Method for predicting slaughtering performance of local pigs, storage medium and electronic equipment
By collecting and filtering live body size data of local pigs, random forest and support vector machine models were constructed to solve the problem of difficulty in predicting slaughter performance caused by the small number of local pig populations. This resulted in high-precision slaughter performance prediction and supported accurate prediction of carcass weight and lean meat weight.
Patent Information
- Application Number
- CN202511494129.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-20
- Publication Date
- 2026-01-30
AI Technical Summary
Existing technologies struggle to accurately predict slaughter performance in local pig populations with relatively small numbers, due to issues such as insufficient sample size, lack of representativeness, and limited model generalization ability.
By collecting live body size data of local pigs, using the Akaike Information Criterion to filter the body size data, constructing random forest and support vector machine models, and selecting the optimal prediction model to predict slaughter performance, including carcass weight, lean meat weight, and fat weight.
It has achieved high-precision prediction of the slaughter performance of local pigs, with prediction errors of less than 5% for carcass weight and less than 10% for lean meat weight, providing a scientific basis for precision feeding management and breeding selection.
Smart Images
Figure CN121435181A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of livestock breeding technology, and in particular to a method, storage medium, and electronic device for predicting the slaughter performance of local pigs based on live body size data of local pigs. Background Technology
[0002] In modern pig farming and meat production, slaughter performance indicators such as carcass weight, lean meat weight, and fat weight are crucial for measuring individual production performance and determining meat grading and market value. Accurate and reliable prediction of these slaughter performance indicators during the live stage not only helps optimize feeding management and slaughter scheduling but also enables the early identification of high-yielding individuals in breeding selection, thereby improving economic benefits and production efficiency.
[0003] Considering that Hunan local pigs are a typical high-quality local livestock and poultry genetic resource, possessing excellent characteristics such as tender meat, unique flavor, and strong adaptability, their limited population size, slow breeding rate, and significant individual differences in growth and development pose challenges to the accurate prediction of slaughter performance and production decisions.
[0004] Currently, methods for predicting pig slaughter performance using live pig body size data (body size data refers to phenotypic indicators obtained by measuring live pigs, such as weight, body length, body height, chest circumference, back height, chest depth, abdominal circumference, and backfat thickness. These indicators are correlated with slaughter performance and can therefore be used as predictive factors) have a certain research foundation in introduced breeds or large-scale farming populations. However, their application in local pig populations with relatively small numbers remains significantly limited. This is because the relatively small number of local pig populations results in a limited number of training samples, leading to the following shortcomings in building big data models: 1) Insufficient sample size leads to a decrease in model accuracy and stability. For example, a small sample size makes complex models (such as deep learning) prone to overfitting, and simple models prone to underfitting, making it difficult to accurately characterize the complex nonlinear relationship between body size indicators and slaughter performance.
[0005] 2) Insufficient sample representativeness leads to prediction bias. For example, local pig data often comes from concentrated sources and covers limited feeding conditions, age structure and production batches, resulting in insufficient sample representativeness and prediction bias.
[0006] 3) The model's generalization ability is limited. For example, a prediction model built for foreign pigs may perform well within its own population, but it is difficult to directly generalize to local pig populations.
[0007] Therefore, it is necessary to provide a method, storage medium, and electronic device for predicting the slaughter performance of local pigs based on live body size data of local pigs, in order to solve the problem that existing slaughter performance prediction methods cannot be applied to local pig populations with relatively small numbers. Summary of the Invention
[0008] The purpose of this invention is to provide a method, storage medium, and electronic device for predicting the slaughter performance of local pigs based on live body size data. The specific technical solution is as follows: In a first aspect, the present invention provides a method for predicting local pig slaughter performance, comprising: Step S1: Collect live body size data of local pigs, remove outliers from the collected body size data, and obtain the body size dataset. Step S2: The Akaike Information Criterion is used to filter each body size data in the body size dataset to obtain a candidate set of body size data related to local pig slaughter performance. Step S3: Using each body size data in the candidate set of body size data as the independent variable and the slaughter performance as the dependent variable, construct a first random forest model; output a first predicted slaughter performance value through the first random forest model; calculate a first error value between the first predicted slaughter performance value and the actual slaughter performance value; if the first error value exceeds an error threshold and the coefficient of determination between the first predicted slaughter performance value and the actual slaughter performance value is less than 0.7, then remove the body size data corresponding to the first predicted slaughter performance value; if the first error value is within the error threshold and the coefficient of determination between the first predicted slaughter performance value and the actual slaughter performance value is greater than or equal to 0.7, then determine the body size data corresponding to the first predicted slaughter performance value as a key variable; form a candidate variable set from each of the key variables. Step S4: Using the body size data from the candidate variable set as independent variables and the slaughter performance as the dependent variable, construct a second random forest model and a support vector machine model respectively; output a second predicted slaughter performance value through the second random forest model; output a third predicted slaughter performance value through the support vector machine model; calculate a second error value of the second predicted slaughter performance value relative to the actual slaughter performance value; calculate a third error value of the third predicted slaughter performance value relative to the actual slaughter performance value; select the minimum error value from the second and third error values, and the model corresponding to the minimum error value is the optimal prediction model for the slaughter performance of local pigs, and predict the slaughter performance of local pigs using the optimal prediction model.
