1. Trang chủ
  2. » Tài Chính - Ngân Hàng

SAS/ETS 9.22 User''''s Guide 298 ppt

10 54 0

Đang tải... (xem toàn văn)

THÔNG TIN TÀI LIỆU

Thông tin cơ bản

Định dạng
Số trang 10
Dung lượng 675,2 KB

Nội dung

2962 ✦ Chapter 47: SAS/ETS Model Editor Window Reference Model Summary of Residual Errors Window This window displays the residual error analysis for each model equation. To open this window, right-click an existing fitted model in the SAS libraries panel and select Residual Errors. Figure 47.29 Model Summary of Residual Errors Window Actual v/s Predicted Plot Window This window displays the actual data set and the predicted data set for viewing and analyzing the model prediction power in graphical plots. To open this window, right-click an existing fitted model in the SAS libraries panel and select Actual v/s Predicted Plot. Figure 47.30 Actual v/s Predicted Plot Window Open Existing Model ✦ 2963 Edit Existing Fitted Model Open Existing Model In the SAS/ETS Model Editor window, click the Sources tab. In the Display Preference list box, select All Models to display all of the predefined models. If necessary, expand items in the Name box. Right-click the model name and select Open . The model window opens. The title of the window corresponds to the structure where the model is located. The model window contains a pane on the left with values such as Variables , Equations , Constraints , and Fit . The contents of the right pane change depending on the selection made in the left pane. Figure 47.31 Open Existing Model The model window always contains the following controls and fields: Details: tab displays details for the selection highlighted in the left pane. 2964 ✦ Chapter 47: SAS/ETS Model Editor Window Reference SAS Code provides the SAS statements for this window. Check Syntax displays any error in the MODEL procedure syntax. OK runs the model and saves the results. Cancel closes the window without implementing any changes. Fit Model—Equations To select which equations to fit, select Equations under Fit in the left pane of the model window. In the right pane, select an entry from the Equations Available to Fit: list and click the right-arrow button . The equation is moved to Equations to Fit: list. Figure 47.32 Define Equations This right pane has the following controls and fields: Equation Available to Fit: list lists equations that are available to be fit. If the list of equations is blank, all model equations that contain parameters are fitted. Equations To Fit: list Fit Model—Input ✦ 2965 displays the equations selected to be fit. To delete an equation, select an existing equation and then click the Delete Row button . Fit Model—Input To specify input data, select Input under Fit in the left pane of the model window. The right pane then displays fields where you can select the library name and data set name, map program symbols to data set variables, and define covariance matrix of equation errors. Figure 47.33 Input Data Set This right pane has the following controls and fields: Library Name: is the name of the library where the current model is located. Data Set Name: is the name of the input data set against which you want the model fitted. Browse: opens the data set selection window for selecting an input data set. ID: specifies a single variable to identify observations in error messages or in other listings and in the output data set. The ID variables are usually SAS date or datetime variables. You can type any valid SAS variable name in this field or select from the list. 2966 ✦ Chapter 47: SAS/ETS Model Editor Window Reference Frequency: is the time interval (data frequency) for the input data set. Specify specifies which frequency in the list is to be used in the time series model. When you click Specify , the Specify Frequency window opens. See the section “Specify Frequency Window” on page 2938 for more information. Sort By: indicates how the fit options are sorted. This field has two list boxes. In the left list box, select the variable by which to sort. In the right list box, select Ascending , Descending , or Not Sorted. Weight: specifies a variable to supply a weighting value to use for each observation in estimating parameters. Missing Value Behavior: specifies whether missing values are tracked on an equation-by-equation basis or the entire observation is omitted from the analysis when any equation has a missing predicted or actual value for the equation. Program Symbols: table lists the name, type, and label of the variables in the existing model. The type of the variables include endogenous, exogenous, and variable. Data Set: table lists the name and label of the variables in the data set. You can identify the type of each variable by mapping program symbols to data set variables. New Mapping button creates a new mapping between a model variable selected in the Program Symbols table and a data set variable selected in the Data Set table. This button becomes available only when a model variable and a data set variable are selected. To create a mapping, click , then click an entry in the Program Symbols: table and drag the pointer to an entry in the Data Set: table. Delete button deletes the selected mapping. Evaluate button evaluates data. Automatic Mapping button automatically maps program variables to data set parameters by using their names. Click this button to map variables between the Program Symbols and Data Set tables by their variable names. If the Program Symbols table contains a variable whose name matches a variable in the Data Set table, these variables are automatically mapped to each other. Covariance Matrix of Equation Errors