Analysis of Variance

Source	Sum of Squares	Df	Mean Square	F-Ratio	P-Value
Model	69227,4		69227,4	296,66	0,0000
Residual	4200,39		233,355
Total (Corr.)	73427,8

Correlation Coefficient = -0,970977

R-squared = 94,2796 percent

R-squared (adjusted for d.f.) = 93,9618 percent

Standard Error of Est. = 15,276

Mean absolute error = 12,4017

Durbin-Watson statistic = 0,17058 (P=0,0000)

Lag 1 residual autocorrelation = 0,715

The StatAdvisor

The output shows the results of fitting a linear model to describe the relationship between В and N. The equation of the fitted model is

В = 216,032 - 10,203*N

Since the P-value in the ANOVA table is less than 0,05, there is a statistically significant relationship between В and N at the 95,0% confidence level.

The R-Squared statistic indicates that the model as fitted explains 94,2796% of the variability in В. The correlation coefficient equals -0,970977, indicating a relatively strong relationship between the variables. The standard error of the estimate shows the standard deviation of the residuals to be 15,276. 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 12,4017 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. Since the P-value is less than 0,05, there is an indication of possible serial correlation at the 95,0% confidence level. Plot the residuals versus row order to see if there is any pattern that can be seen.

Comparison of Alternative Models

Model	Correlation	R-Squared
Exponential	-0,9995	99,91%
Reciprocal-Y squared-X	0,9988	99,77%
Square root-X	-0,9965	99,30%
Squared-Y logarithmic-X	-0,9938	98,76%
Square root-Y	-0,9926	98,52%
Logarithmic-X	-0,9863	97,29%
Logarithmic-Y square root-X	-0,9841	96,84%
Logarithmic-Y squared-X	-0,9729	94,65%
Linear	-0,9710	94,28%
Reciprocal-Y	0,9677	93,64%
Squared-Y square root-X	-0,9608	92,32%
Square root-Y squared-X	-0,9361	87,63%
Multiplicative	-0,9284	86,20%
Squared-Y reciprocal-X	0,9085	82,54%
Squared-Y	-0,9010	81,18%
Squared-X	-0,8874	78,76%
Reciprocal-Y logarithmic-X	0,8269	68,38%
Reciprocal-X	0,8257	68,17%
Double squared	-0,7774	60,44%
Square root-Y reciprocal-X	0,7687	59,10%
S-curve model	0,7040	49,55%
Double reciprocal	-0,5699	32,48%
Double square root	<no fit>
Reciprocal-Y square root-X	<no fit>
Square root-Y logarithmic-X	<no fit>
Logistic	<no fit>
Log probit	<no fit>

The StatAdvisor

This table shows the results of fitting several curvilinear models to the data. Of the models fitted, the exponential model yields the highest R-Squared value with 99,9054%. This is 5,62584% higher than the currently selected linear model. To change models, select the Analysis Options dialog box.

График остатков:

Пересчитаем в экспонентациальную модель

Simple Regression - В vs. N

Dependent variable: В

Independent variable: N

Exponential model: Y = exp(a + b*X)

Дата добавления: 2015-09-06