Nghiên cứu khả năng sinh trưởng, sinh sản, năng suất và chất lượng sữa của bò cái holstein friesian HF thuần, các thế hệ lai F1, F2 và F3 giữa HF và lai sind nuôi tại tỉnh Lâm Đồng

Bảng 5. Hệ số tương quan giữa năng suất sữa thực tế với chất lượng sữa

NSSTT của nhóm bò

VCKKM	Mỡ	Protein	Tỷ trọng
F1	- 0,70	- 0,60	- 0,70	- 0,09
F2	- 0,35	- 0,62	- 0,50	- 0,33
F3	- 0,77	- 0,46	- 0,29	- 0,84
HF	- 0,33	- 0,91	- 0,70	- 0,84

Có thể bạn quan tâm!

Trần Quang Hân (1996), Nghiên Cứu Các Tính Trạng Năng Suất Chủ Yếu Của Lợn Trắng Phú Khánh Và Lợn Lai F1 Yorkshire Trắng Phú Khánh, Luận Án Phó Tiến Sỹ Khoa
Djiko Soetrisno And Mahyuddin M. D. (1994), “Effect Of Graizing Systems On Reproductive Performance And The Net Return Of Milk Yielded By Shiwal – Friesian (Fs) Cows”, Proceeding Of The 7Th
Safi Jahanshahi A., Vaez Torshizi R., Kashan N. E. J. And Sayyad Nejad M. B. (2002), “Genetic Parameters For Milk Production Traits Of Iran Holsteins”, Proc. 7Th World Congress On Genetics
Nghiên cứu khả năng sinh trưởng, sinh sản, năng suất và chất lượng sữa của bò cái holstein friesian HF thuần, các thế hệ lai F1, F2 và F3 giữa HF và lai sind nuôi tại tỉnh Lâm Đồng - 23

Xem toàn bộ 186 trang tài liệu này.

4 KẾT QUẢ CHẠY HÀM GOMPERTZ TRÊN STATGRAPHICS CENTURION XV v 15.1.02

4.1 Kết quả chạy hàm Gompert của nhóm bò theo dõi

Nonlinear Regression - KLF1TD

Dependent variable: KLF1TD Independent variables: TTF1TD

Function to be estimated: M*EXP(-A*EXP(-B*TTF1TD)) Estimation method: Marquardt

Estimation stopped due to convergence of residual sum of squares. Number of iterations: 6

Number of function calls: 28

Estimation Results

		Asymptotic	95.0%
		Asymptotic	Confidence	Interval
Parameter	Estimate	Standard Error	Lower	Upper
M	420.804	4.89388	411.722	430.067
A	2.37423	0.0233495	2.32837	2.4201
B	0.104943	0.00209512	0.100828	0.109058

Analysis of Variance

Source

Sum of Squares	Df	Mean Square
Model	3.06549E7	3	1.02183E7
Residual	113987.	567	201.035
Total	3.07689E7	570
Total (Corr.)	7.47363E6	569

R-Squared = 98.3042 percent

R-Squared (adjusted for d.f.) = 98.2344 percent Standard Error of Est. = 14.1787

Mean absolute error = 11.6272 Durbin-Watson statistic = 0.836953

Lag 1 residual autocorrelation = 0.580415

Residual Analysis

Estimation	Validation
n	570
MSE	201.035
MAE	11.6272
MAPE	11.2176
ME	-0.650562
MPE	-5.75787

The StatAdvisor

The output shows the results of fitting a nonlinear regression model to describe the relationship between KLF1DT and 1 independent variables. The equation of the fitted model is

KLF1TD = 420.804*EXP(-2.37423*EXP(-0.104943*TTF1TD))

In performing the fit, the estimation process terminated successully after 6 iterations, at which point the estimated coefficients appeared to converge to the current estimates.

