of the dairy cow does not meet the standards, the health status of the dairy cow is analyzed.

[0031] Management modules, including: The judgment module is used to judge the health status of dairy cows and whether there are any abnormalities in their activity status based on the analysis results; The determination module is used to judge the problems existing in the dairy cow breeding process according to the determined results, and match corresponding solutions according to the existing problems.

[0032] The management process of the management module includes: The breeding information of the dairy cow is judged according to the analysis result of the analysis module to determine whether there is any abnormality in the breeding information of the dairy cow; If there is an abnormality, the abnormal data is marked, and the specific information in the dairy cattle breeding information represented by the marked abnormal data is determined; According to the determined results, the problems existing in the dairy cow breeding process are judged, and corresponding solutions are matched according to the existing problems.

[0033] Determine the workflow of the module, including: Obtain the influencing factors corresponding to the abnormal information in the historical dairy farming process, formulate corresponding solutions based on the abnormal information of dairy farming, and establish a solution database; Compare the marked abnormal data with the influencing factors corresponding to the abnormal information in the dairy farming process; Match the corresponding solution in the solution data according to the comparison result; For example, when the cow's appetite changes, it means that the cow's feeding amount is not up to standard. The matching solution is to adjust the cow's feeding amount according to the cow's current growth stage. When the cow's fur is dull, it means that the cow grows in a humid environment, is malnourished or infected with parasites. The matching solution is to improve the cow's growth environment, improve its diet or carry out parasite control. The response plan can be matched according to the growth of the cow, and abnormal situations in the cow breeding process can be quickly resolved. The intelligent management of dairy cow breeding information is realized, the quality of dairy cow breeding is improved, and the use of resources is optimized. The judgment module includes: The first judgment module is used to judge the health status of dairy cow breeding according to the growth information and business data of dairy cows; By collecting and analyzing data on the health status and production performance of dairy cows, we can promptly discover problems in feeding management, such as whether the feed formula is reasonable, whether the feeding environment is suitable, etc., so as to adjust the feeding strategy in time, ensure that dairy cows get balanced nutrition and a comfortable living environment, and promote the health of dairy cows. By analyzing the milk production data of dairy cows, we can understand the production performance of dairy cows of different breeds, as well as the impact of environmental factors and feeding factors on milk production, so as to optimize breed selection, improve the feeding environment and adjust the feeding method, increase milk production and milk quality, and increase economic benefits. By monitoring the health data of dairy cows, we can promptly Discover potential disease problems. For example, by analyzing the body temperature, behavior and other data of dairy cows, the occurrence of diseases can be predicted, so that corresponding preventive measures can be taken to reduce the impact of diseases on the health and production performance of dairy cows. According to the individual differences and nutritional needs of dairy cows, personalized feeding plans can be formulated through data analysis to achieve precise feeding, ensuring that each dairy cow can obtain the right amount of nutrition and avoid overnutrition or undernutrition, which is beneficial to the health of dairy cows and the improvement of milk production performance. Through data analysis, the breeding process can be optimized and unnecessary inputs can be reduced, such as reducing feed waste and reducing disease treatment costs, thereby improving breeding efficiency and economic benefits; The second judgment module is used to judge the activity status of dairy cow farming according to the behavior status data of the dairy cow.

[0034] By analyzing the behavior status data of dairy cows, the health status of dairy cows can be effectively monitored. The behavior of dairy cows, such as eating, drinking, resting, and exercising, can directly reflect their physiological and psychological states, and thus their health status can be evaluated. For example, the eating behavior of dairy cows is an important indicator for evaluating their digestive system and overall health. If the amount of food intake of dairy cows suddenly decreases, it may indicate that the dairy cows have health problems, such as digestive system diseases or discomfort. Similarly, the rest and exercise patterns of dairy cows can also provide clues about their health status. For example, if the rest time of dairy cows is too long or too short, it may indicate health problems or environmental discomfort. In addition, the social behavior of dairy cows, such as communication and interaction within the group, is also an important aspect of evaluating their mental health and welfare. Cows communicate through sounds, smells, etc. These behaviors can reflect whether the cows feel comfortable and safe. If cows show abnormal social behaviors, such as excessive tension or increased aggression, it may indicate that they are feeling uneasy or stressed. Statistics and collation modules, including: The data analysis report generation module is used to receive a request for report generation, retrieve the analysis results obtained by the analysis management module according to the received report generation request, and generate corresponding report information; The statistical result display module is used to visually display the report information generated by the report generation module.