specifies the name of the input data set and the name of the library where it is stroed. Fit Model—Iteration ✦ 2967 Fit Model—Method To specify how the model parameters are to be estimated, select Method under Fit in the left pane of the model window. For more information about available methods of parameters estimation, see Chapter 18, “The MODEL Procedure,” (SAS/ETS User’s Guide). Figure 47.34 Model Fit Method This right pane has the following controls and fields: Estimation Methods check boxes enable you to select the corresponding estimation method. The check boxes correspond to the methods implemented in PROC MODEL, all of which aim to minimize an objective function. Divisor of Variance specifies a degrees-of-freedom correction in estimating the variance matrix. This could be the sum of the weights or the sum of the weights minus the model degrees of freedom. Fit Model—Iteration To specify the parameters for minimizing the objective function, select Iteration under Fit in the left pane of the model window. 2968 ✦ Chapter 47: SAS/ETS Model Editor Window Reference Figure 47.35 Model Estimation Iteration This right pane has the following controls and fields: Method: options specifies the Gauss method or Marquardt method. Gauss is the default. For the Gauss method, the Gauss-Newton parameter-change vector for a system is computed at the end of each iteration. The objective function is then computed at the changed parameter values at the start of the next iteration. For the Marquardt method, at each iteration, the objective function is evaluated at the parameters changed by the Marquardt-Levenberg parameter-change vector. For more information about available methods of parameter estimation, see Chapter 18, “The MODEL Procedure” (SAS/ETS User’s Guide). Fit Model—Tests ✦ 2969 Maximum Iterations: specifies the maximum number of Newton iterations performed at each observation and each replication of Monte Carlo simulations. Maximum number of step halvings: specifies the maximum number of subiterations allowed for an iteration. For the Gauss method, this value limits the number of step halvings. For the Marquardt method, this value limits the number of times step parameter  can be increased. The default is MAXSUBITER=30. See Chapter 18, “The MODEL Procedure” (SAS/ETS User’s Guide), for more information. Minimization Tuning specifies the Convergence Criteria and Singularity Criteria. Fit Model—Tests The SAS/ETS Model Editor enables you to perform tests of nonlinear hypotheses on the model parameters. To define statistical tests that are associated with the model parameters, select Tests under Fit in the left pane of the model window. Any previously defined parameter tests for the model are listed. Figure 47.36 Fit Tests This right pane has the following controls and fields: 2970 ✦ Chapter 47: SAS/ETS Model Editor Window Reference Tests lists tests of nonlinear hypotheses to be performed on the model parameters. Test expressions can be composed of parameter names, arithmetic operators, functions, and constants. To add a test, click the Add button and enter the parameters for the test. To delete a test, select the test and then click the Delete button . Fit Model—Results To specify which results you want and to save them to an output data set, select Results under Fit in the left pane of the model window. The right pane then provides fields where you can specify which results you want to save and where you can specify the library and data set name where the output is saved. Figure 47.37 Output Selection This right pane has the following controls and fields: Output specifies the results from your model fitting that are to be displayed. Select the check box for each analysis you want to include in the output data set or select All to display all analyses for each model equation. Define the Model Parameters and Variables ✦ 2971 Estimated Covariance of the Equation Errors specifies that the cross-equation covariance are to be displayed and also saved in the output data set specified in Data Set Name: in the folder specified in Library Name: . In addition, the determinant of the matrices is displayed. Parameter Estimates specifies that the parameter estimates are to be displayed and also saved in the output data set specified in Data Set Name: in the folder specified in Library Name: . Select Include Covariance Matrix of Estimates to also display and save the estimated covariance and correlations matrix. Library Name: is the name of the library where the model fit results are to be saved. Data Set Name: is the data set in which you want the model fit results to be saved . Browse: opens the data set selection window for selecting an output data set. Define the Model Parameters and Variables To select input variables for use by the SAS/ETS Model Editor and add them to the Variable and Parameter Definitions list, select Parameters and Variables in the left pane of the model window. Figure 47.38 Create New Model Template—Equations . available methods of parameter estimation, see Chapter 18, “The MODEL Procedure” (SAS/ETS User’s Guide) . Fit Model—Tests ✦ 296 9 Maximum Iterations: specifies the maximum number of Newton iterations performed. “The MODEL Procedure” (SAS/ETS User’s Guide) , for more information. Minimization Tuning specifies the Convergence Criteria and Singularity Criteria. Fit Model—Tests The SAS/ETS Model Editor enables. fields: Details: tab displays details for the selection highlighted in the left pane. 296 4 ✦ Chapter 47: SAS/ETS Model Editor Window Reference SAS Code provides the SAS statements for this window. Check

Ngày đăng: 02/07/2014, 15:20

TỪ KHÓA LIÊN QUAN