[0009] Optionally, in step S1, the body measurements include weight, body length, chest circumference, body height, back height, chest depth, abdominal circumference, tube circumference, and backfat thickness.
[0010] Optionally, in step S1, outliers are removed from the collected body size data using the 3x standard deviation method.
[0011] Optionally, in step S2, if the output value of a certain body size data in the body size dataset after being substituted into the Akaike Information Criterion is greater than the criterion threshold, then the body size data is removed; if the output value of a certain body size data in the body size dataset after being substituted into the Akaike Information Criterion is less than or equal to the criterion threshold, then it is included in the candidate set of body size data; the criterion threshold is the minimum value of the Akaike Information Criterion in the current body size dataset plus 2.
[0012] Optionally, in step S3, the slaughter performance includes carcass weight, lean meat weight, and fat weight.
[0013] Optionally, in step S3, the first error value includes the root mean square error and the mean absolute error of the first slaughter performance prediction value; wherein, the root mean square error of the first slaughter performance prediction value is calculated using equation (1); and the mean absolute error of the first slaughter performance prediction value is calculated using equation (2). (1); (2); in, Indicates the first Predicted first slaughter performance of a live pig; Indicates the first Actual slaughter performance of a live pig; This represents the total number of local pigs collected. The value range is 1~ Any natural number in the range; RMSE This represents the root mean square error of the first predicted slaughter performance value, and its corresponding error threshold range is 5% to 15% of the arithmetic mean of the actual slaughter performance values of all live pigs. MAE This represents the mean absolute error of the first predicted slaughter performance value, with the corresponding error threshold range being 3% to 10% of the arithmetic mean of the actual slaughter performance values of all live pigs. The coefficient of determination is calculated using equation (3): (3); in, This represents the arithmetic mean of the actual slaughter performance values of all locally sourced pigs collected. R 2 The coefficient of determination represents the first predicted slaughter performance value relative to the actual slaughter performance value, and its value is greater than or equal to 0.7. In the first predicted slaughter performance value RMSE and MAE All are within their respective error thresholds, and the determination coefficient of the first predicted slaughter performance value relative to the actual slaughter performance value is... R 2If the value is greater than or equal to 0.7, then the body size data corresponding to the first predicted slaughter performance value is determined as the key variable.
[0014] Optionally, in step S4, the second error value is the comprehensive error value of the second slaughter performance prediction value, which is obtained by weighted summation of the normalized root mean square error value and the mean absolute percentage error value; wherein, the normalized root mean square error value of the second slaughter performance prediction value is calculated by combining equations (4) and (5); the mean absolute percentage error value of the second slaughter performance prediction value is calculated by equation (6); (4); (5); (6); in, This represents the second predicted slaughter performance value; This represents the normalized root mean square error of the second slaughter performance prediction value; This represents the root mean square error of the second predicted slaughter performance value; This represents the mean absolute percentage error of the second slaughter performance prediction value.
[0015] Optionally, in step S4, the third error value is the comprehensive error value of the third slaughter performance prediction value, which is obtained by weighted summation of the normalized root mean square error value and the mean absolute percentage error value; wherein, the normalized root mean square error value of the third slaughter performance prediction value is calculated by combining equations (7) and (8); the mean absolute percentage error value of the third slaughter performance prediction value is calculated by equation (9); (7); (8); (9); in, This represents the third predicted slaughter performance value; This represents the normalized root mean square error of the third predicted slaughter performance value. This represents the root mean square error of the third predicted slaughter performance value; This represents the mean absolute percentage error of the third slaughter performance prediction value.
