Tuesday, 6 September 2011


Rather than discussing about the various methods taught in class, I thought today I’ll review some of the facts related to regression; which is the basic knowledge required to draw conclusions from the output of the “Discriminant Analysis”.

Regression analysis is used to produce an equation that will predict a dependent variable using one or more independent variables. This equation has the form

Y = b1X1 + b2X2 + ... + A

where Y is the dependent variable being predicted;

X1, X2 and so on are the independent variables being used to predict it;

b1, b2 and so on are the coefficients or multipliers that describe the size of the effect the independent variables are having on the dependent variable Y, and

A is the value Y is predicted to have when all the independent variables are equal to zero.

Suppose we have a regression equation for the dependent variable,

PRICE = -294.1955 (mpg) + 1767.292 (foreign) + 11905.42; telling us that price is predicted to increase by 1767.292 when the foreign variable goes up by one, decrease by 294.1955 when mpg goes up by one, and is predicted to be 11905.42 when both mpg and foreign are zero.

Coming up with a prediction equation like this is useful, only if the independent variables in the dataset have some correlation with the dependent variable. So in addition to the prediction components of our equation--the coefficients on our independent variables (betas) and the constant (alpha)--we need some measure to tell us how strongly each independent variable is associated with our dependent variable.

When running the regression model, we are trying to discover whether the coefficients on our independent variables are really different from 0 (so the independent variables are having a genuine effect on our dependent variable) or if alternatively any apparent differences from 0 are just due to random chance. The null (default) hypothesis is always that each independent variable is having absolutely no effect (has a coefficient of 0) and we are looking for a reason to reject this theory.


In simple or multiple linear regression, the size of the coefficient for each independent variable gives the magnitude of the effect that variable is having on the dependent variable, and the sign on the coefficient (positive or negative) gives the direction of the effect. In regression with a single independent variable, the coefficient tells how much the dependent variable is expected to increase (if the coefficient is positive) or decrease (if the coefficient is negative) when that independent variable increases by one. In regression with multiple independent variables, the coefficient tells how much the dependent variable is expected to increase when that independent variable increases by one, holding all the other independent variables constant. We would have to keep in mind the units which our variables are measured in.

R-Squared and overall significance of the regression

The R-squared of the regression is the fraction of the variation in the dependent variable that is accounted for (or predicted by) the independent variables. (In regression with a single independent variable, it is the same as the square of the correlation between your dependent and independent variable.)

With this blog, I hope to have covered some of the intrinsic underlying facts behind the regression model and its analysis.

Posted by-


Operations Group 2

Discriminant Analysis – Application Area

Yesterday, I had written about conjoint analysis. A few days back, when Discriminant analysis was being taught in class, due to some reasons I had to skip the class. So today I thought, I would read on discriminant analysis and would write on that topic.

From what I could understand, the major application area of discriminant analysis is where a distinction between two or three set of objects, or people,, based on knowledge of some of their characteristic is required. For Ex. For the selection process of a job, the admission process of an educational programme in a college, or dividing a group of people into different classes, like buyers and non buyers. Discriminant Analysis can be, and is in fact, used by credit rating agencies to rate the credit risk. In fact being students of finance, we had done an analysis by taking a paper where Risk analysis of SMEs was done using Discriminant Analysis and even a Risk model was developed. This technique can be very handy in deciding and classifying them into good lending risk and bad lending risk.

From what I could make out by going through various articles and blogs on the net, that discriminant analysis is very similar to the multiple regression technique. It also gives quite a similar equation as we get in a regression equation, where Y is the dependent variables and x1 and x2 are the independent variables. K1 and k2 are the unstandardised discriminant function coefficients.

To summarize, we can use linear discriminant analysis when we have to classify objects into two or more groups based on knowledge of some variables related to them. Typically these groups would be users/non-users, potentially successful salesman/potentially unsuccessful salesman, high risk/low risk customers or can be on any such similar lines.



My Choice : Factor Analysis

Today’s class covered revision of almost all the topics and techniques covered in Business Analytics course so far. It was a good learning to understand the application of different techniques appropriately to analyze the given data and to come up with strategies.