[0035] Working principle: When using the intelligent management system for dairy cattle breeding information of the present invention, according to Figure 1 and Figure 2 , including the following steps: Step 1: Preset the collection frequency, obtain dairy cow breeding data according to the preset collection frequency, and clean the real-time collected data; Step 2: Classify the data collected in real time, and analyze the growth information and behavior status data of each cow according to the classification results. According to the analysis results of the analysis module, judge the breeding information of the cow to determine whether there is any abnormality in the breeding information of the cow; Step 3: If there is an abnormality, the abnormal data is marked, and the specific information in the dairy cattle breeding information represented by the marked abnormal data is determined; Step 4: Determine the problems existing in the dairy cow breeding process based on the determined results, and match corresponding solutions according to the existing problems; Step 5: Summarize, organize and count various dairy farming information and corresponding solutions, and generate corresponding query passwords based on the statistical results. By counting the basic information of dairy farming, users can be provided with more intuitive and convenient information management.

[0036] Step 6: Query the dairy cow breeding information through the browser. According to the query results, the dairy cow breeding information and the display solution for the dairy cows with abnormal breeding status are displayed. At the same time, the location of the dairy cow can be quickly located according to the dairy cow number, realizing the visual query of dairy cow breeding information.

[0037] By cleaning the collected data, errors, missing values, duplicate values, outliers, etc. in the data can be eliminated, thereby improving data quality and credibility, making the cleaned data more standardized and consistent, reducing obstacles in the data analysis process, improving analysis efficiency, and enabling data analysts to draw valuable insights more quickly. The cleaned data is more reliable and accurate, and can better support business and decision-making, thereby enhancing the value and utilization of data. It can match response plans according to the growth of dairy cows, quickly resolve abnormal situations that occur during the breeding process of dairy cows, and realize intelligent management of dairy cow breeding information, thereby improving the quality of dairy cow breeding and optimizing resource utilization. By managing the breeding information of dairy cows, scientific breeding can be achieved, effectively reducing feeding costs, increasing the milk production of dairy cows, and increasing the economic income of dairy farms. At the same time, it ensures that the cows have sufficient nutrients to supplement the energy consumed by dairy cows when producing milk, thereby improving the quality and output of milk.

[0038] The analysis and management module classifies the data collected in real time, including: Construct the target loss learning function of the data classification model; Classify the real-time collected data based on the target loss learning function; Construct the target loss learning function of the data classification model, including: Get the construction data set; divide each construction data in the construction data set into sub-data, obtain the data category of each sub-data, and then construct the initial loss learning function :

[0039] in, Initial loss learning function for construction; is the bias parameter after initialization; is a natural constant; To build data Middle The category of each sub-data; Structuring data for prediction Middle The probability that the category of each sub-data is accurate; is the category of the labeled sub-data; Adjust the initialized bias parameters during the construction process;

[0040] in, is the adjusted bias parameter; is the loss learning rate of the loss learning function, with an initial value of 0.002; Learn a function for the initial loss After initialization, the bias parameters Conduct partial guidance; According to the adjusted bias parameters, the initial loss learning function is adjusted to obtain the target loss learning function ;

[0041] in, For The set of categories predicted by each sub-data; is the category of the unlabeled sub-data.

[0042] The working principle and beneficial effects of the above technical solution: The analysis and management module classifies the real-time collected data based on the target loss learning function of the data classification model, which facilitates accurate data classification. When constructing the target loss learning function of the data classification model, first obtain the construction data set for building and training the classification model. Each data item in the construction data set is divided into N sub-data, and the true category label of each sub-data is determined. The initial loss learning function is constructed based on the difference between the predicted category and the true category. During the construction process, the initialization bias parameter is adjusted; according to the adjusted bias parameter, the initial loss learning function is adjusted to obtain the target loss learning function, which is based on the target loss learning function to facilitate accurate classification of the real-time collected data.

[0043] Preferably, the management module includes: The first processing module is used for: Extracting first data to be labeled from the dairy cow breeding information according to pre-set keywords; According to a clustering algorithm with a distance threshold and a density threshold set, the first data to be labeled are clustered to obtain each first cluster set after clustering, and a first labeling process is performed; wherein the distance threshold is the maximum distance between each first data to be labeled in each type of the first cluster set after clustering; and the density threshold is the maximum value of the total number of each first data to be labeled in each type of the first cluster set; The second processing module is used for: Determine the associated data of each first cluster set as the second cluster set; Determine the hyperparameters of the single-class support vector machine algorithm and build a prediction model; Inputting the second cluster set into the prediction model, and outputting the prediction value corresponding to each second cluster set; Screening out a second cluster set whose predicted value is not within a preset prediction range, and performing a second marking process; The determination module is used to determine a final result of marking the abnormal data based on the first marking processing result and the second marking processing result.

