Chicken Coop Production Prediction Method Based on Optimal Delay Detrending and Stepwise Regression

Abstract

The present invention relates to a chicken coop production prediction method based on optimal delay detrending and stepwise regression, belonging to the technical field of production information prediction in the breeding industry. The present invention includes the following steps: S1: Establishment of a chicken coop production prediction model: The establishment process is divided into the following two dimensions: S11: Production data time series, S12: Environmental data time series; S2: Preprocessing of the chicken coop production prediction model: S21: DFA detrending processing of the production data time series, S22: SCCF analysis of the optimal delay of the environmental data time series; S3: Establishment of a stepwise regression equation for environmental factors, S4: Fitting of the regression curve of egg production fluctuations, S5: Prediction of future chicken coop production information. The present invention widely combines the technologies of predicting future production by integrating chicken coop environment and production information, can accurately predict the production information of chickens, greatly improves the quality and quantity of breeding products, and increases the income of farmers.

Description

Technical Field

[0001] The present invention relates to a chicken coop production prediction method based on optimal delay detrending and stepwise regression, belonging to the technical field of aquaculture production information prediction. Background Art

[0002] The current aquaculture technology is a combination of traditional and modern aquaculture technologies, and the proportion of traditional aquaculture technology is relatively large. Due to the low awareness of scientific aquaculture in some areas and the immature mastery of aquaculture technology, the growth environment of chicken coops is poor, resulting in a significant reduction in the quality and quantity of aquaculture products and affecting the final income. At present, the technology of predicting future production by combining chicken coop environment and production information is not widely applied. Therefore, it is very meaningful to invent a method that can accurately predict the production information of chickens. Summary of the Invention

[0003] In view of the above-mentioned defects existing in the prior art, the present invention proposes a chicken coop production prediction method based on optimal delay detrending and stepwise regression.

[0004] The chicken coop production prediction method based on optimal delay detrending and stepwise regression according to the present invention includes the following steps:

[0005] S1: Establishment of a chicken coop production prediction model: The establishment process is divided into the following two dimensions:

[0006] S11: Obtain the daily egg production data of the chicken coop within the historical time period, form a production data time series, and store it in a computer;

[0007] S12: Obtain the temperature, humidity, carbon dioxide concentration, negative pressure, and wind speed data within the historical time period, form an environmental data time series, and store it in a computer;

[0008] S2: Preprocessing of the chicken coop production prediction model: Eliminate abnormal data in the chicken coop production prediction model through data preprocessing technology, where: the data preprocessing technology is divided into the following small steps:

[0009] S21: DFA detrending processing of the production data time series, including the following small steps:

[0010] Establish a new sequence y(i) of egg production in the chicken coop:

[0011]

[0012] In the formula: <x>is the mean of the production sequence <>, of the original chicken coop; x k is the egg production of the chicken coop at time k; N is the length of the interval of the chicken coop production sequence;

[0013] Divide the new sequence y(i) of the egg production of the chicken coop with an interval length of N into non-overlapping sub-intervals of equal length S;

[0014] Perform polynomial regression fitting on the data of each sub-interval to obtain the local trend function y v (i) of the chicken coop production information, to eliminate the trend within each sub-interval of the time series of the chicken coop production data, and the formula for calculating its variance mean is as follows:

[0015]

[0016]

[0017] In the formula: v is the label of different data segments; s is the length of the new data sequence; y is the egg production sequence of the chicken coop data; Ns is the number of intervals after sequence reconstruction;

[0018] Determine the q-order fluctuation function of the entire sequence, and calculate the detrended egg production fluctuation sequence F 0 (s):

[0019]

[0020] In the formula: exp{} is the exponential calculation formula with e as the base; 2Ns is twice the number of intervals; 1 / q is the reciprocal of the order of the fluctuation function in DFA;

[0021] S22: The optimal delay of the SCCF analysis of the environmental data time series, including the following small steps:

[0022] Obtain the time series of the production data after DFA detrending processing, and analyze its correlation with the local environmental data time series, including the following situations:

[0023] Situation 1: Through the sliding analysis of the data sequence window, when the correlation between the environmental data time series and the production data time series is less than a certain threshold, continue to slide and search;

[0024] Situation 2: Through the sliding analysis of the data sequence window, when the correlation between the environmental data time series and the production data time series is greater than a certain threshold, this data sequence window is the optimal time delay to be eliminated;

[0025] Eliminate the optimal time delay, which is the environmental data time series without delay information;

[0026] S3: Establishment of stepwise regression equation for environmental factors: Perform non-linear normalization transformation on the environmental data after removing the time delay, convert it into environmental factors that are easy to analyze, and finally establish a stepwise regression equation between the environmental factors and the production data series;

[0027] S4: Fitting of the regression curve of egg production fluctuations: Substitute the processed environmental variables and production variables into the stepwise regression equation in order of significance, and fit the regression curve of egg production fluctuations;

[0028] S5: Prediction of future chicken house production information: Add this fluctuation value to the stable egg production value of the chicken house, and a prediction model of the chicken house production data can be obtained, realizing the prediction of future chicken house production information using environmental data.