As the freedom is to us to choose any topic for blog today, I prefer to write about Factor Analysis as it is my favorite topic. Factor analysis is a good technique to reduce data/variables to a smaller number of factor or components. Let us take for example we want to examine banking preferences of customers : There will be around 30-40 variables like which would define customer preferences. However it gets time consuming to analyze so many variables. Thus, factor analysis helps to reduce these variables into dimensions in such a way that variables which have high degree of correlation come under one dimension. Thus, in this example 30-40 variables can be reduced to say around 7 or 8 variables under the label: quality, reliability, accessibility, user friendliness, etc.

With the help of SPSS, factor analysis can help us find out the correlation between different variables and also define the extent to which one variable can be extracted to form a single component of different variables. Another advantage of this technique is that correlation between groups can be found.

Once the variables are reduced to components, various other statistical tools and techniques can be applied to groups to come up with findings and solutions. In our earlier example, we can find the preferences for customers for various dimensions and thus come up with strategies.

Let us take an example where an insurance company wants to understand the responsiveness of customers with respect to product communication. It can prepare a questionnaire and collect data from a set of customers. For example: Will you buy insurance product if a celebrity endorses? , Product communication is better through brochures, product communication is better through audios, etc.

Thus, these independent variables can be grouped under various heads having high correlations. The main labels can be: Communication through flyer, communication through videos, communication through website, etc. Thus, after grouping data under various heads, we can find customer preferences of different set of customers for product communication in insurance industry.

The study can help us find out if people in rural areas prefer communication through agents? People in metro prefer communication through website, etc. These conclusions can help the company decide their product communication and marketing strategy.



The saga of Neural Networks

All of us while doing the assignments for the Business Analytics might have come across the concept of Neural Networks. Let’s have a sneak peek into the basics of Neural networks and some of its applications.

Neural networks are state-of-the-art, trainable algorithms that emulate certain major aspects in the functioning of the human brain. This gives them a unique, self-training ability, the ability to formalize unclassified information and, most importantly, the ability to make forecasts based on the historical information they have at their disposal.

Neural networks have been used increasingly in a variety of business applications, including forecasting and marketing research solutions. In some areas, such as fraud detection or risk assessment, they are the indisputable leaders. The major fields in which neural networks have found application are financial operations, enterprise planning, trading, business analytics and product maintenance. Neural networks can be applied gainfully by all kinds of traders, so if you're a trader and you haven't yet been introduced to neural networks, we'll take you through this method of technical analysis and show you how to apply it to your trading style.

Neural networks do not make any forecasts. Instead, they analyze price data and uncover opportunities. Using a neural network, you can make a trade decision based on thoroughly analyzed data, which is not necessarily the case when using traditional technical analysis methods. For a serious, thinking trader, neural networks are a next-generation tool with great potential that can detect subtle non-linear interdependencies and patterns that other methods of technical analysis are unable to uncover.

Some of the application of Neural networks are

1. Character Recognition : The idea of character recognition has become very important as handheld devices like the Palm Pilot are becoming increasingly popular. Neural networks can be used to recognize handwritten characters.

2. Image Compression – Neural networks can receive and process vast amounts of information at once, making them useful in image compression. With the Internet explosion and more sites using more images on their sites, using neural networks for image compression is worth a look.

3. Stock Market Prediction -The day-to-day business of the stock market is extremely complicated. Many factors weigh in whether a given stock will go up or down on any given day. Since neural networks can examine a lot of information quickly and sort it all out, they can be used to predict stock prices.

4. Traveling Saleman's Problem- Interestingly enough, neural networks can solve the traveling salesman problem, but only to a certain degree of approximation.

5. Medicine, Electronic Nose, Security, and Loan Applications - These are some applications that are in their proof-of-concept stage, with the acception of a neural network that will decide whether or not to grant a loan, something that has already been used more successfully than many humans.

6. Miscellaneous Applications - These are some very interesting (albeit at times a little absurd) applications of neural networks.



Author: Surya Narayana Adiga

Group : Finance 2

Various Conjoint Analysis: Which one is the best?

Conjoint analysis is one of the terms used to describe a broad range of techniques for estimating the value people place on the attributes or features that define products and services. Discrete Choice, Choice Modeling, Hierarchical Choice, Card Sorts, Tradeoff Matrices, Preference Based Conjoint and Pairwise Comparisons are some of the names used for various forms of conjoint analysis.