[0044] The working principle of the above technical solution: In this embodiment, the keywords include key information related to dairy farming, such as "abnormal milk production", "abnormal feed consumption", "abnormal health status", "abnormal breeding records", etc. Traverse the entire dairy farming information database, use text mining or regular expression techniques, filter out records or data fragments containing these keywords according to predefined keywords, and extract the first data to be marked. Distance threshold: defines the farthest distance at which a point can be considered to belong to a cluster. That is, if the distance between a point and the center of the cluster is less than or equal to this threshold, it is considered to be part of the cluster. Density threshold: defines the minimum number of data points in a cluster. If the number of points in an area exceeds this threshold, the area forms a cluster. Clustering algorithms such as DBSCAN cluster the first data to be marked to obtain each first cluster set after clustering, which is convenient for classifying direct abnormal data and marking it.

[0045] In this embodiment, according to business needs and data characteristics, determine which data is closely related to the first cluster set. For example, if the first cluster set is divided based on the abnormal milk production of dairy cows, then the associated data may include the breed, age, feed type, etc. of the dairy cows. A one-class support vector machine (One-Class SVM) algorithm is used to detect the second cluster set that is abnormal or does not conform to the expected pattern. The main hyperparameters of the one-class SVM include kernel functions (such as RBF, polynomials, etc.), penalty parameters (nu or C), and parameters of the kernel function (such as gamma). Based on the distribution of the prediction problem and the data, the cross-validation method is used to select appropriate nu and gamma values, and the one-class SVM model, that is, the prediction model, is trained based on the RBF kernel function to output the prediction value corresponding to each second cluster set. The preset prediction range represents the corresponding data range when normal. The second cluster set whose prediction value is not within the preset prediction range is screened out, and a second labeling process is performed; it is used to determine indirect abnormal data and perform labeling processing.

[0046] The beneficial effects of the above technical solution are as follows: direct abnormal data is determined from the breeding information of dairy cows, and the first marking process is performed; the associated data of each first cluster set is determined as the second cluster set, and the indirect abnormal data is determined from the second cluster set, and the second marking process is performed. The final result of marking the abnormal data is determined based on the first marking process result and the second marking process result, so as to achieve comprehensiveness and accuracy of abnormal data marking.

[0047] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0048] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. An intelligent management system for dairy cattle breeding information, characterized in that: include: An information collection module is used to preset a collection frequency and obtain dairy cattle breeding data according to the preset collection frequency; An analysis management module is used to construct a target loss learning function of a data classification model, classify the real-time collected data based on the target loss learning function, and analyze the growth information and behavior status data of each dairy cow. According to the analysis results, direct abnormal data is determined from the breeding information of the dairy cow, and a first labeling process is performed. The associated data of each first cluster set is determined as a second cluster set, and indirect abnormal data is determined from the second cluster set. A second labeling process is performed, and a final result of labeling the abnormal data is determined based on the first labeling process result and the second labeling process result; The statistical arrangement module is used to summarize, arrange and count various dairy cow breeding information and generate corresponding query passwords; The browsing query module is used to query and display dairy cow breeding information and quickly locate the location of dairy cows.

2. The intelligent management system for dairy cattle breeding information according to claim 1 is characterized in that ,The information acquisition module includes: A real-time collection module is used to preset a collection frequency and obtain dairy cow breeding data according to the preset collection frequency. The dairy cow breeding data includes environmental influencing factors, dairy cow growth information, dairy cow business data and dairy cow behavior status data; The data cleaning module is used to judge the collected data, determine whether the collected data is missing or repeated, and process the missing and repeated data; Data cleaning module, including: Fill in missing values: Determine whether the real-time collected data is missing. If so, use the preset rules to fill in the missing values. ‌Delete Duplicate Data‌: Determine whether there are duplicates in the real-time collected data. If there are duplicates, choose to keep the first or last row of data in each group of duplicate rows and delete the remaining rows; ‌Handling duplicate values‌: Check and remove invalid duplicate data through case processing, or mark the first duplicate value that appears; ‌Identify and handle outliers‌: Use the MAD method to identify outliers and replace them with appropriate values; ‌Data normalization‌: Eliminate dimensional differences between data through min-max normalization method‌.

3. The intelligent management system for dairy cattle breeding information according to claim 1 is characterized in that ,The analysis management module includes: A storage module is used to store the cleaned data according to environmental factors, growth information of dairy cows, business data of dairy cows and behavior status data of dairy cows, and set corresponding labels; An analysis module is used to analyze the health status of dairy cows based on environmental factors, growth information of dairy cows, business data of dairy cows, and behavioral status data of dairy cows; The management module is used to determine whether there are any abnormalities in the dairy cow breeding process according to the health status of the dairy cows, and take corresponding management measures based on the results.