[0029] Preferably, in the step S1, the data acquisition system is installed in the chicken house. The data acquisition system detects temperature and humidity through a temperature and humidity sensor, detects carbon dioxide concentration through a carbon dioxide sensor, detects pressure data through a pressure gauge, detects wind speed data through an anemometer, and stores them in a computer in the cloud through a wireless communication module.

[0030] Preferably, in the preprocessing of the chicken house production prediction model in step S2, the following pre-steps are further included:

[0031] Obtain abnormal data within the historical time period: Use data processing technology to process data outliers and store them in the computer for the next data processing.

[0032] Preferably, in the step S21, a sequence of length N needs to be divided into N s = N / S subintervals. Since the sequence length N may not be divisible by s, in order to ensure that the original sequence information is not lost, the same operation is performed on the reverse order of the sequence to obtain 2N subintervals.

[0033] Preferably, in the step S21, the local trend function y v (i) of the chicken house production information is a polynomial of the first order, second order or higher order.

[0034] Preferably, in the step S21, the DFA detrending process uses the detrended fluctuation analysis method. By analyzing the fluctuation curve of the egg production in the chicken house, some growth factors brought about by the growth and development of chickens are effectively discriminated for the egg production trend of chickens. After removing the trend information, the remaining fluctuations are caused by the environment, and these fluctuations can better describe the environmental factors.

[0035] Preferably, in the step S22, the optimal delay of the SCCF analysis uses the delay cross-correlation analysis method to calculate the optimal delay correlation between the environmental data and the production data, calculate the optimal delay and remove it.

[0036] Preferably, in the step S3, the egg production fluctuation regression model F(x) is as follows:

[0037]

[0038] In the formula: X represents the production variable, that is, X = {xi|i∈V}, V is the processed environmental variable, g(X) is the variable combination function, and β 0 , β 1 , β 2 , β 3 are the parameters to be regressed.

[0039] The chicken coop production prediction system based on optimal delay detrending and stepwise regression according to the present invention includes a memory and a processor connected to each other. Among them, the memory is used to store a computer program, and when the computer program is executed by the processor, it is used to implement the chicken coop production prediction method based on optimal delay detrending and stepwise regression.

[0040] The computer-readable storage medium according to the present invention is used to store a computer program, and when the computer program is executed by a processor, it is used to implement the chicken coop production prediction method based on optimal delay detrending and stepwise regression.

[0041] The beneficial effects of the present invention are as follows: The present invention widely combines the technologies of chicken coop environment and production information to predict future production, can accurately predict the production information of chickens, greatly improves the quality and quantity of breeding products, and increases the income of farmers. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 is the chicken coop production prediction flowchart of the present invention.

[0043] Figure 2 is the structure diagram of the data processing part of the present invention.

[0044] Figure 3 is the analysis flowchart of the delay cross-correlation analysis method of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0045] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention.

[0046] The purpose of the present invention is to provide an efficient and accurate chicken house production prediction method, which is used to predict the future production of the chicken house and understand the future production of the chicken. The temperature, humidity, negative pressure, carbon dioxide concentration, and wind speed information inside the chicken house are collected by a data acquisition system as an environmental data sequence, and the daily egg production of the chicken is counted as a production data sequence. The improved DFA algorithm and the delayed cross-correlation analysis method are used to eliminate the trend and delay respectively. Finally, the optimal regression equation is established by combining environmental factors and production factors through stepwise regression, and the production regression curve is fitted to complete the prediction of future egg production.

[0047] like Figure 1 The chicken house production forecasting flow chart shown briefly describes the steps of the entire forecasting model.

[0048] First, the environmental data inside the chicken house is collected through the data acquisition system inside the chicken house, including: temperature, humidity, carbon dioxide concentration, negative pressure, wind speed, and the chicken house environmental data time series is stored in the computer; the daily egg production of the chickens is counted and stored in the computer as the chicken house production data time series.

[0049] Secondly, data preprocessing technology is used to eliminate abnormal data in environmental data sequences and production data sequences.

