HPA Clostridium difficile analysis

ARIMA
Notes
Output Created
Comments
Input
Missing Value
Handling
02-FEB-2007 10:19:14
Data
Filter
Weight
Split File
N of Rows in Working
Data File
Date
Definition of Missing
Cases Used
Syntax
Resources
Use
Predict
Elapsed Time
From
To
From
To
Time Series
Amount of Output
Settings (TSET) Saving New Variables
Maximum Number of
Lags in Autocorrelation or
Partial Autocorrelation
Plots
Maximum Number of
Lags in Cross-Correlation
Plots
Maximum Number of New
Variables Generated Per
Procedure
Maximum Number of New
Cases Per Procedure
Treatment of
User-Missing Values
Confidence Interval
Percentage Value
Tolerance for Entering
Variables in Regression
Equations
D:\My Documents\Cdiff\HPA Cdiff
data.sav
<none>
<none>
<none>
12
<none>
User-defined missing values are
treated as missing.
Statistics are based on all cases,
including cases with any missing
value.
ARIMA Cdiff
/MODEL=( 1 0 0 )CONSTANT
/MXITER= 10
/PAREPS= .001
/SSQPCT= .001
/FORECAST= EXACT .
0:00:00.25
First observation
Last observation
First observation following the use
period
Last observation
PRINT = DEFAULT
NEWVAR = ALL
MXAUTO = 16
MXCROSS = 7
MXNEWVAR = 60
MXPREDICT = 1000
MISSING = EXCLUDE
CIN = 95
TOLER = .0001
Maximum Iterative
Parameter Change
Method of Calculating
Std. Errors for
Autocorrelations
Length of Seasonal
Period
Variable Whose Values
Label Observations in
Plots
Equations Include
FIT_1
Variables
Created or
Modified
ERR_1
LCL_1
UCL_1
SEP_1
CNVERGE = .001
ACFSE = IND
Unspecified
Unspecified
CONSTANT
Fit for Cdiff from ARIMA, MOD_1,
CON
Error for Cdiff from ARIMA, MOD_1,
CON
95% LCL for Cdiff from ARIMA,
MOD_1, CON
95% UCL for Cdiff from ARIMA,
MOD_1, CON
SE of Fit for Cdiff from ARIMA,
MOD_1, CON
[DataSet0]
Model Description(a)
Model Name
Dependent Series
MOD_1
Transformation
Constant
AR
Non-Seasonal
Differencing
MA
None
Included
1
Cdiff
0
None
Applying the model specifications from MOD_1
a Since there is no seasonal component in the model, the seasonality of the data will be ignored.
Iteration Termination Criteria
Maximum Parameter
Change Less Than
Maximum Marquardt
Constant Greater Than
Sum of Squares
Percentage Change
Less Than
Number of Iterations
Equal to
.001
10000000
00
.001%
10
Case Processing Summary
Series Length
Number of Cases
Skipped Due to
Missing Values
11
At the Beginning of the
Series
At the End of the
Series
Number of Cases with Missing Values within
the Series
0
1
0(a)
Number of Forecasted Cases
1
Number of New Cases Added to the Current
Working File
0
a Melard's Algorithm will be used for estimation.
Requested Initial Configuration
Non-Seasonal Lags
Constant
AR1
AUTO
AUTO(a)
a The prior parameter value is invalid and is reset to 0.1.
Iteration History
Non-Seas
onal Lags
Adjusted Sum
Marquardt
Constant
of Squares
Constant
0
12562.91
.317
30760558.512
.001
6
1
12565.00 30742414.894
.293
.001
2
(a)
Melard's algorithm was used for estimation.
a The estimation terminated at this iteration, because the sum of squares decreased by less
than .001%.
AR1
Residual Diagnostics
Number of Residuals
Number of Parameters
Residual df
Adjusted Residual Sum of
Squares
Residual Sum of Squares
Residual Variance
Model Std. Error
Log-Likelihood
Akaike's Information
Criterion (AIC)
11
1
9
30742274
.161
30760558
.512
3387591.
382
1840.541
-97.342
198.684
Schwarz's Bayesian
Criterion (BIC)
199.479
Parameter Estimates
Estimates
.295
12564.822
Melard's algorithm was used for estimation.
Non-Seasonal Lags
Constant
AR1
Std Error
.314
759.108
t
.939
16.552
Approx Sig
.372
.000
Correlation Matrix
Non-Seas
onal Lags
AR1
Constant
1.000
0(a)
0(a)
1.000
Melard's algorithm was used for estimation.
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.
Non-Seasonal Lags
Constant
AR1
Covariance Matrix
Non-Seas
onal Lags
Non-Seasonal Lags
Constant
AR1
AR1
.099
0(a)
Constant
0(a)
576245.5
74
Melard's algorithm was used for estimation.
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.
ARIMA
Notes
Output Created
Comments
Input
02-FEB-2007 10:22:08
Data
Filter
Weight
Split File
N of Rows in Working
Data File
D:\My Documents\Cdiff\HPA Cdiff
data.sav
<none>
<none>
<none>
12
Missing Value
Handling
Date
Definition of Missing
Cases Used
Syntax
Resources
Use
Predict
Elapsed Time
From
To
From
To
Time Series
Amount of Output
Settings (TSET) Saving New Variables
Maximum Number of
Lags in Autocorrelation or
Partial Autocorrelation
Plots
Maximum Number of
Lags in Cross-Correlation
Plots
Maximum Number of New
Variables Generated Per
Procedure
Maximum Number of New
Cases Per Procedure
Treatment of
User-Missing Values
Confidence Interval
Percentage Value
Tolerance for Entering
Variables in Regression
Equations
Maximum Iterative
Parameter Change
Method of Calculating
Std. Errors for
Autocorrelations
Length of Seasonal
Period
Variable Whose Values
Label Observations in
Plots
Equations Include
Variables
FIT_2
Created or
Modified
ERR_2
<none>
User-defined missing values are
treated as missing.
Statistics are based on all cases,
including cases with any missing
value.
ARIMA Cdiff
/MODEL=( 1 1 0 )CONSTANT
/MXITER= 10
/PAREPS= .001
/SSQPCT= .001
/FORECAST= EXACT .
0:00:00.03
First observation
Last observation
First observation following the use
period
Last observation
PRINT = DEFAULT
NEWVAR = ALL
MXAUTO = 16
MXCROSS = 7
MXNEWVAR = 60
MXPREDICT = 1000
MISSING = EXCLUDE
CIN = 95
TOLER = .0001
CNVERGE = .001
ACFSE = IND
Unspecified
Unspecified
CONSTANT
Fit for Cdiff from ARIMA, MOD_2,
CON
Error for Cdiff from ARIMA, MOD_2,
CON
LCL_2
95% LCL for Cdiff from ARIMA,
MOD_2, CON
95% UCL for Cdiff from ARIMA,
MOD_2, CON
SE of Fit for Cdiff from ARIMA,
MOD_2, CON
UCL_2
SEP_2
[DataSet0]
Model Description(a)
Model Name
Dependent Series
MOD_2
Transformation
Constant
AR
Non-Seasonal
Differencing
MA
None
Included
1
Cdiff
1
None
Applying the model specifications from MOD_2
a Since there is no seasonal component in the model, the seasonality of the data will be ignored.