4. The intelligent management system for dairy cattle breeding information according to claim 3 is characterized in that ,The analysis module specifically includes: A database establishment module is used to establish a dairy cattle breeding information management database based on the standard growth information and behavior status data of dairy cows at different growth stages; A retrieval module is used to retrieve the breeding information of the dairy cow at this stage according to the growth stage of the dairy cow, obtain the standard breeding information of the dairy cow, and retrieve the breeding information of the dairy cow; The comparison module is used to compare the dairy cow breeding information collected in real time with the retrieved dairy cow standard breeding information to determine the comparison result.

5. The intelligent management system for dairy cattle breeding information according to claim 4 is characterized in that ,The management module includes: The judgment module is used to judge the health status of dairy cows and whether there are any abnormalities in their activity status based on the analysis results; A determination module is used to determine the problems existing in the dairy cow breeding process according to the determined results, and match corresponding solutions according to the existing problems; Determine the workflow of the module, including: Obtain the influencing factors corresponding to the abnormal information in the historical dairy farming process, formulate corresponding solutions based on the abnormal information of dairy farming, and establish a solution database; Compare the marked abnormal data with the influencing factors corresponding to the abnormal information in the dairy farming process; According to the comparison results, the corresponding solution is matched in the solution data.

6. The intelligent management system for dairy cattle breeding information according to claim 5, characterized in that ,The management process of the management module specifically includes: The breeding information of the dairy cow is judged according to the analysis result of the analysis module to determine whether there is any abnormality in the breeding information of the dairy cow; If there is an abnormality, the abnormal data is marked, and the specific information in the dairy cattle breeding information represented by the marked abnormal data is determined; According to the determined results, the problems existing in the dairy cow breeding process are judged, and corresponding solutions are matched according to the existing problems.

7. The intelligent management system for dairy cattle breeding information according to claim 5, characterized in that ,The judgment module includes: The first judgment module is used to judge the health status of dairy cow breeding according to the growth information and business data of dairy cows; The second judgment module is used to judge the activity status of dairy cow farming according to the behavior status data of the dairy cow.

8. The intelligent management system for dairy cattle breeding information according to claim 1, characterized in that ,The statistical arrangement module includes: The data analysis report generation module is used to receive a request for report generation, retrieve the analysis results obtained by the analysis management module according to the received report generation request, and generate corresponding report information; The statistical result display module is used to visually display the report information generated by the report generation module.

9. The intelligent management system for dairy cattle breeding information according to claim 1, characterized in that ,The target loss learning function of constructing the data classification model includes: Get the construction data set; divide each construction data in the construction data set into sub-data, obtain the data category of each sub-data, and then construct the initial loss learning function : ； in, Initial loss learning function for construction; is the bias parameter after initialization; is a natural constant; To build data Middle The category of each sub-data; Structuring data for prediction Middle The probability that the category of each sub-data is accurate; is the category of the labeled sub-data; Adjust the initialized bias parameters during the construction process; ； in, is the adjusted bias parameter; is the loss learning rate of the loss learning function, with an initial value of 0.002; Learn a function for the initial loss After initialization, the bias parameters Conduct partial guidance; According to the adjusted bias parameters, the initial loss learning function is adjusted to obtain the target loss learning function ; ； in, For The set of categories predicted by each sub-data; is the category of the unlabeled sub-data.

10. The intelligent management system for dairy cattle breeding information according to claim 3, characterized in that , the management module includes: The first processing module is used for: Extracting first data to be labeled from the dairy cow breeding information according to pre-set keywords; According to a clustering algorithm with a distance threshold and a density threshold set, the first data to be labeled are clustered to obtain each first cluster set after clustering, and a first labeling process is performed; wherein the distance threshold is the maximum distance between each first data to be labeled in each type of the first cluster set after clustering; and the density threshold is the maximum value of the total number of each first data to be labeled in each type of the first cluster set; The second processing module is used for: Determine the associated data of each first cluster set as the second cluster set; Determine the hyperparameters of the single-class support vector machine algorithm and build a prediction model; Inputting the second cluster set into the prediction model, and outputting the prediction value corresponding to each second cluster set; Screening out a second cluster set whose predicted value is not within a preset prediction range, and performing a second marking process; The determination module is used to determine a final result of marking the abnormal data based on the first marking processing result and the second marking processing result.

Images