 |
||||||||||||||
 |
 |
 |
 |
|
 |
|
 |
|
 |
|
 |
|
 |
 |
 |
 |
 |
 |
|
 |
|
 |
|
 |
|
 |
|
 |
 |
|
|
||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||
|
 |
As a service to our customers, QMS posts current shipping versions of key files for all of our software, as well as assorted data files. You may update your copy of EViews 4.0 to the latest shipping version by downloading and running the EViews 4.0 update executable. Note that the update installer will only work if you have a previously installed copy of EViews 4.0 on your system. 11-14-2001 (update installer) Remove limits on the number of series that can be graphed in a single line or pie graph. 11-02-2001 Minor changes to date labeling of graphs, and model options. 10-19-2001 Improve handing of math errors when opening workfiles containing groups with expressions. In some circumstances,
EViews would generate an error and fail when attempting to open a workfile with large numbers of these errors. 10-16-2001 A bug in the t-distribution was introduced in the 10-11 EViews update. If you downloaded this update or your copy of EViews is dated 10-11, please be sure to update to a newer version. We apologize for the inconvenience. 10-11-2001 Significant speed improvements in floating point operations. Speed improvements in the calculation of the t
and F cumulative distribution functions for certain combinations of parameters. Fix right-mouse-click open of multiple
series windows. 08-29-2001 Fix labeling of axes when performing scatterplots with fit. The labels were removing relevant lag information.
Also fix bug in sample processing for the scatterplot fits. 07-26-2001 Enables the "Propagate NAs" frequency conversion option for Haver Databases. This update is only relevant
for Enterprise Edition versions of EViews. 07-25-2001 Improve memory usage when allocating large numbers of small objects. Fix reported problems with maintaining settings on graphs that were saved to disk without ever being displayed. Improve handling of system iteration settings when switching between command line and dialog estimation. Fix problem when exporting a subsample of gradients from count models. 07-10-2001 Fix parsing of some of the new VAR command line procs/views to correctly handle expressions containing the "@"
character. Additional error checking when running VARs on large numbers of series. Fix initialization of binary
residual exporting for out-of-sample observations. Improvements in the efficiency of string memory allocation. 06-29-2001 Fix labeling of tickmarks in reverse scale graphs. Minor improvements to the analytic mathematics engine. Improvements
to the exception.rc errorhandling. 06-22-2001 Fix problem with hfetch leaving open file handles. Note that hfetch is not the preferred method of working with
Haver databases. Extend the use of the @resample and @permute functions to work with vectors and coefficients. 06-08-2001 Fix minor problem with the dialog for selecting VAR structural factorizations. Some fixes involving empty group
objects. 06-06-2001 Fix incompatibility with some foreign language enabled versions of Windows. This fix has been tested with the
Hebrew enabled version of Windows 9x/Me and should solve similar problems observed when using the Arabic version
of Windows. 06-05-2001 Corrected posting for fix as described below for 5-31-2001. 05-31-2001 Fix problems in display of the Actual/Fitted/Residual graph in the presence of noncontiguous samples. Fix problem with processing sample if statements in random effects pool estimation. 05-10-2001 Improve text labeling of estimation methods when estimating systems. 05-08-2001 Improve use of pre-sample information when performing in-sample forecasting and model simulation with ARMA errors. This improvement means that in some circumstances forecasts can be made for earlier periods since the initialization (run-in) period has been shortened. 05-03-2001 Enable font selection for legends in scatterplots. Additional improvements in the code which should improve stability when printing group views. 05-02-2001 Add additional error checking for samples used in model stochastic simulation. 05-01-2001 Fix crash in stochastic model simulation that occurred when using certain samples to compute the residual covariance. Fix system estimation dialog so that it correctly reads the option for BHHH estimation. 04-16-2001 Minor fix to improve handling of multiple untitled workfiles when performing a workfile extraction (subsetting). 04-13-2001 Improve ARCH estimator robustness in regions where the process is nearly explosive and where the log-likelihood is unstable. 04-03-2001 Fix problem where, when estimated via the dialog, system estimation always used the global default settings. 03-30-2001 Improve the algorithm for displaying log scaling in graphs. Other minor improvements to the graphics engine. 03-23-2001 Add features to registration system. Fix problem that can sometimes occur when setting up samples with missing values in Pool estimation. While this problem did not affect estimation results, for complicated sample structures, it could lead to minor errors in the residual covariance matrix view. 03-07-2001 Fix interpreter problem when using the store command with the "::" syntax to specify the database name. Fix problem that sometimes occurred when you copy-and-paste text containing specifications into the least squares dialog. The existence of extraneous linefeeds in the copied text could cause problems with correctly interpreting the specification, leading to misleading results. Improve labeling of daily data in graph displays for years beyond 2000. The default setting for GMM estimation from the program line was changed to HAC to match the documentation. Improve ARCH estimation output when estimated using an expression for the mean equation, and when displaying estimation settings. Fix problem with the rndseed command not always correctly setting the random number seed. Two fixes for incorrect error messages: one occurring while checking data for scaling using Historical X11, and one sometimes occurring when assigning the result of a square-root to a scalar or matrix object. Enable merging of graphs using the MERGE proc. Also add the ability for users to set a default text font to be used when adding to a graph. Improve options handling for systems. Enable iteration control for single equation GMM. Improve copy and paste of groups between workfiles with better handling of auto-series, and retention of dated data table options. Fix to handling of the directory paths for X12 automatic model selection file. The dialog options for the GMM kernel type were reversed. There is a syntax issue associated with how Windows handles double- clicking on files (with spaces in the full path) that created some incompatibilities with previous EViews syntax. As a result, double-clicking conflicted with the method of launching EViews from a command window while automatically running a program file. To launch EViews from a command window, enclose the name of the executable in double quotes (if necessary), and enclose the name of the program name in double quotes (if necessary), followed by any arguments. Reduce memory use in pool spreadsheet view for pools with large numbers of cross-section units. Fix tracking of unnamed groups when performing principal components analysis. Fix the default dialog settings in the File I/O dialogs. The build date on the last update of EViews did not properly reflect the new version so that it appeared to be dated 01-26. Improve handling of named samples in dialogs. Fixes a bug introduced on 01-25-2001 in which closing all previously opened workiles could lead to instability in the program when performing program, database or workfile operations with no open workfiles. We strongly recommend that all users who have a copy of EViews built on 01-25 or 01-26 upgrade their copy as soon as convenient. Fix bug in command line option processing for exponential smoothing. Improve handling of blank lines in RTF copy-and-paste. Provide better handling of multiple untitled workfiles and placement of dates in graphs. Fix of out-of-sample initializations in forecasting and fit commands, and out-of-sample static ARCH forecasts. Add ability to initialize excluded variables in model using aliased series. Provide additional copy-and-paste and file saving options for metafiles. Fix problem with output from serial correlation LM test. The test statistic was reported correctly, but the auxiliary regression output had misplaced coefficient labels. Fix problem with dialog for equation forecasting. The dialog incorrectly stripped out whitespace from the sample
statement. As a result, the forecast procedure generated incorrect error messages. |
|
|
[ EViews 홈 | EViews 5 | 제품 정보 | 가격 | 문의 | 다운로드 | 제품/서적 | EViews 4.1 ] |
| 필 사이언스 홈 | 소프트웨어 | 한글 홈페이지 | 기술지원 | 회사 | 주문 | 질의/응답 |
| 경기도 고양시 일산동구 백석동 1141-2 전화 (031) 905-7754 팩스 (031) 905-7724 Copyright ⓒ 2000~2008 Phil Science, Inc. All rights reserved. E-mail: |