[0050] Then, the trend information of increased egg production due to the growth of laying hens in the production sequence is removed by using the improved detrending fluctuation analysis method (DFA), and the optimal delay between the environmental data time series and the production data time series is analyzed and eliminated by the delayed cross-correlation analysis method.

[0051] Based on the currently known optimal environmental values, the optimal chicken house environmental factors are calculated, and the optimal regression equation is established using the stepwise regression method. The optimal chicken house environmental factors are substituted into the regression equation according to the significance, and the chicken house production curve is fitted. The chicken house production information is predicted based on this curve.

[0052] like Figure 2 The data processing part structure diagram shown in the figure adopts the improved trend elimination fluctuation analysis method to eliminate the trend information of the production time series, that is, to eliminate the fluctuation of the increase in egg production caused by the growth of chickens, so that the production data series only has fluctuations caused by the environment. The DFA method is a method based on the theory of random processes, which is often used to analyze the trend information in time series. From a dynamic point of view, the transformed sequence in this method still has traces of the original sequence and maintains the same persistence as the original sequence. After improvement, the DFA algorithm can better filter out the trend component of its own growth, and the remaining deviation sequence is mainly the egg production fluctuation component caused by the chicken house environment. The use of the DFA method can avoid the misjudgment of the sequence environment fluctuation information. The implementation steps of the improved DFA are as follows:

[0053] (1) Establish a new sequence y(i) of the egg production in the chicken coop:

[0054]

[0055] Where: <x>is the mean of the original chicken coop production sequence <x k >.

[0056] (2) Divide the new egg production sequence y(i) of the chicken coop into non - overlapping equal - length sub - intervals of length S. A sequence of length N needs to be divided into N s = N / S sub - intervals. Since the sequence length N may not be divisible by S, in order to ensure that the information of the original sequence is not lost, the same operation is performed on the reverse order of the sequence. In this way, a total of 2N sub - intervals can be obtained.

[0057] (3) Perform polynomial regression fitting on the data of each sub - interval to obtain the local trend function y v (i) of the chicken coop production information, which can be a first - order, second - order or higher - order polynomial. Eliminate the trend within each sub - interval of the chicken coop production sequence, and calculate the formula for its mean variance as follows:

[0058]

[0059]

[0060] (4) Determine the q - order fluctuation function of the entire sequence, and calculate the detrended egg production fluctuation sequence:

[0061]

[0062] As Figure 3 shown in the delayed cross - correlation analysis method delay analysis flow chart. The specific method is to first read the local environmental sequence and production sequence information, remove the abnormal data values in each time series through a certain method, analyze the correlation between the environmental sequence and the production sequence, and perform a sliding window analysis on the data sequence. When the correlation between the environmental data and the production data is greater than a certain threshold, usually 0.5, we consider that there is a high correlation between the environmental data and the production data at this time. The duration of the sliding of this data sequence window is the optimal time delay to be removed. After removing it, the environmental data without delay information can be obtained, and the detrended production data sequence and the environmental sequence without delay information can be obtained. Perform a non - linear normalization transformation on the environmental data without delay, convert it into environmental factors that are easy to analyze, and finally establish a step - by - step regression equation between the environmental factors and the production data sequence. Substitute the processed environmental variables and production variables into the step - by - step regression equation in order of significance, fit the regression curve of the egg production fluctuation, and finally add the fluctuation value to the stable egg production value of the chicken coop to obtain the prediction model of the chicken coop production data, realizing the prediction of future chicken coop production information using environmental data.

[0063] In summary, the method can comprehensively integrate the production data and environmental data sequences of the chicken coop, and accurately predict the future production information of the chicken coop, which has great practical value for evaluating the feeding environment inside the chicken coop and improving production quality and quantity. The above specific implementation manners have described this invention in detail and are not limited to the above manners. For those of ordinary skill in the art, improvements, substitutions, uses, etc. of this method based on the above invention principles are within the protection scope of this invention.