Iteration Termination Criteria
Maximum Parameter
Change Less Than
Maximum Marquardt
Constant Greater Than
Sum of Squares
Percentage Change
Less Than
Number of Iterations
Equal to
.001
10000000
00
.001%
10
Case Processing Summary
Series Length
Number of Cases
Skipped Due to
Missing Values
11
At the Beginning of the
Series
At the End of the
Series
Number of Cases with Missing Values within
the Series
Number of Forecasted Cases
Number of New Cases Added to the Current
Working File
0
1
0(a)
1
0
a Melard's Algorithm will be used for estimation.
Requested Initial Configuration
Non-Seasonal Lags
Constant
AR1
AUTO
AUTO(a)
a The prior parameter value is invalid and is reset to 0.1.
Iteration History
Non-Seas
onal Lags
Adjusted Sum
Marquardt
of Squares
Constant
0
45790771.240
-.010
84.169
.001
(a)
Melard's algorithm was used for estimation.
a The estimation terminated at this iteration, because the sum of squares decreased by less
than .001%.
AR1
Constant
Residual Diagnostics
Number of Residuals
Number of Parameters
Residual df
Adjusted Residual Sum of
Squares
Residual Sum of Squares
Residual Variance
Model Std. Error
Log-Likelihood
Akaike's Information
Criterion (AIC)
Schwarz's Bayesian
Criterion (BIC)
10
1
8
45790758
.869
45790771
.240
5723783.
069
2392.443
-90.990
185.980
186.585
Parameter Estimates
Estimates
Non-Seasonal Lags AR1
-.010
Constant
84.354
Melard's algorithm was used for estimation.
Correlation Matrix
Std Error
.371
749.548
t
-.028
.113
Approx Sig
.978
.913
Non-Seas
onal Lags
AR1
Constant
1.000
0(a)
0(a)
1.000
Melard's algorithm was used for estimation.
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.
Non-Seasonal Lags
Constant
AR1
Covariance Matrix
Non-Seas
onal Lags
Non-Seasonal Lags
Constant
AR1
AR1
.138
0(a)
Constant
0(a)
561822.9
09
Melard's algorithm was used for estimation.
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.
ARIMA
Notes
Output Created
Comments
Input
Missing Value
Handling
02-FEB-2007 10:23:17
Data
Filter
Weight
Split File
N of Rows in Working
Data File
Date
Definition of Missing
Cases Used
Syntax
Resources
Use
Elapsed Time
From
D:\My Documents\Cdiff\HPA Cdiff
data.sav
<none>
<none>
<none>
12
<none>
User-defined missing values are
treated as missing.
Statistics are based on all cases,
including cases with any missing
value.
ARIMA Cdiff
/MODEL=( 1 1 1 )CONSTANT
/MXITER= 10
/PAREPS= .001
/SSQPCT= .001
/FORECAST= EXACT .
0:00:00.03
First observation
Predict
To
From
To
Time Series
Amount of Output
Settings (TSET) Saving New Variables
Maximum Number of
Lags in Autocorrelation or
Partial Autocorrelation
Plots
Maximum Number of
Lags in Cross-Correlation
Plots
Maximum Number of New
Variables Generated Per
Procedure
Maximum Number of New
Cases Per Procedure
Treatment of
User-Missing Values
Confidence Interval
Percentage Value
Tolerance for Entering
Variables in Regression
Equations
Maximum Iterative
Parameter Change
Method of Calculating
Std. Errors for
Autocorrelations
Length of Seasonal
Period
Variable Whose Values
Label Observations in
Plots
Equations Include
Variables
FIT_3
Created or
Modified
ERR_3
LCL_3
UCL_3
SEP_3
Last observation
First observation following the use
period
Last observation
PRINT = DEFAULT
NEWVAR = ALL
MXAUTO = 16
MXCROSS = 7
MXNEWVAR = 60
MXPREDICT = 1000
MISSING = EXCLUDE
CIN = 95
TOLER = .0001
CNVERGE = .001
ACFSE = IND
Unspecified
Unspecified
CONSTANT
Fit for Cdiff from ARIMA, MOD_3,
CON
Error for Cdiff from ARIMA, MOD_3,
CON
95% LCL for Cdiff from ARIMA,
MOD_3, CON
95% UCL for Cdiff from ARIMA,
MOD_3, CON
SE of Fit for Cdiff from ARIMA,
MOD_3, CON
[DataSet0]
Warnings
Our tests have determined that the estimated model lies close to the boundary of
the invertibility region. Although the moving average parameters are probably
correctly estimated, their standard errors and covariances should be considered
suspect.
Model Description(a)
Model Name
Dependent Series
MOD_3
Transformation
Constant
AR
Non-Seasonal
Differencing
MA
None
Included
1
Cdiff
1
1
Applying the model specifications from MOD_3
a Since there is no seasonal component in the model, the seasonality of the data will be ignored.
Iteration Termination Criteria
Maximum Parameter
Change Less Than
Maximum Marquardt
Constant Greater Than
Sum of Squares
Percentage Change
Less Than
Number of Iterations
Equal to
.001
10000000
00
.001%
10
Case Processing Summary
Series Length
Number of Cases
Skipped Due to
Missing Values
11
At the Beginning of the
Series
At the End of the
Series
Number of Cases with Missing Values within
the Series
Number of Forecasted Cases
Number of New Cases Added to the Current
Working File
a Melard's Algorithm will be used for estimation.
Requested Initial Configuration
0
1
0(a)
1
0
Non-Seasonal
Lags
AR1
MA1
AUTO
AUTO
Constant
AUTO(a)
a The prior parameter value is invalid and is reset to 0.1.
Iteration History
Non-Seasonal Lags
Adjusted Sum
Marquardt
AR1
MA1
Constant
of Squares
Constant
0
.994
.921
66.339 55302248.049
.001
1
.942
.946
81.562 45653153.077
.001
2
.623
.946
173.494 36880980.440
.010
3
.578
.946
185.210 35948128.503
10.000
4
.481
.946
207.914 34214420.692
1.000
5
.451
.946
214.172 33765263.005
10.000
6
.430
.946
218.465 33467846.305
1.000
7
.406
.946
223.244 33149139.518
10.000
8
.403
.974
224.490 32971062.669
100.000
9
.382
.974
228.505 32718222.307
10.000
10(a)
.379
.974
228.975 32689412.691
100.000
Melard's algorithm was used for estimation.
a The estimation terminated at this iteration, because the maximum number of iterations 10 was
reached.
Residual Diagnostics
Number of Residuals
Number of Parameters
Residual df
Adjusted Residual Sum of
Squares
Residual Sum of Squares
Residual Variance
Model Std. Error
Log-Likelihood
Akaike's Information
Criterion (AIC)
Schwarz's Bayesian
Criterion (BIC)
10
2
7
32689412
.691
55302248
.049
4035793.
482
2008.928
-89.293
184.586
185.494
Parameter Estimates
Non-Seasonal
Lags
AR1
MA1
Estimates
.379
.974
Std Error
.875
8.328
t
.433
.117
Approx Sig
.678
.910
Constant
228.975
Melard's algorithm was used for estimation.
261.858
.874
.411
Correlation Matrix
Non-Seasonal Lags
AR1
MA1
Constant
Non-Seasonal
AR1
1.000
.871
0(a)
Lags
MA1
.871
1.000
0(a)
Constant
0(a)
0(a)
1.000
Melard's algorithm was used for estimation.