The R-Squared statistic indicates that the model as fitted explains 98.3042% of the variability in KLF1TD. The adjusted

R-Squared statistic, which is more suitable for comparing models with different numbers of independent variables, is 98.2344%. The standard error of the estimate shows the standard deviation of the residuals to be 14.1787. This value can be used to construct prediction limits for new observations by selecting the Forecasts option from the text menu. The mean absolute error (MAE) of 11.6272 is the average value of the residuals. The Durbin-Watson (DW) statistic tests the residuals to determine if there is any significant correlation based on the order in which they occur in your data file.

The output also shows aymptotic 95.0% confidence intervals for each of the unknown parameters. These intervals are approximate and most accurate for large sample sizes. You can determine whether or not an estimate is statistically significant by examining each interval to see whether it contains the value 0. Intervals covering 0 correspond to coefficients which may well be removed form the model without hurting the fit substantially.

Plot of Fitted Model

500

450

K h o i l u o n g b o F 1 ( k g )

400

350

300

250

200

150

100

0 6 12 18 24

Thang tuoi

Nonlinear Regression - KLF2TD

Dependent variable: KLF2TD Independent variables: TTF2TD

Function to be estimated: M*EXP(-A*EXP(-B*TTF2TD)) Estimation method: Marquardt

Estimation stopped due to convergence of residual sum of squares. Number of iterations: 6

Number of function calls: 28

Estimation Results

		Asymptotic	95.0%
		Asymptotic	Confidence	Interval
Parameter	Estimate	Standard Error	Lower	Upper
M	441.949	5.15766	431.999	452.258
A	2.35978	0.0240481	2.31255	2.40701
B	0.104381	0.00218541	0.100088	0.108673

Analysis of Variance

Source

Sum of Squares	Df	Mean Square
Model	3.49492E7	3	1.16497E7
Residual	146839.	587	250.151
Total	3.5096E7	590
Total (Corr.)	8.50234E6	589

R-Squared = 98.463 percent

R-Squared (adjusted for d.f.) = 98.271 percent Standard Error of Est. = 15.8162

Mean absolute error = 12.6605 Durbin-Watson statistic = 0.586476

Lag 1 residual autocorrelation = 0.705686

Residual Analysis

Estimation	Validation
n	590
MSE	250.151
MAE	12.6605
MAPE	11.8795
ME	-0.730073
MPE	-6.17391

The StatAdvisor

The output shows the results of fitting a nonlinear regression model to describe the relationship between KLF2TD and 1 independent variables. The equation of the fitted model is

KLF2TD = 441.949*EXP(-2.35978*EXP(-0.104381*TTF2TD))

In performing the fit, the estimation process terminated successully after 6 iterations, at which point the estimated coefficients appeared to converge to the current estimates.

The R-Squared statistic indicates that the model as fitted explains 98.463% of the variability in KLF2TD. The adjusted R-Squared statistic, which is more suitable for comparing models with different numbers of independent variables, is 98.271%. The standard error of the estimate shows the standard deviation of the residuals to be 15.8162. This value can be used to construct prediction limits for new observations by selecting the Forecasts option from the text menu. The mean absolute error (MAE) of 12.6605 is the average value of the residuals. The Durbin-Watson (DW) statistic tests the residuals to determine if there is any significant correlation based on the order in which they occur in your data file.

Plot of Fitted Model

500

K h o i lu o n g b o F 2 ( k g )

450

400

350

300

250

200

150

100

0 6 12 18 24

Thang tuoi

Nonlinear Regression - KLF3TD

Dependent variable: KLF3TD Independent variables: TTF3TD

Function to be estimated: M*EXP(-A*EXP(-B*TTF3TD)) Estimation method: Marquardt

Estimation stopped due to convergence of parameter estimates. Number of iterations: 7

Number of function calls: 31

Estimation Results

		Asymptotic	95.0%
		Asymptotic	Confidence	Interval
Parameter	Estimate	Standard Error	Lower	Upper
M	478.554	4.61922	469.703	487.848
A	2.36299	0.020509	2.32271	2.40328
B	0.105528	0.00184746	0.101899	0.109156