[0016] In a second aspect, the present invention provides a storage medium having stored computer program instructions thereon, which, when executed by a processor, implement the method for predicting local pig slaughter performance.
[0017] In a third aspect, the present invention provides an electronic device comprising: at least one processor, at least one memory, and computer program instructions stored in the memory, wherein the computer program instructions, when executed by the processor, implement the method for predicting local pig slaughter performance.
[0018] The application of the technical solution of the present invention has at least the following beneficial effects: This invention proposes a method for predicting the slaughter performance of local pigs, applicable to the prediction of slaughter performance in local pig populations with relatively small numbers. Specifically, this invention employs step S1 to collect a live body size dataset of local pigs; step S2 to obtain a candidate set of body size data associated with the slaughter performance of local pigs; step S3 to obtain a candidate variable set composed of the aforementioned key variables; and step S4 to obtain an optimal prediction model for the slaughter performance of local pigs. The optimal prediction model predicts the slaughter performance of local pigs, achieving prediction errors of less than 5% for carcass weight and less than 10% for lean meat weight, demonstrating the high accuracy and reliability of the optimal prediction model in practical applications. Therefore, this invention can quickly and accurately predict the slaughter performance of pigs, such as carcass weight and lean meat weight, thereby enabling early assessment of the potential meat production performance of individuals and providing a scientific basis for precision feeding management, phased feeding optimization, slaughter planning optimization, and breeding selection.
[0019] In addition to the objectives, features, and advantages described above, the present invention has other objectives, features, and advantages. The invention will now be described in further detail with reference to the figures. Attached Figure Description
[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.
[0021] Figure 1 The optimal prediction model predicts the carcass weight of local pigs. Figure 2 The optimal prediction model predicts the fat weight of local pigs. Figure 3 The result of the optimal prediction model for the lean weight of local pigs. 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 a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.
[0023] Example: A method for predicting the slaughter performance of local pigs includes: Step S1: Collect live body size data of local pigs (specifically, Hunan local pigs, a total of 120 individuals), remove outliers from the collected body size data to obtain the body size dataset; the measuring tools and precision used when collecting body size data are as follows: soft ruler or calipers, with a precision of 0.1cm; weight scale, with a precision of 0.1kg; Step S2: The Akaike Information Criterion is used to filter each body size data in the body size dataset to obtain a candidate set of body size data associated with the local pig slaughter performance (the specific body size data are pre-slaughter weight, height, length, chest circumference, chest depth, back height, abdominal circumference, cannon bone circumference, and backfat thickness). Step S3: Using each body size data in the candidate set of body size data as the independent variable and the slaughter performance as the dependent variable, construct a first random forest model; output a first predicted slaughter performance value through the first random forest model; calculate a first error value between the first predicted slaughter performance value and the actual slaughter performance value; if the first error value exceeds an error threshold and the coefficient of determination between the first predicted slaughter performance value and the actual slaughter performance value is less than 0.7, then remove the body size data corresponding to the first predicted slaughter performance value; if the first error value is within the error threshold and the coefficient of determination between the first predicted slaughter performance value and the actual slaughter performance value is greater than or equal to 0.7, then determine the body size data corresponding to the first predicted slaughter performance value as a key variable; form a candidate variable set from each of the key variables. Step S4: Using the body size data from the candidate variable set as independent variables and the slaughter performance as the dependent variable, construct a second random forest model and a support vector machine model respectively; output a second predicted slaughter performance value through the second random forest model; output a third predicted slaughter performance value through the support vector machine model; calculate a second error value of the second predicted slaughter performance value relative to the actual slaughter performance value; calculate a third error value of the third predicted slaughter performance value relative to the actual slaughter performance value; select the minimum error value from the second and third error values, and the model corresponding to the minimum error value is the optimal prediction model for the slaughter performance of local pigs (specifically, the optimal prediction model for lean meat weight is the support vector machine model, and the optimal prediction model for carcass weight and fat weight is the second random forest model), and predict the slaughter performance of local pigs using the optimal prediction model.