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.
Covariance Matrix
Non-Seasonal Lags
Non-Seasonal
Lags
AR1
MA1
Constant
AR1
.766
6.353
MA1
6.353
69.349
0(a)
0(a)
Constant
0(a)
0(a)
68569.74
7
Melard's algorithm was used for estimation.
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.
The following new variables are being created:
Name
Label
YEAR_
QUARTER_
DATE_
YEAR, not periodic
QUARTER, period 4
Date. Format: "QQ YYYY"
ARIMA
Notes
Output Created
Comments
Input
02-FEB-2007 10:26:53
Data
Filter
Weight
Split File
N of Rows in Working
Data File
Date
D:\My Documents\Cdiff\HPA Cdiff
data.sav
<none>
<none>
<none>
12
YEAR, not periodic, QUARTER,
period 4
Missing Value
Handling
Definition of Missing
Cases Used
Syntax
Resources
Use
Predict
Elapsed Time
From
To
From
To
Time Series
Amount of Output
Settings (TSET) Saving New Variables
Maximum Number of
Lags in Autocorrelation or
Partial Autocorrelation
Plots
Maximum Number of
Lags in Cross-Correlation
Plots
Maximum Number of New
Variables Generated Per
Procedure
Maximum Number of New
Cases Per Procedure
Treatment of
User-Missing Values
Confidence Interval
Percentage Value
Tolerance for Entering
Variables in Regression
Equations
Maximum Iterative
Parameter Change
Method of Calculating
Std. Errors for
Autocorrelations
Length of Seasonal
Period
Variable Whose Values
Label Observations in
Plots
Equations Include
Variables
FIT_4
Created or
Modified
ERR_4
User-defined missing values are
treated as missing.
Statistics are based on all cases,
including cases with any missing
value.
ARIMA Cdiff
/MODEL=( 1 1 1 )( 0 0 0 )
CONSTANT
/MXITER= 10
/PAREPS= .001
/SSQPCT= .001
/FORECAST= EXACT .
0:00:00.06
First observation
Last observation
First observation following the use
period
Last observation
PRINT = DEFAULT
NEWVAR = ALL
MXAUTO = 16
MXCROSS = 7
MXNEWVAR = 60
MXPREDICT = 1000
MISSING = EXCLUDE
CIN = 95
TOLER = .0001
CNVERGE = .001
ACFSE = IND
PERIOD = 4(a)
Unspecified
CONSTANT
Fit for Cdiff from ARIMA, MOD_4,
CON
Error for Cdiff from ARIMA, MOD_4,
CON
LCL_4
95% LCL for Cdiff from ARIMA,
MOD_4, CON
95% UCL for Cdiff from ARIMA,
MOD_4, CON
SE of Fit for Cdiff from ARIMA,
MOD_4, CON
UCL_4
SEP_4
a Imputed by SPSS.
[DataSet0]
Warnings
Since there is no seasonal component in the model, the seasonality of the data will
be ignored.
Our tests have determined that the estimated model lies close to the boundary of
the invertibility region. Although the moving average parameters are probably
correctly estimated, their standard errors and covariances should be considered
suspect.
Model Description(a)
Model Name
Dependent Series
MOD_4
Transformation
Constant
AR
Non-Seasonal
Differencing
MA
None
Included
1
Cdiff
1
1
Applying the model specifications from MOD_4
a Since there is no seasonal component in the model, the seasonality of the data will be ignored.
Iteration Termination Criteria
Maximum Parameter
Change Less Than
Maximum Marquardt
Constant Greater Than
Sum of Squares
Percentage Change
Less Than
Number of Iterations
Equal to
.001
10000000
00
.001%
10
Case Processing Summary
Series Length
Number of Cases
Skipped Due to
11
At the Beginning of the
Series
0
Missing Values
At the End of the
Series
Number of Cases with Missing Values within
the Series
1
0(a)
Number of Forecasted Cases
1
Number of New Cases Added to the Current
Working File
0
a Melard's Algorithm will be used for estimation.
Requested Initial Configuration
Non-Seasonal
Lags
AR1
MA1
AUTO
AUTO
Constant
AUTO(a)
a The prior parameter value is invalid and is reset to 0.1.
Iteration History
Non-Seasonal Lags
Adjusted Sum
Marquardt
AR1
MA1
Constant
of Squares
Constant
0
.994
.921
66.339 55302248.049
.001
1
.942
.946
81.562 45653153.077
.001
2
.623
.946
173.494 36880980.440
.010
3
.578
.946
185.210 35948128.503
10.000
4
.481
.946
207.914 34214420.692
1.000
5
.451
.946
214.172 33765263.005
10.000
6
.430
.946
218.465 33467846.305
1.000
7
.406
.946
223.244 33149139.518
10.000
8
.403
.974
224.490 32971062.669
100.000
9
.382
.974
228.505 32718222.307
10.000
10(a)
.379
.974
228.975 32689412.691
100.000
Melard's algorithm was used for estimation.
a The estimation terminated at this iteration, because the maximum number of iterations 10 was
reached.
Residual Diagnostics
Number of Residuals
Number of Parameters
Residual df
Adjusted Residual Sum of
Squares
Residual Sum of Squares
Residual Variance
10
2
7
32689412
.691
55302248
.049
4035793.
Model Std. Error
Log-Likelihood
Akaike's Information
Criterion (AIC)
Schwarz's Bayesian
Criterion (BIC)
482
2008.928
-89.293
184.586
185.494
Parameter Estimates
Estimates
Std Error
.379
.875
.974
8.328
Constant
228.975
261.858
Melard's algorithm was used for estimation.
Non-Seasonal
Lags
AR1
MA1
t
.433
.117
.874
Approx Sig
.678
.910
.411
Correlation Matrix
Non-Seasonal Lags
AR1
MA1
Constant
1.000
.871
0(a)
.871
1.000
0(a)
Constant
0(a)
0(a)
1.000
Melard's algorithm was used for estimation.
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.
Non-Seasonal
Lags
AR1
MA1
Covariance Matrix
Non-Seasonal Lags
Non-Seasonal
Lags
AR1
MA1
Constant
AR1
.766
6.353
MA1
6.353
69.349
0(a)
0(a)
Constant
0(a)
0(a)
68569.74
7
Melard's algorithm was used for estimation.
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.
ARIMA
Notes
Output Created
Comments
Input
02-FEB-2007 10:27:28
Data
D:\My Documents\Cdiff\HPA Cdiff
data.sav
Filter
Weight
Split File
N of Rows in Working
Data File
Date
Missing Value
Handling
Definition of Missing
Cases Used
Syntax
Resources
Use
Predict
Elapsed Time
From
To
From
To
Time Series
Amount of Output
Settings (TSET) Saving New Variables
Maximum Number of
Lags in Autocorrelation or
Partial Autocorrelation
Plots
Maximum Number of
Lags in Cross-Correlation
Plots
Maximum Number of New
Variables Generated Per
Procedure
Maximum Number of New
Cases Per Procedure
Treatment of
User-Missing Values
Confidence Interval
Percentage Value
Tolerance for Entering
Variables in Regression
Equations
Maximum Iterative
Parameter Change
Method of Calculating
Std. Errors for
Autocorrelations
Length of Seasonal
Period
<none>
<none>
<none>
12
YEAR, not periodic, QUARTER,
period 4
User-defined missing values are
treated as missing.