Analysis of Variance

Source

Sum of Squares	Df	Mean Square
Model	4.07927E7	3	1.35976E7
Residual	119364.	577	206.87
Total	4.09121E7	580
Total (Corr.)	9.86366E6	579

R-Squared = 98.7329 percent

R-Squared (adjusted for d.f.) = 98.7237 percent Standard Error of Est. = 14.383

Mean absolute error = 11.8348 Durbin-Watson statistic = 0.90785

Lag 1 residual autocorrelation = 0.54493

Residual Analysis

Estimation	Validation
n	580
MSE	206.87
MAE	11.8348
MAPE	10.7227
ME	-0.748078
MPE	-5.66541

The StatAdvisor

The output shows the results of fitting a nonlinear regression model to describe the relationship between KLF3TD and 1 independent variables. The equation of the fitted model is

KLF3TD = 478.554*EXP(-2.36299*EXP(-0.105528*TTF3TD))

In performing the fit, the estimation process terminated successully after 7 iterations, at which point the residual sum of squares appeared to approach a minimum.

The R-Squared statistic indicates that the model as fitted explains 98.7329% of the variability in KLF3TD. The adjusted R-Squared statistic, which is more suitable for comparing models with different numbers of independent variables, is 98.7237%. The standard error of the estimate shows the standard deviation of the residuals to be 14.383. This value can be used to construct prediction limits for new observations by selecting the Forecasts option from the text menu. The mean absolute error (MAE) of 11.8348 is the average value of the residuals. The Durbin-Watson (DW) statistic tests the residuals to determine if there is any significant correlation based on the order in which they occur in your data file.

Plot of Fitted Model

500

450

K h o i l u o n g F 3 ( k g )

400

350

300

250

200

150

100

0 6 12 18 24

Thang tuoi

Nonlinear Regression - KLHFTD

Dependent variable: KLHFTD Independent variables: TTHFTD

Function to be estimated: M*EXP(-A*EXP(-B*TTHFTD)) Estimation method: Marquardt

Estimation stopped due to convergence of parameter estimates. Number of iterations: 7

Number of function calls: 31

Estimation Results

		Asymptotic	95.0%
		Asymptotic	Confidence	Interval
Parameter	Estimate	Standard Error	Lower	Upper
M	498.823	3.38154	492.612	505.867
A	2.36657	0.0151509	2.33188	2.39127
B	0.107524	0.00135165	0.104875	0.110173

Analysis of Variance

Source

Sum of Squares	Df	Mean Square
Model	1.00592E8	3	3.35307E7
Residual	344281.	1282	268.55
Total	1.00936E8	1285
Total (Corr.)	2.43181E7	1284

R-Squared = 98.5843 percent

R-Squared (adjusted for d.f.) = 98.5821 percent Standard Error of Est. = 16.3875

Mean absolute error = 13.4649 Durbin-Watson statistic = 0.883896

Lag 1 residual autocorrelation = 0.557709

Residual Analysis

Estimation	Validation
n	1285
MSE	268.55
MAE	13.4649
MAPE	11.0368
ME	-0.788923
MPE	-5.67459

The StatAdvisor

The output shows the results of fitting a nonlinear regression model to describe the relationship between KLHFTD and 1 independent variables. The equation of the fitted model is

KLHFTD = 498.823*EXP(-2.36657*EXP(-0.107524*TTHFTD))

In performing the fit, the estimation process terminated successully after 7 iterations, at which point the residual sum of squares appeared to approach a minimum.

The R-Squared statistic indicates that the model as fitted explains 98.5843% of the variability in KLHFTD. The adjusted R-Squared statistic, which is more suitable for comparing models with different numbers of independent variables, is 98.5821%. The standard error of the estimate shows the standard deviation of the residuals to be 16.3875. This value can be used to construct prediction limits for new observations by selecting the Forecasts option from the text menu. The mean absolute error (MAE) of 13.4649 is the average value of the residuals. The Durbin-Watson (DW) statistic tests the residuals to determine if there is any significant correlation based on the order in which they occur in your data file.