There are different types of conjoint analysis:

· Adaptive conjoint analysis:

· Choice based conjoint analysis:

· Discrete choice analysis:

Which Conjoint Methodology is Best?

The answer depends upon the circumstances. However In marketing the Choice-based Conjoint or Discrete Choice Modeling has become the most popular methodology over the last 4 or 5 years.

Advantages of Choice-Based Conjoint/Discrete Choice Modeling/Choice Modeling

1.) More closely resembles the decision process customers make in the market place where they look at all the alternatives available and pick the one they most prefer. It is believed, though difficult to prove, that the more closely a research task mimics real behavior the more valid and reliable the results.

2.) Allows respondent to choose "none of these." In most purchase decisions, one of the alternatives is to walk away without buying anything. Choice-based conjoint allows you to include this response in the model and account for it in the calculation of utilities.

3.) More product/service profiles are seen by each survey respondent because choice-based conjoint typically presents 3 or more alternatives in each choice set.

4.) Easier to calculate attribute interactions like price and brand. Based on aggregate level analyses, attribute interactions can be included without dramatically increasing the complexity of the research design for choice-based conjoint exercises.

Group- HR1

Author- Ankita Kanojia


Talking about our everyday take away from the class, there may be certain things that you tend to miss out on. Well, if you have missed the presentation which one of my group members had given, here it is.

The main objective of this analysis is to infer the relationship between the randomness levels in behavior of the phone users in a cellular network. While individual phone user’s calling behavior is random, some users might be more predictable than others. Being more predictable can also mean being less random. To quantify the randomness or amount of predictable structure in an individual calling pattern, the information entropy can be used. The information entropy or Shannon’s entropy is a measure of uncertainty of a random variable.

The calling pattern is understood from the Calling Time, Interconnected Time, Talk time & Location of the call. These calling patterns and the factors affecting them are studied using correlation coefficient and factor analysis. 

If a factor has a low Eigen value, then it is contributing little to the explanation of variances in the variables and may be ignored as redundant with more important factors. If a factor has Eigen Value of more than 1, then it is considered as effective component and is not redundant. Scree Plot helps in selecting the number of factors to be retained in order to account for most of the variation. The Scree Plot denotes the relationship between the factors of variance. It helps you determine which factors to retain. 
  • What we understood from the analysis is that User’s randomness level based on location has high correlation to the randomness level in time of making phone calls and vice-versa.
  • Randomness level based on user’s inter-connected time has a high correlation to the randomness level in time spent talking on each phone call
Author: Mohamed Sahle
Group: Marketing 5

Factor Analysis and Conjoint Analysis

Factor analysis is a collection of methods used to examine how underlying constructs influence the responses on a number of measured variables. There are two types of factor analysis: exploratory and confirmatory. Exploratory factor analysis (EFA) attempts to discover the nature of the constructs influencing a set of responses while Confirmatory factor analysis (CFA) tests whether a specified set of constructs is influencing responses in a predicted way.

Both types of factor analyses are based on the Common Factor Model. This model proposes that each observed response is influenced partially by underlying common factors and partially by underlying unique factors. The strength of the link between each factor and each measure varies, such that a given factor influences some measures more than others.

Conjoint analysis is a statistical technique used in market research to determine how people value different features that make up an individual product or service.

The objective of conjoint analysis is to determine what combination of a limited number of attributes is most influential on respondent choice or decision making. A controlled set of potential products or services is shown to respondents and by analyzing how they make preferences between these products, the implicit valuation of the individual elements making up the product or service can be determined. These implicit valuations can be used to create market models that estimate market share, revenue and even profitability of new designs.

In a conjoint analysis, the respondent may be asked to arrange a list of combinations of product attributes in decreasing order of preference. Once this ranking is obtained, a computer is used to find the utilities of different values of each attribute that would result in the respondent's order of preference. This method is efficient in the sense that the survey does not need to be conducted using every possible combination of attributes. The utilities can be determined using a subset of possible attribute combinations. From these results one can predict the desirability of the combinations that were not tested.