Statistics are based on all cases,
including cases with any missing
value.
ARIMA Cdiff
/MODEL=( 1 1 1 )( 1 0 0 )
CONSTANT
/MXITER= 10
/PAREPS= .001
/SSQPCT= .001
/FORECAST= EXACT .
0:00:00.03
First observation
Last observation
First observation following the use
period
Last observation
PRINT = DEFAULT
NEWVAR = ALL
MXAUTO = 16
MXCROSS = 7
MXNEWVAR = 60
MXPREDICT = 1000
MISSING = EXCLUDE
CIN = 95
TOLER = .0001
CNVERGE = .001
ACFSE = IND
PERIOD = 4(a)
Variables
Created or
Modified
Variable Whose Values
Label Observations in
Plots
Equations Include
FIT_5
ERR_5
LCL_5
UCL_5
SEP_5
Unspecified
CONSTANT
Fit for Cdiff from ARIMA, MOD_5,
CON
Error for Cdiff from ARIMA, MOD_5,
CON
95% LCL for Cdiff from ARIMA,
MOD_5, CON
95% UCL for Cdiff from ARIMA,
MOD_5, CON
SE of Fit for Cdiff from ARIMA,
MOD_5, CON
a Imputed by SPSS.
[DataSet0]
Warnings
Our tests have determined that the estimated model lies close to the boundary of
the invertibility region. Although the moving average parameters are probably
correctly estimated, their standard errors and covariances should be considered
suspect.
Model Description
Model Name
Dependent Series
Transformation
Constant
AR
Non-Seasonal
Differencing
MA
Seasonal AR
Seasonal Differencing
MOD_5
Cdiff
None
Included
1
Seasonal MA
Length of Seasonal
Period
None
1
1
1
0
4
Applying the model specifications from MOD_5
Iteration Termination Criteria
Maximum Parameter
Change Less Than
Maximum Marquardt
Constant Greater Than
.001
10000000
00
Sum of Squares
Percentage Change
Less Than
Number of Iterations
Equal to
.001%
10
Case Processing Summary
Series Length
Number of Cases
Skipped Due to
Missing Values
11
At the Beginning of the
Series
At the End of the
Series
0
1
Number of Cases with Missing Values within
the Series
0(a)
Number of Forecasted Cases
1
Number of New Cases Added to the Current
Working File
0
a Melard's Algorithm will be used for estimation.
Requested Initial Configuration
Non-Seasonal
Lags
Seasonal Lags
AR1
MA1
Seasonal AR1
AUTO
AUTO
AUTO
Constant
AUTO(a)
a The prior parameter value is invalid and is reset to 0.1.
Iteration History
Non-Seasonal Lags
0
1
2
3
4
5
6
7
8
9
10(a)
AR1
.994
.971
.826
.822
.754
.739
.751
.743
.742
.735
.734
MA1
.921
.962
.873
.873
.954
.979
.979
.979
.981
.981
.985
Seasonal
Lags
Seasonal
AR1
.461
.931
.873
.869
.872
.876
.905
.904
.904
.903
.903
Constant
275.574
542.077
473.215
469.504
427.220
423.000
433.040
430.909
430.496
428.708
428.345
Adjusted Sum
of Squares
35279136.655
19293829.818
16905542.530
16857035.607
14162692.960
13682672.788
13668154.252
13631952.254
13589437.097
13561749.753
13522603.405
Marquardt
Constant
.001
.001
.000
.100
1.000
10.000
1.000
10.000
100.000
10.000
100.000
Melard's algorithm was used for estimation.
a The estimation terminated at this iteration, because the maximum number of iterations 10 was
reached.
Residual Diagnostics
Number of Residuals
Number of Parameters
Residual df
Adjusted Residual Sum of
Squares
Residual Sum of Squares
Residual Variance
Model Std. Error
Log-Likelihood
Akaike's Information
Criterion (AIC)
Schwarz's Bayesian
Criterion (BIC)
10
3
6
13522603
.405
35279136
.655
1334135.
826
1155.048
-84.776
177.553
178.763
Parameter Estimates
Estimates
.734
.985
Std Error
.529
.676
t
1.387
1.456
Approx Sig
.215
.196
.903
.149
6.048
.001
Constant
428.345
Melard's algorithm was used for estimation.
304.285
1.408
.209
Non-Seasonal
Lags
Seasonal Lags
AR1
MA1
Seasonal AR1
Correlation Matrix
Non-Seasonal Lags
Non-Seasonal
Lags
Seasonal Lags
AR1
MA1
Seasonal AR1
Seasonal
Lags
AR1
1.000
.871
MA1
.871
1.000
Seasonal
AR1
.491
.518
Constant
0(a)
0(a)
.491
.518
1.000
0(a)
Constant
0(a)
0(a)
0(a)
1.000
Melard's algorithm was used for estimation.
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.
Covariance Matrix
Non-Seasonal Lags
Non-Seasonal
Lags
Seasonal Lags
AR1
MA1
Seasonal AR1
Seasonal
Lags
AR1
.280
.311
MA1
.311
.457
Seasonal
AR1
.039
.052
Constant
0(a)
0(a)
.039
.052
.022
0(a)
0(a)
0(a)
0(a)
92589.58
6
Constant
Melard's algorithm was used for estimation.
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.
ARIMA
Notes
Output Created
Comments
Input
02-FEB-2007 10:28:46
Data
Filter
Weight
Split File
N of Rows in Working
Data File
Date
Missing Value
Handling
Definition of Missing
Cases Used
Syntax
Resources
Use
Predict
Elapsed Time
From
To
From
To
Time Series
Amount of Output
Settings (TSET) Saving New Variables
D:\My Documents\Cdiff\HPA Cdiff
data.sav
<none>
<none>
<none>
12
YEAR, not periodic, QUARTER,
period 4
User-defined missing values are
treated as missing.
Statistics are based on all cases,
including cases with any missing
value.
ARIMA Cdiff
/MODEL=( 2 1 0 )( 1 0 0 )
CONSTANT
/MXITER= 10
/PAREPS= .001
/SSQPCT= .001
/FORECAST= EXACT .
0:00:00.05
First observation
Last observation
First observation following the use
period
Last observation
PRINT = DEFAULT
NEWVAR = ALL
Variables
Created or
Modified
Maximum Number of
Lags in Autocorrelation or
Partial Autocorrelation
Plots
Maximum Number of
Lags in Cross-Correlation
Plots
Maximum Number of New
Variables Generated Per
Procedure
Maximum Number of New
Cases Per Procedure
Treatment of
User-Missing Values
Confidence Interval
Percentage Value
Tolerance for Entering
Variables in Regression
Equations
Maximum Iterative
Parameter Change
Method of Calculating
Std. Errors for
Autocorrelations
Length of Seasonal
Period
Variable Whose Values
Label Observations in
Plots
Equations Include
FIT_6
ERR_6
LCL_6
UCL_6
SEP_6
MXAUTO = 16
MXCROSS = 7
MXNEWVAR = 60
MXPREDICT = 1000
MISSING = EXCLUDE
CIN = 95
TOLER = .0001
CNVERGE = .001
ACFSE = IND
PERIOD = 4(a)
Unspecified
CONSTANT
Fit for Cdiff from ARIMA, MOD_6,
CON
Error for Cdiff from ARIMA, MOD_6,
CON
95% LCL for Cdiff from ARIMA,
MOD_6, CON
95% UCL for Cdiff from ARIMA,
MOD_6, CON
SE of Fit for Cdiff from ARIMA,
MOD_6, CON
a Imputed by SPSS.