[0024] In step S1, the body measurements include weight, body length (from the tip of the pig's nose to the base of its tail), chest circumference (the maximum circumference of the thoracic cavity at the front shoulder of the pig), body height (from the highest point of the pig's shoulder blade to the ground), back height (from the highest point of the pig's spine to the ground), chest depth (from the highest point of the pig's shoulder blade to the bottom of the sternum), abdominal circumference (the maximum circumference of the pig's abdomen), cannon bone circumference (the circumference of the middle part of the cannon bone of the pig's foreleg), and backfat thickness (the thickness of subcutaneous fat in the middle of the spine between the 3rd and 4th ribs of the pig).
[0025] In step S1, outliers are removed from the collected body size data using the three-standard-deviation method. Specifically, the three-standard-deviation method uses the following formula to remove outliers from the collected body size data: ; in, Indicates the first The first live pig Individual ruler data measurement values; This indicates all the first-order live pigs collected. The mean of individual scale data; This indicates all the first-order live pigs collected. Standard deviation of individual scale data; The value range is 1~ Any natural number in the range; This represents the total number of local pigs collected, specifically 120. The value range is 1~ Any natural number in the range; The total number of body measurements collected (specifically, in this embodiment, weight, length, chest circumference, height, back height, chest depth, abdominal circumference, cannon circumference, and backfat thickness, totaling 9 items).
[0026] In step S2, if the output value of a certain body size data in the body size dataset after being substituted into the Akaike Information Criterion is greater than the criterion threshold, then the body size data is removed; if the output value of a certain body size data in the body size dataset after being substituted into the Akaike Information Criterion is less than or equal to the criterion threshold, then it is included in the candidate set of body size data; the criterion threshold is the minimum value of the Akaike Information Criterion in the current body size dataset plus 2.
[0027] The Akaike Information Criterion is expressed by the following formula: ; in, AIC This represents the value output by the Akaike Information Criterion; express + , This represents the total number of data items collected in body size data. This represents the intercept, which takes a value of 1. L This represents the maximum value of the likelihood function obtained by performing a log-likelihood fit on each body size data in the current body size dataset.
[0028] In step S3, the slaughter performance includes carcass weight, lean meat weight, and fat weight.
[0029] In step S3, the first error value includes the root mean square error and the mean absolute error of the first slaughter performance prediction value; wherein, the root mean square error of the first slaughter performance prediction value is calculated using equation (1); and the mean absolute error of the first slaughter performance prediction value is calculated using equation (2). (1); (2); in, Indicates the first Predicted first slaughter performance of a live pig; Indicates the first Actual slaughter performance of a live pig; This represents the total number of local pigs collected. The value range is 1~ Any natural number in the range; RMSE This represents the root mean square error of the first predicted slaughter performance value, and its corresponding error threshold range is 5% to 15% of the arithmetic mean of the actual slaughter performance values of all live pigs. MAE This represents the mean absolute error of the first predicted slaughter performance value, with the corresponding error threshold range being 3% to 10% of the arithmetic mean of the actual slaughter performance values of all live pigs. The coefficient of determination is calculated using equation (3): (3); in, This represents the arithmetic mean of the actual slaughter performance values of all locally sourced pigs (i.e., all live pigs collected). R 2 The coefficient of determination represents the first predicted slaughter performance value relative to the actual slaughter performance value, and its value is greater than or equal to 0.7. In the first predicted slaughter performance value RMSE and MAE All are within their respective error thresholds, and the determination coefficient of the first predicted slaughter performance value relative to the actual slaughter performance value is... R 2 If the value is greater than or equal to 0.7, then the body size data corresponding to the first predicted slaughter performance value is determined as the key variable.
[0030] In step S4, the second error value is the comprehensive error value of the second slaughter performance prediction value, which is obtained by weighted summation of the normalized root mean square error value and the mean absolute percentage error value; wherein, the normalized root mean square error value of the second slaughter performance prediction value is calculated by combining equations (4) and (5); the mean absolute percentage error value of the second slaughter performance prediction value is calculated by equation (6); (4); (5); (6); in, This represents the second predicted slaughter performance value; This represents the normalized root mean square error of the second slaughter performance prediction value; This represents the root mean square error of the second predicted slaughter performance value; This represents the mean absolute percentage error of the second slaughter performance prediction value.
