Aditya Kannan

HR Batch 2011-2013

Group G

We started today’s business analytics classes with an overview of another final exam paper in the subject in another institute. We saw that the instructions are given very clearly about how to obtain results for various analyses such as first level, frequencies, cross tabs, cluster and factor analysis. The steps involved in arriving at collated results in table forms, in a crisp manner were given. The professor has made it easy for us by telling us not to get worked up about the steps. I completely agree with his point of view that the strategies to be formed and the final analysis of the data and results are much more important than the steps. Yes, the steps are important, but even if we make a mistake in the steps but using the data and results obtained, if our analysis is correct then we get awarded with marks. If we perform the steps mentioned in the paper correctly we do get marks for them. That was comforting.

We worked a lot on the Bank loan file. Using this file we worked on factor, cluster, frequencies and discriminant analysis. We tried to analyze the credit worthiness of the people, what different factors help us in deciding whether the person will default on the loan or not. Some of the factors we saw are age groups, number of years on his job, his income and address. We had made some hypotheses about the number of people defaulting or not and we saw how close our predictions are. As learnt from Wikipedia, Discriminant or Linear discriminant analysis is a method used in statistics, pattern recognition and machine learning to find a linear combination of features which characterizes or separates two or more classes of objects or events. The resulting combination may be used as a linear classifier, or, more commonly, for dimensionality reduction before later classification. Linear discriminant analysis is used in bankruptcy prediction, face recognition and marketing.

We worked on Utility analysis. In this we try to find out that from certain data what is more important to people, their preferences. For example we worked on what are people’s preferences when it comes to buying a laptop computer. We tried to find the utility value for three factors, price of the machine, battery life and its weight. We did this by putting all possible combinations of the factors and then ranking them on a scale of 1 to 8 with 8 being most preferred and 1 being least preferred. For all combinations we then added all the rankings and then subtracted the totals for each variable.

We started work on Conjoint. It is about getting answers for choices such as for example a direct or flight with a stopover, what people really want, what they can afford, what he would like to trade-off with the other choice. The questions are asked jointly as to what is more important to people. We worked on generating orthogonal design which helps to create the choices to choose from, what is more preferred.

For conjoint we formed a separate SPSS file called plan, which concerns career choices. We got the following after forming a file on SPSS.

Company	Industry	Salary	Growth	Satisfaction	WLBalance	Security
National	Service	Above average	Medium	Moderate	No	No
National	Manufacturing	Below average	Fast	Moderate	No	Yes
MNC	Manufacturing	Above average	Medium	High	No	Yes
National	Service	Above average	Fast	High	Yes	Yes
MNC	Service	Below average	Medium	Moderate	Yes	Yes
National	Manufacturing	Below average	Medium	High	Yes	No
MNC	Manufacturing	Above average	Fast	Moderate	Yes	No
MNC	Service	Below average	Fast	High	No	No

It was found that certain combinations were not there such as the following:-

MNC

Service

Above average

Fast

High

Yes

Yes

Regarding the absence of the above, we will work on it in the next class.