[DataSet0]
Model Description
Model Name
Dependent Series
Transformation
Constant
AR
Non-Seasonal
Differencing
MOD_6
Cdiff
None
Included
1, 2
1
MA
Seasonal AR
Seasonal Differencing
None
1
Seasonal MA
Length of Seasonal
Period
None
0
4
Applying the model specifications from MOD_6
Iteration Termination Criteria
Maximum Parameter
Change Less Than
Maximum Marquardt
Constant Greater Than
Sum of Squares
Percentage Change
Less Than
Number of Iterations
Equal to
.001
10000000
00
.001%
10
Case Processing Summary
Series Length
Number of Cases
Skipped Due to
Missing Values
11
At the Beginning of the
Series
At the End of the
Series
Number of Cases with Missing Values within
the Series
0
1
0(a)
Number of Forecasted Cases
1
Number of New Cases Added to the Current
Working File
a Melard's Algorithm will be used for estimation.
Requested Initial Configuration
Non-Seasonal
Lags
Seasonal Lags
AR1
AR2
Seasonal AR1
AUTO
AUTO
AUTO
Constant
AUTO(a)
a The prior parameter value is invalid and is reset to 0.1.
Iteration History
0
Seasonal
Lags
Non-Seasonal Lags
0
1
2
3
4
5
6
7
8
9
AR1
-.016
.176
.364
.413
.447
.461
.471
.478
.482
AR2
-.661
-.449
-.466
-.531
-.544
-.559
-.566
-.572
-.576
Seasonal
AR1
.461
.901
.892
.913
.916
.920
.922
.923
.924
Constant
343.202
408.620
387.492
377.337
371.995
368.808
366.668
365.271
364.318
.485
-.578
.924
363.671
Adjusted Sum
of Squares
17616952.451
13950374.841
13232339.424
13138679.439
13117238.472
13110021.281
13107110.527
13105828.188
13105243.862
13104972.067
(a)
Marquardt
Constant
.001
.001
.000
.000
.000
.000
.000
.000
.000
.000
Melard's algorithm was used for estimation.
a The estimation terminated at this iteration, because the sum of squares decreased by less
than .001%.
Residual Diagnostics
Number of Residuals
Number of Parameters
Residual df
Adjusted Residual Sum of
Squares
Residual Sum of Squares
Residual Variance
Model Std. Error
Log-Likelihood
Akaike's Information
Criterion (AIC)
Schwarz's Bayesian
Criterion (BIC)
10
3
6
13104844
.119
17616952
.451
907237.4
20
952.490
-85.003
178.005
179.215
Parameter Estimates
Estimates
.487
-.580
Std Error
.267
.219
t
1.821
-2.654
Approx Sig
.118
.038
.925
.069
13.479
.000
Constant
363.227
Melard's algorithm was used for estimation.
1117.990
.325
.756
Non-Seasonal
Lags
Seasonal Lags
AR1
AR2
Seasonal AR1
Correlation Matrix
Non-Seasonal Lags
Non-Seasonal
Lags
Seasonal Lags
AR1
AR2
Seasonal AR1
Seasonal
Lags
AR1
1.000
-.424
AR2
-.424
1.000
Seasonal
AR1
.585
-.476
Constant
0(a)
0(a)
.585
-.476
1.000
0(a)
Constant
0(a)
0(a)
0(a)
1.000
Melard's algorithm was used for estimation.
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.
Covariance Matrix
Non-Seasonal Lags
Non-Seasonal
Lags
Seasonal Lags
AR1
AR2
Seasonal AR1
Seasonal
Lags
AR1
.071
-.025
AR2
-.025
.048
Seasonal
AR1
.011
-.007
Constant
0(a)
0(a)
.011
-.007
.005
0(a)
0(a)
0(a)
0(a)
1249901.
744
Constant
Melard's algorithm was used for estimation.
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.
ARIMA
Notes
Output Created
Comments
Input
02-FEB-2007 10:29:30
Data
Filter
Weight
Split File
N of Rows in Working
Data File
Date
Missing Value
Handling
Definition of Missing
Cases Used
D:\My Documents\Cdiff\HPA Cdiff
data.sav
<none>
<none>
<none>
12
YEAR, not periodic, QUARTER,
period 4
User-defined missing values are
treated as missing.
Statistics are based on all cases,
including cases with any missing
value.
Syntax
Resources
Use
Predict
ARIMA Cdiff
/MODEL=( 2 1 1 )( 1 0 0 )
CONSTANT
/MXITER= 10
/PAREPS= .001
/SSQPCT= .001
/FORECAST= EXACT .
Elapsed Time
From
To
From
To
Time Series
Amount of Output
Settings (TSET) Saving New Variables
Maximum Number of
Lags in Autocorrelation or
Partial Autocorrelation
Plots
Maximum Number of
Lags in Cross-Correlation
Plots
Maximum Number of New
Variables Generated Per
Procedure
Maximum Number of New
Cases Per Procedure
Treatment of
User-Missing Values
Confidence Interval
Percentage Value
Tolerance for Entering
Variables in Regression
Equations
Maximum Iterative
Parameter Change
Method of Calculating
Std. Errors for
Autocorrelations
Length of Seasonal
Period
Variable Whose Values
Label Observations in
Plots
Equations Include
Variables
FIT_7
Created or
Modified
ERR_7
LCL_7
UCL_7
0:00:00.05
First observation
Last observation
First observation following the use
period
Last observation
PRINT = DEFAULT
NEWVAR = ALL
MXAUTO = 16
MXCROSS = 7
MXNEWVAR = 60
MXPREDICT = 1000
MISSING = EXCLUDE
CIN = 95
TOLER = .0001
CNVERGE = .001
ACFSE = IND
PERIOD = 4(a)
Unspecified
CONSTANT
Fit for Cdiff from ARIMA, MOD_7,
CON
Error for Cdiff from ARIMA, MOD_7,
CON
95% LCL for Cdiff from ARIMA,
MOD_7, CON
95% UCL for Cdiff from ARIMA,
MOD_7, CON
SEP_7
SE of Fit for Cdiff from ARIMA,
MOD_7, CON
a Imputed by SPSS.
[DataSet0]
Warnings
Our tests have determined that the estimated model lies close to the boundary of
the invertibility region. Although the moving average parameters are probably
correctly estimated, their standard errors and covariances should be considered
suspect.
Model Description
Model Name
Dependent Series
Transformation
Constant
AR
Non-Seasonal
Differencing
MA
Seasonal AR
Seasonal Differencing
MOD_7
Cdiff
None
Included
1, 2
Seasonal MA
Length of Seasonal
Period
None
1
1
1
0
4
Applying the model specifications from MOD_7
Iteration Termination Criteria
Maximum Parameter
Change Less Than
Maximum Marquardt
Constant Greater Than
Sum of Squares
Percentage Change
Less Than
Number of Iterations
Equal to
.001
10000000
00
.001%
10
Case Processing Summary
Series Length
Number of Cases
Skipped Due to
11
At the Beginning of the
Series
0
Missing Values
At the End of the
Series
1
Number of Cases with Missing Values within
the Series
0(a)
Number of Forecasted Cases
1
Number of New Cases Added to the Current
Working File
0
a Melard's Algorithm will be used for estimation.
