EViews 5 새로운 점 - 필 사이언스(주)

QMS is pleased to announce the eagerly anticipated release of EViews 5. Today QMS continues its tradition of innovation by blending the best of modern spreadsheet and relational database technology with the traditional tasks of statistical software. The result is a program that provides unprecedented power within the familiar easy to use EViews interface.

Whether you work with cross-section data, panel data, irregularly dated financial data, or simple time series data, EViews 5 simplifies the process of managing your data. Need to merge data between multiple sources or want your data to be updated whenever you begin a new EViews session? No problem. New linking and updating technologies allow you to embed within your workfile relationships based on formula, match merges, frequency conversion, and references to external data. And if your work requires more than simple numbers, new alphanumeric and date handling features will cover your needs.

Data import and export have never been easier. If your data are currently in Excel, SPSS, SAS, Stata, Rats, TSP, or over a dozen other popular file formats, simply drag-and drop your file onto EViews and the data will automatically appear. EViews automatically detects the number of rows and columns in your dataset, determines the data type of each column, and even automatically analyzes the date structure, frequency and range of the incoming data. EViews 5 also understands HTML, making importing data from web pages a snap. And support for ODBC in the EViews Enterprise Edition makes it easy to exchange data with your company database, including products such as Oracle, Microsoft SQL Server and IBM DB2.

But EViews 5 is not simply about data management. Powerful new statistical techniques have been added to the already impressive EViews toolkit. For time series analysis, new GARCH estimators, band-pass filters, and equation diagnostic tools are now available. Panel data analysis is supported by a new panel workfile structure, with new panel unit root tests, dynamic panel data estimators, and a full compliment of fixed and random effects specifications. New functions for dealing with categorical data, including but not limited to the automatic generation of categorical dummy variables, simplify the process of analyzing cross-section and panel data.

The following list outlines some of the most important new features of EViews 5.

Workfile
General Data
Alpha series and String Support
Date series and Data Support
Panel and Pool Data
Garch Estimation
Statistics and Econometric
Graph and Tables
Speed

Workfile

Multi-page workfiles.

Support for complex data structures including irregular dated data, cross-section data with observation identifiers, dated and undated panel data.

Match merge, join, append, subset, resize, sort, and reshape (stack and unstack) workfiles.

Convert data between EViews and various spreadsheet, statistical, and database formats, including: Microsoft Access files, Gauss Dataset files, ODBC Dsn files (Enterprise Version), ODBC Query files (Enterprise Version), SAS Transport files, native SPSS files, SPSS Portable files, Stata files, Excel files, raw ASCII text or binary files, HTML, or ODBC Databases and queries (ODBC support is provided only the Enterprise Edition). (Access to SAS native format files is available using SAS ODBC drivers which must be purchase separately from SAS.)

Drag-and-drop support for reading data; simply drop files into EViews for automatic conversion of foreign data into EViews workfile format.

^Top

General Data

Alphanumeric (string) series, with an extensive library of string manipulation functions.

Date series, with an extensive library of date manipulation functions.

Dynamic frequency conversion and match merging using link objects. Frequency conversion and match merge links will be updated whenever the underlying data change.

Auto-updating series that depend upon a formula are automatically recalculated whenever the underlying data change.

Value labels (e.g., "High", "Med", "Low", corresponding to 2, 1, 0) may be used with numeric and alpha series. Function support allows you to work with either the underlying or the mapped values.

Improved sample object processing allows for the direct use of sample objects in series expressions. In addition, set operators may now be used with sample objects, allowing you to form samples using the operators AND, OR, and NOT.

New functions facilitate assigning values from the computation of by-group descriptive statistics to individual observations.

Automatic creation of sets of dummy variables for use in estimation.

^Top

Alpha Series and String Support

New library of string functions and operators.

Additional functions for converting between string representations of dates and EViews numeric date values.

^Top

Date series and Date support

Full support for calendar dates with extensive library of functions for manipulating dates and EViews numeric date values.

Added functions for converting between EViews numeric date values and string or numeric representations of dates.

^Top

Panel and Pool data

General

Workfile tools reshape data to and from panel (stacked) and pool (unstacked) workfile structures.

Panel unit root tests: Levin-Lin-Chu, Breitung, Im-Pesaran-Shin, Fisher-type tests using ADF and PP tests (Maddala-Wu, Choi), Hadri.

Linear equation estimation with additive cross-section and period effects (fixed or random). Two-way random and mixed effects models supported for balanced data only, most others for both balanced and unbalanced data.

Quadratic unbiased estimators (QUEs) for component variances in random effects models: Swamy-Arora, Wallace-Hussain, Wansbeek-Kapteyn.

Generalized least squares for models with cross-section or period heteroskedastic and correlated specifications. Support for both one-step and iterative weighting.

Two-stage least squares (2SLS) / Instrumental variables (IV) estimation with cross-section and period fixed or random effects. Generalized 2SLS/IV estimation of GLS specifications.

Most specifications support estimation with AR errors using nonlinear least squares on the transformed specification.

Robust standard error calculations including seven types of robust White and Panel-corrected standard errors (PCSE).

Panel Specific

Structured workfiles support large cross-section panels.

Panel data graphs. Various plots by cross-section in multiple graphs or combined. Graphs of summary values across cross-section.

Nonlinear equation estimation with additive effects.

GMM estimation for models with cross-section or period heteroskedastic and correlated specifications. Support for both one-step and iterative weighting.

Linear dynamic panel data estimation using first differences or orthogonal deviations, with period specific instruments (Arellano-Bond one-step, one-step robust, two-step, iterated). Flexible specification of instrument lists.

Pool Specific

Define groups of cross-sections for dummy variable processing.

Support for period specific effects, coefficients, instruments and weights.

^Top

Garch Estimation

Student's t and Generalized Error Distribution support with optional fixed distribution parameter.

More flexible EGARCH and TARCH specifications allow for estimation of a wider range of econometric models.

Power ARCH specifications with optional fixed power parameter.

^Top

Statistics/Econometrics

Confidence ellipses showing the joint confidence region of any two functions of estimated parameters from an EViews estimation object.

ARMA equation diagnostics. Display the inverse roots of the AR and/or MA characteristic polynomial; compare the theoretical (estimated) autocorrelation pattern with the actual correlation pattern for the structure residuals; display the ARMA impulse response to an innovation shock.

Band-pass (frequency) filters for a series object. EViews can compute the Baxter-King, Christiano-Fitzgerald fixed length, and the Christiano-Fitzgerald asymmetric full sample filters.

Cointegration tests now use the MacKinnon-Haug-Michelis critical values and p-values. Output displays the critical value for any user-specified size.

Bivariate normal density and cumulative distribution function evaluation support is now provided.

^Top

Graph and Tables

Filled area graphs.

Boxplots.

Formatted spreadsheet display allows you to customize the display of your series and group data values.

Enhanced table customization with control over font face, font size and color, cell background color, and borders, with cell merging and annotation.

Improved interactive and program interface for working with tables. Selecting cells, resizing columns, and changing numeric and other display formats should be much more straightforward and intuitive.

Write graphs as PostScript files. Improved Windows Metafile support now comes with control over output sizing.

Tables may be written to HTML and RTF files.

^Top

Speed

Greatly improved speed of operation.

^Top