Plot of Fitted Model

500

450

K h o i l u o n g b o H F ( k g )

400

350

300

250

200

150

100

0 6 12 18 24

Thang tuoi

4.2 Kết quả chạy hàm Gompert của nhóm bò nuôi thí nghiệm

Nonlinear Regression - KLF1NTN

Dependent variable: KLF1NTN Independent variables: TTNTN

Function to be estimated: M*EXP(-A*EXP(-B*TTNTN)) Estimation method: Marquardt

Estimation stopped due to convergence of parameter estimates. Number of iterations: 7

Number of function calls: 31

Estimation Results

		Asymptotic	95.0%
		Asymptotic	Confidence	Interval
Parameter	Estimate	Standard Error	Lower	Upper
M	444.484	7.62082	429.544	459.834
A	2.30287	0.0347721	2.23385	2.37188
B	0.104687	0.00327434	0.0981884	0.111186

Analysis of Variance

Source

Sum of Squares	Df	Mean Square
Model	6.13112E6	3	2.04371E6
Residual	9480.44	97	97.7365
Total	6.1406E6	100
Total (Corr.)	1.43782E6	99

R-Squared = 99.3406 percent

R-Squared (adjusted for d.f.) = 99.327 percent Standard Error of Est. = 9.88618

Mean absolute error = 8.64538 Durbin-Watson statistic = 0.352689

Lag 1 residual autocorrelation = 0.817525

Residual Analysis

Estimation	Validation
n	100
MSE	97.7365
MAE	8.64538
MAPE	9.70995
ME	-0.700724
MPE	-5.2761

The StatAdvisor

The output shows the results of fitting a nonlinear regression model to describe the relationship between KLF1NTN and 1 independent variables. The equation of the fitted model is

KLF1NTN = 444.484*EXP(-2.30287*EXP(-0.104687*TTNTN))

In performing the fit, the estimation process terminated successully after 7 iterations, at which point the residual sum of squares appeared to approach a minimum.

The R-Squared statistic indicates that the model as fitted explains 99.3406% of the variability in KLF1NTN. The adjusted R-Squared statistic, which is more suitable for comparing models with different numbers of independent variables, is 99.327%. The standard error of the estimate shows the standard deviation of the residuals to be 9.88618. This value can be used to construct prediction limits for new observations by selecting the Forecasts option from the text menu. The mean absolute error (MAE) of 8.64538 is the average value of the residuals. The Durbin-Watson (DW) statistic tests the residuals to determine if there is any significant correlation based on the order in which they occur in your data file.

Plot of Fitted Model

500

450

K h o i l u o n g b o F 1 ( k g )

400

350

300

250

200

150

100

0 6 12 18 24

Thang tuoi

Nonlinear Regression - KLF2NTN

Dependent variable: KLF2NTN Independent variables: TTNTD

Function to be estimated: M*EXP(-A*EXP(-B*TTNTN)) Estimation method: Marquardt

Estimation stopped due to convergence of parameter estimates. Number of iterations: 7

Number of function calls: 31

Estimation Results

		Asymptotic	95.0%
		Asymptotic	Confidence	Interval
Parameter	Estimate	Standard Error	Lower	Upper
M	468.184	8.03365	452.489	484.422
A	2.37464	0.0410211	2.29322	2.45605
B	0.107303	0.00358498	0.102788	0.117018

Analysis of Variance

Source

Sum of Squares	Df	Mean Square
Model	7.04729E6	3	2.3491E6
Residual	12596.6	97	129.862
Total	7.05989E6	100
Total (Corr.)	1.69594E6	99

R-Squared = 99.2372 percent

R-Squared (adjusted for d.f.) = 99.2219 percent Standard Error of Est. = 11.3957

Mean absolute error = 9.7617