Requested Initial Configuration
Non-Seasonal
Lags
Seasonal Lags
AR1
AR2
MA1
Seasonal AR1
AUTO
AUTO
AUTO
AUTO
Constant
AUTO(a)
a The prior parameter value is invalid and is reset to 0.1.
Iteration History
Seasonal
Lags
Non-Seasonal Lags
Seasonal
Adjusted Sum
Marquardt
AR1
AR2
MA1
AR1
Constant
of Squares
Constant
0
.377
-.657
.957
.461
288.797 11557279.393
.001
1
.670
-.530
.957
.969
330.159 11250895.554
.001
2
.835
-.584
.927
.883
323.735
8927115.823
.000
3
.951
-.635
.927
.950
332.471
8596064.385
.100
4
.979
-.657
.977
.947
330.429
7910856.534
1.000
5
.978
-.663
.977
.948
329.473
7881716.415
10.000
6
.995
-.682
.977
.956
328.907
7813515.619
1.000
7
.994
-.688
.999
.956
327.877
7689790.156
10.000
8
1.008
-.703
.999
.960
326.880
7639227.606
1.000
9
1.007
-.708
.999
.959
325.984
7620033.729
10.000
10(a)
1.015
-.720
.999
.961
324.872
7583259.752
1.000
Melard's algorithm was used for estimation.
a The estimation terminated at this iteration, because the maximum number of iterations 10 was
reached.
Residual Diagnostics
Number of Residuals
Number of Parameters
Residual df
10
4
5
Adjusted Residual Sum of
Squares
Residual Sum of Squares
Residual Variance
Model Std. Error
Log-Likelihood
Akaike's Information
Criterion (AIC)
Schwarz's Bayesian
Criterion (BIC)
7583259.
752
11557279
.393
626672.6
81
791.627
-81.981
173.961
175.474
Parameter Estimates
Estimates
1.015
-.720
.999
Std Error
.227
.191
3.141
t
4.467
-3.780
.318
Approx Sig
.007
.013
.763
.961
.059
16.207
.000
Constant
324.872
Melard's algorithm was used for estimation.
116.982
2.777
.039
Non-Seasonal
Lags
Seasonal Lags
AR1
AR2
MA1
Seasonal AR1
Correlation Matrix
Seasonal
Lags
Non-Seasonal Lags
Non-Seasonal
Lags
Seasonal Lags
AR1
AR2
MA1
Seasonal AR1
AR1
1.000
-.518
-.033
AR2
-.518
1.000
.369
MA1
-.033
.369
1.000
Seasonal
AR1
.380
-.384
-.040
Constant
0(a)
0(a)
0(a)
.380
-.384
-.040
1.000
0(a)
Constant
0(a)
0(a)
0(a)
0(a)
1.000
Melard's algorithm was used for estimation.
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.
Covariance Matrix
Seasonal
Lags
Non-Seasonal Lags
Non-Seasonal
Lags
AR1
AR2
MA1
AR1
.052
-.022
-.024
AR2
-.022
.036
.221
MA1
-.024
.221
9.864
Seasonal
AR1
.005
-.004
-.007
Constant
0(a)
0(a)
0(a)
Seasonal Lags
Seasonal AR1
Constant
.005
-.004
-.007
.004
0(a)
0(a)
0(a)
0(a)
0(a)
13684.75
5
Melard's algorithm was used for estimation.
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.
The following new variables are being created:
Name
Label
YEAR_
QUARTER_
DATE_
YEAR, not periodic
QUARTER, period 4
Date. Format: "QQ YYYY"
ARIMA
Notes
Output Created
Comments
Input
02-FEB-2007 10:38:33
Data
Filter
Weight
Split File
N of Rows in Working
Data File
Date
Missing Value
Handling
Definition of Missing
Cases Used
Syntax
Resources
Use
Predict
Elapsed Time
From
To
From
To
Time Series
Amount of Output
Settings (TSET) Saving New Variables
D:\My Documents\Cdiff\HPA MRSA
data.sav
<none>
<none>
<none>
11
YEAR, not periodic, QUARTER,
period 4
User-defined missing values are
treated as missing.
Statistics are based on all cases,
including cases with any missing
value.
ARIMA MRSA
/MODEL=( 1 1 1 )( 0 0 0 )
CONSTANT
/MXITER= 10
/PAREPS= .001
/SSQPCT= .001
/FORECAST= EXACT .
0:00:00.00
First observation
Last observation
First observation following the use
period
Last observation
PRINT = DEFAULT
NEWVAR = ALL
Variables
Created or
Modified
Maximum Number of
Lags in Autocorrelation or
Partial Autocorrelation
Plots
Maximum Number of
Lags in Cross-Correlation
Plots
Maximum Number of New
Variables Generated Per
Procedure
Maximum Number of New
Cases Per Procedure
Treatment of
User-Missing Values
Confidence Interval
Percentage Value
Tolerance for Entering
Variables in Regression
Equations
Maximum Iterative
Parameter Change
Method of Calculating
Std. Errors for
Autocorrelations
Length of Seasonal
Period
Variable Whose Values
Label Observations in
Plots
Equations Include
FIT_1
ERR_1
LCL_1
UCL_1
SEP_1
MXAUTO = 16
MXCROSS = 7
MXNEWVAR = 5
MXPREDICT = 1000
MISSING = EXCLUDE
CIN = 95
TOLER = .0001
CNVERGE = .001
ACFSE = IND
PERIOD = 4(a)
Unspecified
CONSTANT
Fit for MRSA from ARIMA, MOD_8,
CON
Error for MRSA from ARIMA,
MOD_8, CON
95% LCL for MRSA from ARIMA,
MOD_8, CON
95% UCL for MRSA from ARIMA,
MOD_8, CON
SE of Fit for MRSA from ARIMA,
MOD_8, CON
a Imputed by SPSS.
[DataSet1]
Warnings
Since there is no seasonal component in the model, the seasonality of the data will
be ignored.
Model Description(a)
Model Name
MOD_8
Dependent Series
Transformation
Constant
AR
Non-Seasonal
Differencing
MA
MRSA
None
Included
1
1
1
Applying the model specifications from MOD_8
a Since there is no seasonal component in the model, the seasonality of the data will be ignored.
Iteration Termination Criteria
Maximum Parameter
Change Less Than
Maximum Marquardt
Constant Greater Than
Sum of Squares
Percentage Change
Less Than
Number of Iterations
Equal to
.001
10000000
00
.001%
10
Case Processing Summary
Series Length
Number of Cases
Skipped Due to
Missing Values
11
At the Beginning of the
Series
At the End of the
Series
Number of Cases with Missing Values within
the Series
0
0
0(a)
Number of Forecasted Cases
0
Number of New Cases Added to the Current
Working File
a Melard's Algorithm will be used for estimation.
Requested Initial Configuration
Non-Seasonal
Lags
AR1
MA1
AUTO
AUTO
Constant
AUTO(a)
a The prior parameter value is invalid and is reset to 0.1.