[0031] In step S4, the third error value is the comprehensive error value of the third slaughter performance prediction value, which is obtained by weighted summation of the normalized root mean square error value and the mean absolute percentage error value; wherein, the normalized root mean square error value of the third slaughter performance prediction value is calculated by combining equations (7) and (8); the mean absolute percentage error value of the third slaughter performance prediction value is calculated by equation (9); (7); (8); (9); in, This represents the third predicted slaughter performance value; The normalized root mean square error of the third slaughter performance prediction value reflects the overall deviation between the predicted value and the actual value. This represents the root mean square error of the third predicted slaughter performance value; This represents the mean absolute percentage error of the third slaughter performance prediction value, reflecting the average level of the relative error of the prediction value.
[0032] See Table 1 and Figures 1-3 The results of the optimal prediction model for predicting the slaughter performance of local pigs, as determined in this embodiment, show that the lower the values of the normalized root mean square error and the mean absolute percentage error, the higher the accuracy of the optimal prediction model. Specifically, the prediction error for carcass weight is less than 5%, demonstrating excellent prediction accuracy; the prediction error for lean meat weight is controlled within 10%, indicating that the optimal prediction model has good fitting ability and generalization performance on this indicator.
[0033] The optimal prediction model for local pig slaughter performance established in this invention can predict fat weight. Results show that the normalized root mean square error (RMSE) for fat weight is 12.31%, and the mean absolute percentage error (MASE) is 10.02%, slightly higher than the commonly used accuracy standards for some slaughter performance indicators, but still reaching a reliable prediction level. This is because fat weight is significantly affected by genetic background, sex, feeding stage, and environmental factors, with significant individual differences. Compared to carcass weight or lean meat weight, body size data has lower explanatory power for fat weight, leading to a larger prediction error. However, the optimal prediction model used in the embodiments of this invention can effectively capture the overall trend and population differences in fat weight, making the prediction results still have practical reference value in population management, breeding selection, and feeding decisions. Therefore, although the prediction error for fat weight is relatively large in the embodiments of this invention, it can still provide reliable trend information in practical applications, meeting the needs of local pig slaughter performance assessment and management.
[0034] Table 1. Results of the optimal prediction model for the slaughter performance of local pigs.
[0035] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural transformations made using the contents of the present invention's specification and drawings under the inventive concept of the present invention, or direct / indirect applications in other related technical fields, are included within the patent protection scope of the present invention.
Claims
1. A method for predicting the slaughter performance of local pigs, characterized in that, include: Step S1: Collect live body size data of local pigs, remove outliers from the collected body size data, and obtain the body size dataset. Step S2: The Akaike Information Criterion is used to filter each body size data in the body size dataset to obtain a candidate set of body size data related to local pig slaughter performance. Step S3: Using each body size data in the candidate set of body size data as the independent variable and the slaughter performance as the dependent variable, construct a first random forest model; output a first predicted slaughter performance value through the first random forest model; calculate a first error value between the first predicted slaughter performance value and the actual slaughter performance value; if the first error value exceeds an error threshold and the coefficient of determination between the first predicted slaughter performance value and the actual slaughter performance value is less than 0.7, then remove the body size data corresponding to the first predicted slaughter performance value; if the first error value is within the error threshold and the coefficient of determination between the first predicted slaughter performance value and the actual slaughter performance value is greater than or equal to 0.7, then determine the body size data corresponding to the first predicted slaughter performance value as a key variable; form a candidate variable set from each of the key variables. Step S4: Using the body size data from the candidate variable set as independent variables and the slaughter performance as the dependent variable, construct a second random forest model and a support vector machine model respectively; output a second slaughter performance prediction value through the second random forest model; output a third slaughter performance prediction value through the support vector machine model; Calculate the second error value between the second predicted slaughter performance value and the actual slaughter performance value; Calculate the third error value of the predicted third slaughter performance value relative to the actual slaughter performance value; select the minimum error value from the second and third error values, and the model corresponding to the minimum error value is the optimal prediction model for the slaughter performance of local pigs, and the slaughter performance of local pigs is predicted by the optimal prediction model.