[0064] Example 1:

[0065] This example gives the detailed steps of the prediction method of the present invention:

[0066] S1: Establishment of the chicken coop production prediction model: The establishment process is divided into the following two dimensions:

[0067] S11: Obtain the daily egg production data of the chicken coop within the historical time period, form the production data time series, and store it in the computer;

[0068] S12: Obtain the temperature, humidity, carbon dioxide concentration, negative pressure, and wind speed data within the historical time period, form the environmental data time series, and store it in the computer;

[0069] S2: Preprocessing of the chicken coop production prediction model: Eliminate the abnormal data in the chicken coop production prediction model through data preprocessing technology, where: the data preprocessing technology is divided into the following small steps:

[0070] S21: DFA detrending processing of the production data time series, including the following small steps:

[0071] Establish a new sequence y(i) of the egg production of the chicken coop:

[0072]

[0073] In the formula: <x>is the mean of the production sequence of the original chicken coop; x k is the egg production of the chicken coop at time k; N is the interval length of the chicken coop production sequence;

[0074] Divide the new sequence y(i) of the egg production of the chicken coop with an interval length of N into non-overlapping sub-intervals of equal length S;

[0075] Perform polynomial regression fitting on the data of each sub-interval to obtain the local trend function y v (i) of the chicken coop production information to eliminate the trend within each sub-interval of the time series of the chicken coop production data. The formula for calculating its variance mean is as follows:

[0076]

[0077]

[0078] In the formula: v is the label of different data segments; s is the length of the new data sequence; y is the egg production sequence of the chicken coop data; Ns is the number of intervals after sequence reconstruction;

[0079] Determine the q-order fluctuation function of the entire sequence, and calculate the detrended egg production fluctuation sequence F 0 (s):

[0080]

[0081] In the formula: exp{} is the exponential calculation formula with e as the base; 2Ns is twice the number of intervals; 1 / q is the reciprocal of the order of the fluctuation function in DFA;

[0082] S22: The optimal delay of the SCCF analysis of the environmental data time series, including the following steps:

[0083] Obtain the time series of the production data after DFA detrending processing, and analyze its correlation with the local environmental data time series, including the following situations:

[0084] Situation 1: Through the sliding analysis of the data sequence window, when the correlation between the environmental data time series and the production data time series is less than a certain threshold, continue to slide and search;

[0085] Situation 2: Through the sliding analysis of the data sequence window, when the correlation between the environmental data time series and the production data time series is greater than a certain threshold, this data sequence window is the optimal time delay to be eliminated;

[0086] Eliminate the optimal time delay, which is the environmental data time series without delay information;

[0087] S3: Establishment of stepwise regression equation for environmental factors: Perform non-linear normalization transformation on the environmental data after removing the time delay, convert it into environmental factors that are easy to analyze, and finally establish a stepwise regression equation between the environmental factors and the production data series;

[0088] S4: Fitting of the regression curve of egg production fluctuation: Substitute the processed environmental variables and production variables into the stepwise regression equation in order of significance, and fit the regression curve of egg production fluctuation;

[0089] S5: Prediction of future chicken house production information: Add this fluctuation value to the stable egg production value of the chicken house, and a prediction model of the chicken house production data can be obtained, realizing the prediction of future chicken house production information using environmental data.

[0090] Embodiment 2:

[0091] This embodiment gives the relevant descriptions of the prediction system and computer storage medium of the present invention:

[0092] The chicken house production prediction system based on optimal time-delay detrending and stepwise regression according to the present invention includes a memory and a processor connected to each other. Among them, the memory is used to store a computer program, and when the computer program is executed by the processor, it is used to implement the chicken house production prediction method based on optimal time-delay detrending and stepwise regression.

[0093] The computer-readable storage medium according to the present invention is used to store a computer program, and when the computer program is executed by a processor, it is used to implement the chicken house production prediction method based on optimal time-delay detrending and stepwise regression.

[0094] The beneficial effects of the present invention are as follows: The present invention widely combines the technologies of predicting future production by chicken house environment and production information, can accurately predict the production information of chickens, greatly improve the quality and quantity of breeding products, and increase the income of farmers.

[0095] The present invention can be widely applied to the occasion of predicting production information in the breeding industry.< / x> < / x> < / x>