Iteration History
0
Non-Seasonal Lags
Adjusted Sum
Marquardt
AR1
MA1
Constant
of Squares
Constant
0
-.871
-.057
-21.313
210331.644
.001
1
-.647
-.240
-22.450
184454.357
.001
2
-.917
-.481
-23.021
180013.309
.000
3
-.818
-.426
-22.908
176435.038
.000
4
-.850
-.431
-22.919
175973.574
.000
5
-.840
-.429
-22.915
175924.537
.000
6
-.844
-.430
-22.917 175918.427(a)
.000
Melard's algorithm was used for estimation.
a The estimation terminated at this iteration, because the sum of squares decreased by less
than .001%.
Residual Diagnostics
Number of Residuals
Number of Parameters
Residual df
Adjusted Residual Sum of
Squares
Residual Sum of Squares
Residual Variance
Model Std. Error
Log-Likelihood
Akaike's Information
Criterion (AIC)
Schwarz's Bayesian
Criterion (BIC)
10
2
7
175917.6
46
210331.6
44
23808.11
0
154.299
-63.273
132.546
133.453
Parameter Estimates
Estimates
Std Error
-.842
.298
-.430
.537
Constant
-22.916
38.484
Melard's algorithm was used for estimation.
Non-Seasonal
Lags
AR1
MA1
t
-2.826
-.801
-.595
Correlation Matrix
Non-Seasonal Lags
AR1
1.000
.800
Constant
0(a)
Melard's algorithm was used for estimation.
Non-Seasonal
Lags
AR1
MA1
MA1
.800
1.000
0(a)
Constant
0(a)
0(a)
1.000
Approx Sig
.026
.449
.570
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.
Covariance Matrix
Non-Seasonal Lags
AR1
MA1
Constant
.089
.128
0(a)
.128
.288
0(a)
Constant
0(a)
0(a)
1480.989
Melard's algorithm was used for estimation.
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.
Non-Seasonal
Lags
AR1
MA1
ARIMA
Notes
Output Created
Comments
Input
02-FEB-2007 10:39:03
Data
Filter
Weight
Split File
N of Rows in Working
Data File
Date
Missing Value
Handling
Definition of Missing
Cases Used
Syntax
Resources
Use
Predict
Elapsed Time
From
To
From
To
Time Series
Amount of Output
Settings (TSET) Saving New Variables
D:\My Documents\Cdiff\HPA MRSA
data.sav
<none>
<none>
<none>
11
YEAR, not periodic, QUARTER,
period 4
User-defined missing values are
treated as missing.
Statistics are based on all cases,
including cases with any missing
value.
ARIMA MRSA
/MODEL=( 1 1 1 )( 1 0 0 )
CONSTANT
/MXITER= 10
/PAREPS= .001
/SSQPCT= .001
/FORECAST= EXACT .
0:00:00.02
First observation
Last observation
First observation following the use
period
Last observation
PRINT = DEFAULT
NEWVAR = ALL
Variables
Created or
Modified
Maximum Number of
Lags in Autocorrelation or
Partial Autocorrelation
Plots
Maximum Number of
Lags in Cross-Correlation
Plots
Maximum Number of New
Variables Generated Per
Procedure
Maximum Number of New
Cases Per Procedure
Treatment of
User-Missing Values
Confidence Interval
Percentage Value
Tolerance for Entering
Variables in Regression
Equations
Maximum Iterative
Parameter Change
Method of Calculating
Std. Errors for
Autocorrelations
Length of Seasonal
Period
Variable Whose Values
Label Observations in
Plots
Equations Include
FIT_2
ERR_2
LCL_2
UCL_2
SEP_2
MXAUTO = 16
MXCROSS = 7
MXNEWVAR = 5
MXPREDICT = 1000
MISSING = EXCLUDE
CIN = 95
TOLER = .0001
CNVERGE = .001
ACFSE = IND
PERIOD = 4(a)
Unspecified
CONSTANT
Fit for MRSA from ARIMA, MOD_9,
CON
Error for MRSA from ARIMA,
MOD_9, CON
95% LCL for MRSA from ARIMA,
MOD_9, CON
95% UCL for MRSA from ARIMA,
MOD_9, CON
SE of Fit for MRSA from ARIMA,
MOD_9, CON
a Imputed by SPSS.
[DataSet1]
Model Description
Model Name
Dependent Series
Transformation
Constant
AR
Non-Seasonal
Differencing
MOD_9
MRSA
None
Included
1
1
MA
Seasonal AR
Seasonal Differencing
1
1
Seasonal MA
Length of Seasonal
Period
None
0
4
Applying the model specifications from MOD_9
Iteration Termination Criteria
Maximum Parameter
Change Less Than
Maximum Marquardt
Constant Greater Than
Sum of Squares
Percentage Change
Less Than
Number of Iterations
Equal to
.001
10000000
00
.001%
10
Case Processing Summary
Series Length
Number of Cases
Skipped Due to
Missing Values
11
At the Beginning of the
Series
At the End of the
Series
Number of Cases with Missing Values within
the Series
0
0
0(a)
Number of Forecasted Cases
0
Number of New Cases Added to the Current
Working File
a Melard's Algorithm will be used for estimation.
Requested Initial Configuration
Non-Seasonal
Lags
Seasonal Lags
AR1
MA1
Seasonal AR1
AUTO
AUTO
AUTO
Constant
AUTO(a)
a The prior parameter value is invalid and is reset to 0.1.
Iteration History
0
Seasonal
Lags
Non-Seasonal Lags
Seasonal
Adjusted Sum
Marquardt
AR1
MA1
AR1
Constant
of Squares
Constant
0
-.871
-.057
.264
-18.601
239768.119
.001
1
-.554
-.156
.059
-21.473
188459.539
.001
2
-.946
-.503
.019
-22.799
187143.223
.010
3
-.815
-.439
-.079
-24.026
177947.314
.001
4
-.908
-.492
-.084
-24.199
175358.753
.000
5
-.905
-.534
-.152
-25.194
174503.592
.000
6
-.949
-.588
-.179
-25.648
173862.247
.000
7
-.962
-.662
-.252
-26.685
172586.870
.000
8
-.969
-.667
-.250
-26.663
172229.198
1.000
9
-.987
-.743
-.298
-27.313
170779.850
.100
10(a)
-.986
-.778
-.319
-27.569
170363.629
1.000
Melard's algorithm was used for estimation.
a The estimation terminated at this iteration, because the maximum number of iterations 10 was
reached.
Residual Diagnostics
Number of Residuals
Number of Parameters
Residual df
Adjusted Residual Sum of
Squares
Residual Sum of Squares
Residual Variance
Model Std. Error
Log-Likelihood
Akaike's Information
Criterion (AIC)
Schwarz's Bayesian
Criterion (BIC)
10
3
6
170363.6
29
239768.1
19
23879.69
8
154.531
-63.160
134.321
135.531
Parameter Estimates
Estimates
-.986
-.778
Std Error
.132
.868
t
-7.456
-.896
Approx Sig
.000
.405
-.319
.551
-.579
.583
Constant
-27.569
Melard's algorithm was used for estimation.
37.127
-.743
.486
Non-Seasonal
Lags
Seasonal Lags
AR1
MA1
Seasonal AR1
Correlation Matrix
Non-Seasonal Lags
Non-Seasonal
Lags
Seasonal Lags
AR1
MA1
Seasonal AR1
Seasonal
Lags
AR1
1.000
.962
MA1
.962
1.000
Seasonal
AR1
.644
.600
Constant
0(a)
0(a)
.644
.600
1.000
0(a)
Constant
0(a)
0(a)
0(a)
1.000
Melard's algorithm was used for estimation.
