CORRESPONDENCE ANALYSIS

Presented by

Iqra Siddiqui Farheen Ahmed Quidwai M.Shoaib Farooq Hina Razi Maria Qayyum

Correspondence Analysis Overview

What is it? Why use it?

HISTORY OF CORRESPONDENCE ANALYSIS

These methods were originally developed

primarily in France by Jean-Paul Benzrci in the early 1960's and 1970's.

similar techniques were developed

independently in several countries, where they were known as optimal scaling, reciprocal averaging, optimal scoring, quantification method, or homogeneity analysis).

WHAT IS CORRESPONDENCE ANALYSIS?

Correspondence Analysis is a technique that generates graphical representations of the interactions between object (or "categories") of two categorical variables. Correspondence analysis is a related perceptual mapping technique with similar objectives. Perceptual mapping: it is a set of techniques that attempt to identify the perceived relative image of a set of objects.

EXAMPLE
DETERGENT CATEGORY
INCOME CLASSES UPPER CLASS MIDDLE CLASS LOWER CLASS COLUMN TOTAL SURF EXCEL 40 30 5 75 EXPRESS POWER 18 50 10 78 BONUS 2 10 40 52 ROWS TOTAL 60 90 55 205

EXAMPLE
DETERGENT CATEGORY
INCOME CLASSES UPPER CLASS MIDDLE CLASS LOWER CLASS COLUMN TOTAL SURF EXCEL 40 30 5 75 EXPRESS POWER 18 50 10 78 BONUS 2 10 40 52 ROWS TOTAL 60 90 55 205

EXAMPLE

REASONS FOR USING CORRESPONDENCE ANALYSIS

The primary goal of correspondence analysis is to transform a table of numerical information into a graphical display.

Form Contingency table.

Perform chi-square test on table. Create metric distance measure from chi-

square association measure. Place categories on these dimensions so as to best account for associations.

ADVANTAGES OF CORRESPONDENCE ANALYSIS

The simple cross-tabulation if multiple categorical variables can be represented in a perceptual space. CA portrays not only the relationships between the rows and columns, but also the relationships between the categories of either rows or columns.

DISADVANTAGES OF CORRESPONDENCE ANALYSIS

The technique is descriptive and not at all

appropriate for hypothesis testing.

The technique is quite sensitive to outliers,

in terms of either rows or columns.

DISTINGUISHING CHARACTERISTICS
It is compositional method rather then a decompositional approach.

CA is the basis for developing perceptual maps. CA abilities for simultaneously representing rows and columns.

STAGES IN CORRESPONDENCE ANALYSIS

STAGE 1: OBJECTIVES OF CA

Association among only row or column categories. Association between both row and column categories.

EXAMPLE
DETERGENT CATEGORY
INCOME CLASSES UPPER CLASS MIDDLE CLASS LOWER CLASS COLUMN TOTAL SURF EXCEL 40 30 5 75 EXPRESS POWER 18 50 10 78 BONUS 2 10 40 52 ROWS TOTAL 60 90 55 205

EXAMPLE
DETERGENT CATEGORY
INCOME CLASSES UPPER CLASS MIDDLE CLASS LOWER CLASS COLUMN TOTAL SURF EXCEL 40 30 5 75 EXPRESS POWER 18 50 10 78 BONUS 2 10 40 52 ROWS TOTAL 60 90 55 205

EXAMPLE
DETERGENT CATEGORY
INCOME CLASSES UPPER CLASS MIDDLE CLASS LOWER CLASS COLUMN TOTAL SURF EXCEL 40 30 5 75 EXPRESS POWER 18 50 10 78 BONUS 2 10 40 52 ROWS TOTAL 60 90 55 205

STAGE 2: RESEARCH DESIGN OF CA

Requires only data matrix of non-negative entries e.g. cross tabulations

Categories in both rows and columns must have specific meaning for interpretation purpose.
Objects are rated on a set of characteristics (e.g: attributes).

STAGE 3: ASSUMPTIONS OF CA
Homogeneity:

In correspondence analysis, it is assumed that there is homogeneity between the column variable of the analysis. If homogeneity is not present in the analysis, then the result will be misleading.

STAGE 3: ASSUMPTIONS OF CA
Distributional assumption:

Correspondence analysis is a nonparametric technique that assumes distributional assumptions.

Category assumption: In correspondence

analysis, it is assumed that the discrete data has many categories.

STAGE 3: ASSUMPTIONS OF CA
Negative values: In correspondence analysis,

negative value is not considered. Continuous data: In correspondence analysis, discrete data is used. If we are using continuous data, then the data must be categorized into range. Correspondence analysis is an exploratory technique not a confirmatory technique.

STAGE 4: DERIVING CA RESULTS AND ASSESSING FIT

a. Deriving results

Chi-square values obtained for each cell.

Chi-squared values standardized and

converted to distance measure.

STAGE 4: DERIVING CA RESULTS AND ASSESSING FIT

b. Assessing fit Identify appropriate number of dimension. Assess importance of each dimension by

their Eigen values.

STAGE 5: INTERPRETATION OF THE RESULTS

Identify a categorys association with other categories. Determine whether comparisons are to be made between row or column, row and column categories.

STAGE 6: VALIDATION OF THE RESULTS

Sample objects

RESEARCH PAPER
Abstract:

A survey was conducted at the Faculty of Economics of the Budapest Tech to measure and evaluate 17 technical skills required by employers. The study utilized questionnaires to rate and rank these skills based on student assessment. The research was supported by a multivariate statistical method referred to as correspondence analysis.

RESEARCH PAPER
Introduction:
Higher education needs to be aware of the

changing nature of the workplace and of the requirements of employing organizations. The main purpose of this study is to find the technical skills and the major areas of knowledge that are sought by employers for our industrial engineering and business management students entering the job market.

RESEARCH PAPER
Technical skills :
A = Word processing (Word, etc.) B = Spreadsheets (Excel, etc.) C = Databases D = Operating systems design E = Project management F = Computer software and programming languages G = Inventory management H = Logistics (transportation, distribution, warehousing, suppliers)

RESEARCH PAPER
Technical skills :
I = Quality management J = Resource planning & control K = Web designing, IT, Internet operations L = Telecommunication M = Quantitative analiysis (statistics, optimization, etc.) N = Managerial accounting (budgeting, break-even, cost controlling) O = Finance (balance sheet, cost-benefit, cash-flow, investments) P = Marketing & market research (sales, behaviour, etc.) Q = Entrepreneurship

RESEARCH PAPER
Survey method :

Students were asked to rate from 1 to 7 on a seven point ascending scale these 17 technical skills according to their importance. These competencies are required on entering employment and also for their professional career.

RESEARCH PAPER
Survey method :

1 2 3 4 5 6 7

Not at all important Scarcely important Slightly important Moderately important Usually important Significantly important Extremely important

RESEARCH PAPER
Sample size:

Data is collected from every type of student in Faculty of Economics of the Budapest and the sample size was 242.

RESEARCH PAPER
Contingency Table :

RESEARCH PAPER
Significance :
In terms of the significance of dependencies the value of the chi-square statistic is 2= 925.794 , which at a stated level of =0.05 indicates a significant dependency between the rows (attributes) and the columns (contributions) (p= 0.000).

RESEARCH PAPER
Result :

RESEARCH PAPER
Conclusion :
Efforts must be made by the school to

minimize any gaps between our engineering and economics students perceptions of marketable skills and actual skills expected by employers. The results of this study furnish a good basis for supervising and further developing the present curricula into this direction.