Author: Karthikeyan Dakshinamurthy (13081)

Group: Marketing - Group 4

Market Research with Conjoint Analysis

Market research is frequently concerned with finding out which characteristics of a product or service is most important to consumers. The ideal product or service, of course, would have all the best characteristics, but realistically, tradeoffs have to be made. The product with the most expensive features, for example, cannot have the lowest price.

Conjoint analysis is a technique for measuring consumer preferences about the attributes of a product or service. There are two general approaches to collecting data for conjoint analysis—the two-factor-at-a-time tradeoff method and the multiple factor full-concept method. With the tradeoff method, respondents are asked to rank the cells of a series of matrices, each matrix crossing the levels of one factor with the levels of another.

Why Use Conjoint Analysis?

  • Effective market research is integral to the design, manufacture, and sale of successful products. It identifies the needs and wants of target markets, ensuring that products will sell because they meet the needs of buyers.
  • Conjoint analysis is a market research tool for developing effective product design.
  • Using conjoint analysis, the researcher can answer questions such as: What product attributes is important or unimportant to the consumer? What levels of product attributes are the most or least desirable ones in the consumer’s mind? What is the market share of preference for leading competitors’ products versus our existing or proposed product? Answers to these questions are of crucial importance in the design and launch of a successful product.
  • The virtue of conjoint analysis is that it asks the respondent to make choices in the same fashion as the consumer presumably does—by trading off features, one against another.

For example, suppose that you want to book an airline flight. You have the choice of sitting in a cramped seat or a spacious seat. If this were the only consideration, your choice would be clear. You would probably prefer a spacious seat. Or suppose you have a choice of ticket prices: $225 or $800. On price alone, taking nothing else into consideration, the lower price would be preferable. Finally, suppose you can take either a direct flight, which takes two hours, or a flight with one layover, which takes five hours. Most people would choose the direct flight.

Steps In The Application Of Conjoint Analysis

The main steps involved in the application of Conjoint Analysis are following:

1. Determination of the salient attributes for the given product from the points of view of the consumers

2. Assigning a set of discrete levels or a range of continuous values to each of the attributes.

3. Utilizing Fractional Factorial Design of Experiment for designing the stimuli for experiment.

4. Physically designing the stimuli

5. Ranking or Rating data collection

6. Conjoint analysis and determination of part worth utilities.

7. Applying conjoint analysis output for different marketing decisions

How does Conjoint Analysis Work?
Conjoint analysis involves the measurement of consumer preferences, or acceptability between choice alternatives. The name "Conjoint Analysis" implies the study of the joint effects. In marketing applications, we study the joint effects of multiple product attributes on product choice. When asked to do so outright, many consumers are unable to determine the relative importance that they place on product attributes. For example, when asked which attributes are the more important ones, the response may be that “they all are important”.

It is difficult for a survey respondent to take a list of attributes and mentally construct the preferred combinations of them. The task is easier if the respondent is presented with combinations of attributes that can be visualized as different product offerings. Fortunately, conjoint analysis can facilitate the process. Conjoint analysis is a tool that allows a subset of the possible combinations of product features to be used to determine the relative importance of each feature in the purchasing decision; the relative values of attributes considered jointly can better be measured than when considered in isolation.

Author: Juhi Priyanka Kachhap (13136)

Group: Marketing - Group 4

Sports Endorsements(Brand Fit) using Conjoint Analysis

Brand Fit & Brand Image Transfer – An amazing Application of Conjoint Analysis

With BA ending today, and me being a marketer, i thought to go beyond the conventional Pricing and other applications. My search led me to a paper written by S. Bucker, where he has tried to find out Brand Fit & Brand Image transfer using Conjoint Analysis.
Brand image transfer is the transfer of brand associations, attributed to another entity, to the brand, while brand fit has been defined as a consumer learning process that seeks to match those brand associations held of the relevant brands involved. Since a variety of brands were involved in this study, conjoint analysis was used as a method of assesing brand fit. In particular, choice-based conjoint analysis was singled out because of its capability to allow the relative advantage of brands considered jointly to be ascertained. Brands might not be able to be measured if taken one at a time. Both qualitative and quantitative research methods were employed in order to assess brand fit using conjoint analysis, which was the main reason for this study.