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.
Covariance Matrix
Non-Seasonal Lags
Non-Seasonal
Lags
Seasonal Lags
AR1
MA1
Seasonal AR1
Seasonal
Lags
AR1
.018
.110
MA1
.110
.753
Seasonal
AR1
.047
.287
Constant
0(a)
0(a)
.047
.287
.304
0(a)
Constant
0(a)
0(a)
0(a)
1378.398
Melard's algorithm was used for estimation.
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.
ARIMA
Notes
Output Created
Comments
Input
02-FEB-2007 10:39:29
Data
Filter
Weight
Split File
N of Rows in Working
Data File
Date
Missing Value
Handling
Definition of Missing
Cases Used
D:\My Documents\Cdiff\HPA MRSA
data.sav
<none>
<none>
<none>
11
YEAR, not periodic, QUARTER,
period 4
User-defined missing values are
treated as missing.
Statistics are based on all cases,
including cases with any missing
value.
Syntax
Resources
Use
Predict
ARIMA MRSA
/MODEL=( 2 1 1 )( 1 0 0 )
CONSTANT
/MXITER= 10
/PAREPS= .001
/SSQPCT= .001
/FORECAST= EXACT .
Elapsed Time
From
To
From
To
Time Series
Amount of Output
Settings (TSET) Saving New Variables
Maximum Number of
Lags in Autocorrelation or
Partial Autocorrelation
Plots
Maximum Number of
Lags in Cross-Correlation
Plots
Maximum Number of New
Variables Generated Per
Procedure
Maximum Number of New
Cases Per Procedure
Treatment of
User-Missing Values
Confidence Interval
Percentage Value
Tolerance for Entering
Variables in Regression
Equations
Maximum Iterative
Parameter Change
Method of Calculating
Std. Errors for
Autocorrelations
Length of Seasonal
Period
Variable Whose Values
Label Observations in
Plots
Equations Include
Variables
FIT_3
Created or
Modified
ERR_3
LCL_3
UCL_3
0:00:00.03
First observation
Last observation
First observation following the use
period
Last observation
PRINT = DEFAULT
NEWVAR = ALL
MXAUTO = 16
MXCROSS = 7
MXNEWVAR = 5
MXPREDICT = 1000
MISSING = EXCLUDE
CIN = 95
TOLER = .0001
CNVERGE = .001
ACFSE = IND
PERIOD = 4(a)
Unspecified
CONSTANT
Fit for MRSA from ARIMA,
MOD_10, CON
Error for MRSA from ARIMA,
MOD_10, CON
95% LCL for MRSA from ARIMA,
MOD_10, CON
95% UCL for MRSA from ARIMA,
MOD_10, CON
SEP_3
SE of Fit for MRSA from ARIMA,
MOD_10, CON
a Imputed by SPSS.
[DataSet1]
Warnings
Our tests have determined that the estimated model lies close to the boundary of
the invertibility region. Although the moving average parameters are probably
correctly estimated, their standard errors and covariances should be considered
suspect.
Model Description
Model Name
Dependent Series
Transformation
Constant
AR
Non-Seasonal
Differencing
MA
Seasonal AR
Seasonal Differencing
MOD_10
MRSA
None
Included
1, 2
Seasonal MA
Length of Seasonal
Period
None
1
1
1
0
4
Applying the model specifications from MOD_10
Iteration Termination Criteria
Maximum Parameter
Change Less Than
Maximum Marquardt
Constant Greater Than
Sum of Squares
Percentage Change
Less Than
Number of Iterations
Equal to
.001
10000000
00
.001%
10
Case Processing Summary
Series Length
Number of Cases
Skipped Due to
11
At the Beginning of the
Series
0
Missing Values
At the End of the
Series
0
Number of Cases with Missing Values within
the Series
0(a)
Number of Forecasted Cases
0
Number of New Cases Added to the Current
Working File
0
a Melard's Algorithm will be used for estimation.
Requested Initial Configuration
Non-Seasonal
Lags
Seasonal Lags
AR1
AR2
MA1
Seasonal AR1
AUTO
AUTO
AUTO
AUTO
Constant
AUTO(a)
a The prior parameter value is invalid and is reset to 0.1.
Iteration History
Non-Seasonal Lags
Seasonal
Lags
Seasonal
Adjusted Sum
Marquardt
AR1
AR2
MA1
AR1
Constant
of Squares
Constant
0
-.668
.127
-.042
.264
-18.873
204570.730
.001
1
-.244
.286
.141
.057
-22.473
180604.159
.001
2
-.826
.012
-.395
.095
-21.594
179438.547
.000
3
-1.224
-.245
-.793
-.122
-23.612
169230.350
.001
4
-1.226
-.242
-.814
-.168
-24.182
167647.495
1.000
5
-1.226
-.227
-.950
-.269
-25.504
164190.100
.100
6
-1.226
-.227
-.946
-.262
-25.420
163967.512
1.000
7
-1.226
-.227
-.948
-.266
-25.476
163148.288
10.000
8
-1.226(a)
-.227(a)
-.950(a)
-.275(a) -25.572(a)
163137.296
1.000
Melard's algorithm was used for estimation.
a The estimation terminated at this iteration, because all the parameter estimates changed by
less than .001.
Residual Diagnostics
Number of Residuals
Number of Parameters
Residual df
Adjusted Residual Sum of
Squares
Residual Sum of Squares
10
4
5
163102.4
99
204570.7
Residual Variance
Model Std. Error
Log-Likelihood
Akaike's Information
Criterion (AIC)
Schwarz's Bayesian
Criterion (BIC)
30
25308.89
0
159.088
-63.106
136.212
137.725
Parameter Estimates
Estimates
-1.226
-.227
-.950
Std Error
.065
.059
3.133
t
-18.919
-3.854
-.303
Approx Sig
.000
.012
.774
-.274
.524
-.523
.623
-25.569
Melard's algorithm was used for estimation.
35.301
-.724
.501
Non-Seasonal
Lags
Seasonal Lags
AR1
AR2
MA1
Seasonal AR1
Constant
Correlation Matrix
Seasonal
Lags
Non-Seasonal Lags
Non-Seasonal
Lags
Seasonal Lags
AR1
AR2
MA1
Seasonal AR1
AR1
1.000
-.044
.731
AR2
-.044
1.000
-.713
MA1
.731
-.713
1.000
Seasonal
AR1
.398
-.365
.523
Constant
0(a)
0(a)
0(a)
.398
-.365
.523
1.000
0(a)
Constant
0(a)
0(a)
0(a)
0(a)
1.000
Melard's algorithm was used for estimation.
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.
Covariance Matrix
Seasonal
Lags
Non-Seasonal Lags
Non-Seasonal
Lags
Seasonal Lags
Constant
AR1
AR2
MA1
Seasonal AR1
AR1
.004
.000
.148
AR2
.000
.003
-.132
MA1
.148
-.132
9.815
Seasonal
AR1
.014
-.011
.859
Constant
0(a)
0(a)
0(a)
.014
-.011
.859
.275
0(a)
0(a)
0(a)
0(a)
0(a)
1246.182
Melard's algorithm was used for estimation.
a The ARMA parameter estimate and the regression parameter estimate are asymptotically
uncorrelated.