2. The method for predicting local pig slaughter performance as described in claim 1, characterized in that, In step S1, the body measurements include weight, length, chest circumference, height, back height, chest depth, abdominal circumference, tube circumference, and backfat thickness.
3. The method for predicting local pig slaughter performance as described in claim 1, characterized in that, In step S1, outliers are removed from the collected body size data using the 3x standard deviation method.
4. The method for predicting local pig slaughter performance as described in claim 1, characterized in that, In step S2, if the value output by substituting a certain body size data in the body size dataset into the Akaike Information Criterion is greater than the criterion threshold, then the body size data is removed. If the value output by substituting a certain body size data in the body size dataset into the Akaike Information Criterion is less than or equal to the criterion threshold, then it is filtered into the body size data candidate set; The threshold value is the minimum value of the Akaike Information Criterion in the current body size dataset plus 2.
5. The method for predicting local pig slaughter performance as described in claim 1, characterized in that, In step S3, the slaughter performance includes carcass weight, lean meat weight, and fat weight.
6. The method for predicting local pig slaughter performance as described in claim 1, characterized in that, In step S3, the first error value includes the root mean square error and the mean absolute error of the first slaughter performance prediction value; wherein, the root mean square error of the first slaughter performance prediction value is calculated using equation (1); and the mean absolute error of the first slaughter performance prediction value is calculated using equation (2). （1）； （2）； in, Indicates the first Predicted first slaughter performance of a live pig; Indicates the first Actual slaughter performance of a live pig; This represents the total number of local pigs collected. The value range is 1~ Any natural number in the range; RMSE This represents the root mean square error of the first predicted slaughter performance value, and its corresponding error threshold range is 5% to 15% of the arithmetic mean of the actual slaughter performance values of all live pigs. MAE This represents the mean absolute error of the first predicted slaughter performance value, with the corresponding error threshold range being 3% to 10% of the arithmetic mean of the actual slaughter performance values of all live pigs. The coefficient of determination is calculated using equation (3): （3）； in, This represents the arithmetic mean of the actual slaughter performance values of all locally sourced pigs collected. R 2 The coefficient of determination represents the first predicted slaughter performance value relative to the actual slaughter performance value, and its value is greater than or equal to 0.
7. In the first predicted slaughter performance value RMSE and MAE All are within their respective error thresholds, and the determination coefficient of the first predicted slaughter performance value relative to the actual slaughter performance value is... R 2 If the value is greater than or equal to 0.7, then the body size data corresponding to the first predicted slaughter performance value is determined as the key variable.
7. The method for predicting local pig slaughter performance as described in claim 6, characterized in that, In step S4, the second error value is the comprehensive error value of the second slaughter performance prediction value, which is obtained by weighted summation of the normalized root mean square error value and the mean absolute percentage error value; wherein, the normalized root mean square error value of the second slaughter performance prediction value is calculated by combining equations (4) and (5); the mean absolute percentage error value of the second slaughter performance prediction value is calculated by equation (6); （4）； （5）； （6）； in, This represents the second predicted slaughter performance value; This represents the normalized root mean square error of the second slaughter performance prediction value; This represents the root mean square error of the second predicted slaughter performance value; This represents the mean absolute percentage error of the second slaughter performance prediction value.
8. The method for predicting local pig slaughter performance as described in claim 7, characterized in that, In step S4, the third error value is the comprehensive error value of the third slaughter performance prediction value, which is obtained by weighted summation of the normalized root mean square error value and the mean absolute percentage error value; wherein, the normalized root mean square error value of the third slaughter performance prediction value is calculated by combining equations (7) and (8); the mean absolute percentage error value of the third slaughter performance prediction value is calculated by equation (9); （7）； （8）； （9）； in, This represents the third predicted slaughter performance value; This represents the normalized root mean square error of the third predicted slaughter performance value. This represents the root mean square error of the third predicted slaughter performance value; This represents the mean absolute percentage error of the third slaughter performance prediction value.
9. A storage medium, characterized in that, It stores computer program instructions that, when executed by a processor, implement the method for predicting local pig slaughter performance as described in any one of claims 1 to 8.
10. An electronic device, characterized in that, include: At least one processor, at least one memory, and computer program instructions stored in the memory, which, when executed by the processor, implement the method for predicting local pig slaughter performance as described in any one of claims 1 to 8.