statistical packages

Upload: palaniappan-sellappan

Post on 03-Apr-2018

218 views

Category:

Documents


0 download

TRANSCRIPT

  • 7/28/2019 Statistical Packages

    1/32

    Statistical Packages

  • 7/28/2019 Statistical Packages

    2/32

    SPSS

  • 7/28/2019 Statistical Packages

    3/32

    Statistical Package for Social Science

    Developed as a programming language forconducting statistical analysis

    Uses graphical and syntactical interface

    Can perform simple percentage to complexanalysis

    Can handle simple data, binary data,

    logarithmic data etc., Has powerful data management function and

    programming language

  • 7/28/2019 Statistical Packages

    4/32

    History

    Multinational software company

    In 1968 SPSS Statistical software system was

    Developed by H.Nie, C.Hadlai Hull and Dale

    bent, Three Stanford University graduated

    In 1975 SPSS incorporated with Headquarter

    at Chicago

    In 1984 SPSS/PC+ Introduced for MS-DOS

  • 7/28/2019 Statistical Packages

    5/32

    Uses

    Widely used by

    Market research

    Health research

    Survey

    Educational research

    Government

    Marketing organisation

    Others

  • 7/28/2019 Statistical Packages

    6/32

    Modules

    Follows module system

    SPSS Programmability extension

    SPSS Data validation

    SPSS Regression module

    SPSS advanced module SPSS Classification trees

    SPSS Tables

    SPSS exact tests

    SPSS Trends SPSS Cojoints

    SPSS complex samples

  • 7/28/2019 Statistical Packages

    7/32

    Features

    Accessible via Pull down menu & Programmablewith 4GL command syntax language

    Complex applications can only be programmed

    Pull down menu will also generate syntax Macro language can be used to write command

    language subroutines

    Python programmability extension can access theinformation in the data dictionary and data anddynamically build command syntax programs

  • 7/28/2019 Statistical Packages

    8/32

    Data

    Places constraints on internal file structure

    Datasets has 2dimesional table structure

    Rows represents cases Columns represent measurement

    Numeric and string can be entered

    Data processing occurs sequentially

  • 7/28/2019 Statistical Packages

    9/32

  • 7/28/2019 Statistical Packages

    10/32

    Output

    Can be read and write to external data base

    tables through ODBC and SQL

    Output in proprietary file format (*.SPV file)

    Can be read by stand alone reader

    Can be exported to text or ms word

    Can be captured as PDF, XLS, HTML, XML

  • 7/28/2019 Statistical Packages

    11/32

    Analysis

    Can do

    Descriptive statistics

    Bivariate statistics

    Linear regression

    Identifying Groups

  • 7/28/2019 Statistical Packages

    12/32

    Strengths

    Ease of use

    User friendly

    Covers most routine analysis Presentation quality

    University site license

  • 7/28/2019 Statistical Packages

    13/32

    Weakness

    Positioning graphics

    Speed differs

    Too easy

    Learning curve

    Lacks many Regression analysis

    Difficult to edit output

    Problems with importing and exporting data

    Need to know statistical terminology

  • 7/28/2019 Statistical Packages

    14/32

    Minitab

  • 7/28/2019 Statistical Packages

    15/32

    History

    Developed by Pennsylvania state university teamof researchers in 1972

    Began as a light version of OMNITAB, a statisticalprogram of NIST

    Originally developed to assist faculties to teachstatistics

    Second most popular statistical package

    Used in teaching and research Used in Six sigma, CMMI and in other statistical

    based process improvement methods

  • 7/28/2019 Statistical Packages

    16/32

    Features

    Available in DOS, Macintosh, UNIX, Windows etc.,

    Student edition is distributed internationally

    Ease to learn and ease to use

    Has pull down menus and dialog box

    Input- through minitab data window or imported

    Macros allow creation of custom operation

    Ease transition from macintosh to windows

  • 7/28/2019 Statistical Packages

    17/32

    Analysis

    Can do

    Descriptive statistics, test of association, test

    of difference, control charts, tables etc.,

    Histogram, box plots, regression plots,

    scattegrams, bar charts

  • 7/28/2019 Statistical Packages

    18/32

    SAS

  • 7/28/2019 Statistical Packages

    19/32

    SAS

    Statistical analysis system

    Provided by SAS Institute Inc.,

    Can do Data entry, retrieval, management and mining

    Statistical analysis Report writing and graphics

    Business planning, forecasting and decision support

    Operational research and project management

    Quality improvement

    Applications development Data warehousing

    Remote computing

  • 7/28/2019 Statistical Packages

    20/32

    Used in

    IT management

    Human resource management

    Customer relationship management

    Business intelligence

    Manufacturing

    Telecommunication

    Government

    Environmental Research

    Biotechnology

    Retail

  • 7/28/2019 Statistical Packages

    21/32

    Features

    GUI exist as front end to facilitate generaton

    of SAS program

    SAS components are used through

    programming interface

    SAS Programs- consist of Data step, Procedure

    step, macro language

    Allows sql programming

    Macro programming extension available

  • 7/28/2019 Statistical Packages

    22/32

    Read and write many different file formats

    Process data in many formats

    4GL programming language

    Built-in statistical and random number function

    Comprehensive data and time handling function

    SAS ACCESS module allows communication with

    data through ODBC Can process files containing millions of rows

    Integrated development environment

  • 7/28/2019 Statistical Packages

    23/32

    STATA

  • 7/28/2019 Statistical Packages

    24/32

    Stata

    General purpose statistical software package

    Created by Stata Corp in 1985

    Sold as a complete package, not in modules

    Used by business and academic world Capabilities include

    Data management

    Statistical analysis

    Graphics Simulations

    Custom programming

  • 7/28/2019 Statistical Packages

    25/32

    Features

    Fast, accurate and easy to use

    Broad suite of statistical capabilities

    Complete data management facilities

    Publication-quality graphics

    Responsive and extensible

    Matrix programmingMata is integral part

    Cross platform compatibility

    Complete documentations

  • 7/28/2019 Statistical Packages

    26/32

    S Plus

  • 7/28/2019 Statistical Packages

    27/32

    S-Plus

    For professional statisticians and data analyst

    GUI pull down menus and dialog box

    Extensive programming through Object oriented

    S language S Language developed by researchers from AT& T

    Bell labs

    Flexibility in user defined functions

    Customization of environment

    Extensive graphics and hard copy capability

    Runs on windows and Unix based platform

  • 7/28/2019 Statistical Packages

    28/32

    Can generate

    Random data from 20 different distributionaltypes

    12 different hypothesis testing Including Students t test, wilcoxin test etc.,

    Linear, non-linear project pursuit regression

    Multivariate analysis

    Including graphical methods, cluster analysis,discriminant analysis

    Time series analysis

  • 7/28/2019 Statistical Packages

    29/32

    MATLAB

  • 7/28/2019 Statistical Packages

    30/32

    MATLAB

    Interactive computing environment

    For scientific and statistical data analysis andvisualization

    Basic objectMatrix Numerical analysis like signal processing, image

    processing and statistical analysis can beperformed on matrix

    No programming Different Version available for UNIX and PC

    platform

  • 7/28/2019 Statistical Packages

    31/32

    Features

    Functions are written in M-file function

    M-files can be modified and programmed

    Recent addition MATLAB C compiler and C-Mathlibrary

    Toolbox collection of m-file functions forimplementing algorithms and functions in an specificarea of interest

    2D and 3D visualization of datamesh plot, contour

    plot, image plot, histograms Has graphical objectsallowing users to create GUIs in

    algorithm packages

  • 7/28/2019 Statistical Packages

    32/32