Claims

1. A chicken coop production prediction method based on optimal delay detrending and stepwise regression, characterized in that, it includes the following steps: S1: Establishment of the chicken coop production prediction model: The establishment process is divided into the following two dimensions: S11: Obtain the daily egg production data of the chicken coop within the historical time period, form a production data time series, and store it in the computer; S12: Obtain the temperature, humidity, carbon dioxide concentration, negative pressure, and wind speed data within the historical time period, form an environmental data time series, and store it in the computer; S2: Preprocessing of the chicken coop production prediction model: Eliminate abnormal data in the chicken coop production prediction model through data preprocessing technology, where: the data preprocessing technology is divided into the following small steps: S21: DFA detrending processing of the production data time series, including the following small steps: Establish a new sequence of egg production in the chicken coop : （1） In the formula: <x>is the mean of the production sequence < > of the original chicken coop; x k is the egg production of the chicken coop at time k; N is the interval length of the production sequence of the chicken coop;< / x> For the new sequence of egg production in a chicken coop with an interval length of N Perform partitioning to obtain non-overlapping sub-intervals of equal length S; Perform polynomial regression fitting on the data of each sub-interval to obtain the local trend function of the chicken coop production information , to eliminate the trend within each sub-interval of the time series of chicken coop production data, the formula for calculating its variance mean is as follows: （2） （3） In the formula: v is the label of different data segments; s is the length of the new data sequence; y is the egg production sequence of the chicken coop data; Ns is the number of intervals after sequence reconstruction; Determine the q-order fluctuation function of the full sequence and calculate the detrended egg production fluctuation sequence : （4） In the formula: exp{} is the exponential calculation formula with e as the base; 2Ns is twice the number of intervals; 1 / q is the reciprocal of the order of the fluctuation function in DFA; S22: SCCF analysis of the optimal delay of the environmental data time series, including the following small steps: Obtain the production data time series after DFA detrending processing, and analyze its correlation with the local environmental data time series, including the following situations: Situation 1: Through data sequence window sliding analysis, when the correlation between the environmental data time series and the production data time series is less than a certain threshold, continue to slide and search; Situation 2: Through data sequence window sliding analysis, when the correlation between the environmental data time series and the production data time series is greater than a certain threshold, this data sequence window is the optimal time delay to be eliminated; Eliminating the optimal time delay is the environmental data time series without delay information; S3: Establishment of the stepwise regression equation of environmental factors: Perform non-linear normalization transformation on the environmental data without delay, convert it into environmental factors that are easy to analyze, and finally establish a stepwise regression equation between the environmental factors and the production data sequence; S4: Fitting of the egg production fluctuation regression curve: Substitute the processed environmental variables and production variables into the stepwise regression equation in order of significance, and fit the egg production fluctuation regression curve; S5: Prediction of future chicken coop production information: Add this fluctuation value to the stable egg production value of the chicken coop, and a prediction model of the chicken coop production data can be obtained, realizing the prediction of future chicken coop production information using environmental data.

2. The chicken coop production prediction method based on optimal delay detrending and stepwise regression according to claim 1, characterized in that, in step S1, the data acquisition system is installed in the chicken coop. The data acquisition system detects temperature and humidity through a temperature and humidity sensor, detects carbon dioxide concentration through a carbon dioxide sensor, detects pressure data through a pressure gauge, detects wind speed data through an anemometer, and stores it in a computer in the cloud through a wireless communication module.

3. The chicken coop production prediction method based on optimal delay detrending and stepwise regression according to claim 1, characterized in that, In the preprocessing of the chicken coop production prediction model in step S2, the following preliminary steps are further included: Obtain abnormal data within the historical time period: Use data processing techniques to process data outliers and store them in a computer for the next data processing.

4. The chicken coop production prediction method based on optimal delay detrending and stepwise regression according to claim 3, characterized in that, In the step S21, the sequence with length N needs to be divided into = N / S subintervals. Since the sequence length N may not be divisible by s, in order to ensure that the original sequence information is not lost, the same operation is performed on the reverse order of the sequence to obtain 2N subintervals.

5. The chicken coop production prediction method based on optimal delay detrending and stepwise regression according to claim 3, characterized in that, In the step S21, the local trend function of the chicken coop production information is a polynomial of the first order, second order or higher order.

6. The chicken coop production prediction method based on optimal delay detrending and stepwise regression according to claim 3, characterized in that, In step S21, the DFA detrending process uses the detrended fluctuation analysis method. By analyzing the fluctuation curve of the egg production in the chicken coop, some growth factors brought about by the growth and development of the chickens are effectively discriminated for the egg production trend of the chickens. After removing the trend information, the remaining fluctuations are caused by the environment, and these fluctuations can better describe the environmental factors.

7. The chicken coop production prediction method based on optimal delay detrending and stepwise regression according to claim 1, characterized in that, In step S22, the optimal delay of the SCCF analysis uses the delay cross-correlation analysis method to calculate the optimal delay correlation between the environmental data and the production data, calculate the optimal delay and remove it.

8. The chicken coop production prediction method based on optimal delay detrending and stepwise regression according to claim 1, characterized in that, In step S3, the egg production fluctuation regression model F(x) is: （5） Where: X represents a production variable, i.e., X = {xi|i ∈ V}, V is the processed environmental variable, g(X) is a variable combination function, and β 0 , β 1 , β 2 , β 3 is the parameter to be regressed.

Images