Rugby sponsorships were chosen to asses brand fit, as this particular game is the second most-watched sport in South Africa, with the highest monetary value attached to its sponsorships at the time of this study. The qualitative research was accomplished by using focus groups to determine which brands were typically perceived to be associated with the Springbok rugby brand. The
different industries and brands used in the focus groups were selected on the grounds of their being current, previous, and potential sponsors of the Springbok rugby brand. The quantitative research was conducted by means of an online questionnaire, sent via a link in an email to a chosen database on the social networking site, ‘Facebook’. A screening question served to ensure that only rugby supporters would be eligible to complete the survey. The information was captured in ‘real time’ in the conjoint analysis software, thereby determining which brands were perceived to best fit the Springbok rugby brand.

The realised sample was composed of a younger, more male-dominated group. All respondents were also Springbok supporters who possessed sufficient knowledge on the Springbok brand and sponsors. There were six brands identified to fit the Springbok rugby brand, namely, Castle, Vodacom, SASOL,Canterbury, Nike, and Adidas. These identified brands proved that the study
did indeed assess brand fit using conjoint analysis.

Conclusions were drawn that brand fit could be established in a variety of ways.The most dominant ways were by leveraging the sponsorship, and also by sponsoring on a continuing basis. These two ways serve to inform consumers of the sponsorship, making them aware of the brands, and building the basis of brand fit in their minds. Brand fit was also achieved based on similar brand
images of the two different brands.

The conclusion drawn were to analyze competitors at various levels – Cell Phone, Alcoholic, Sports Gear etc. The conclusion drawn were based on conjoint analysis and also the results of the focus group activities. Each brand’s score decided whether it was a fit or a no-fit in the category.

Neelima Makani
Marketing 2

Conjoint Analysis - Upside & Flipsyde


Conjoint Analysis is concerned with understanding how people make choices between products or services or a combination of product and service, so that businesses can design new products or services that better meet customers’ underlying needs. A key benefit of conjoint analysis is the ability to produce dynamic market models that enable companies to test out what steps they would need to take to improve their market share, or how competitors’ behavior will affect their customers.

To understand how conjoint analysis works, we need to be able to describe the products or services consistently in terms of attributes and levels in order to see what is being traded off. Conjoint analysis is a sophisticated technique and there are technical issues that need to be considered. In particular, the design of attributes is a crucial step in a conjoint project as choices between poorly defined levels can render the exercise meaningless. You should also be aware that there are different flavors of conjoint analysis depending on the application. Adaptive Conjoint Analysis (ACA) is the most common, but there is also Choice-based and Full-profile Conjoint Analysis.

Conjoint analysis proves to be extremely helpful since estimates psychological tradeoffs that consumers make when evaluating several attributes together. Also conjoint analysis is known to measures preferences at the individual level.

On the flipside conjoint analysis has several disadvantages such as respondents are unable to articulate attitudes toward new categories, or may feel forced to think about issues. Conjoint analysis also is said to have a major flaw that it does not take into account the number items per purchase so it can give a poor reading of market share.

Nikhil Kumar
Marketing 2

Conjoint Analysis - A Marketer's Tool

                         CONJOINT ANALYSIS

Conjoint (trade-off) analysis has become one of the most widely-used quantitative methods in Marketing Research. It is used to measure the perceived values of specific product features, to learn how demand for a particular product or service is related to price, and to forecast what the likely acceptance of a product would be if brought to market.
Rather than directly ask survey respondents what they prefer in a product, or what attributes they find most important, conjoint analysis employs the more realistic context of respondents evaluating potential product profiles.  By understanding precisely how people make decisions and what they value in your products and services, you can work out the optimum level of features and services that balance value to the customer against cost to the company. The principle behind conjoint analysis is to break a product or service down into it's constituent parts then to test combinations of these parts to look at what customers prefer. By designing the study appropriately it is then possible to use statistical analysis to work out the value of each part in driving the customer’s decision. By analysing which items are chosen or preferred from the product profiles offered to the customer it is possible to work out statistically both what is driving the preference from the attributes and levels shown, but more importantly, give an implicit numerical valuation for each attribute and level.
By combining these market models with internal project costing, companies can evaluate decisions in terms of Return on Investment (ROI) before going to market.

Dishant Sharma
Marketing 2