慶應義塾大学 経済学部 PEARL入試 志望理由書 提出例(星野 崇宏先生ゼミ向け)
Dr. Takahiro Hoshino
Department of Economics, Econometrics
Dear Professor Hoshino,
I am writing this letter to explain my motivation in applying for Department of Economics at Keio University, specializing in Econometrics and related fields. As we are exposed to more information than ever, the need for us to truly understand and process them only get bigger. I have read a number of your published work which I was very intrigued by. I hope I am able to elaborate on area of studies that I think is very much relevant, and I would be more than grateful if you could kindly give this a consideration.
A semiparametric model is considered a regression model with both a finite and an infinite-dimensional component. In research and statistics projects involving semi-parametric models, the emphasis is almost always on the parametric component of the model. That is usually because this is the part which lends itself well to research, but then why would we use semiparametric models? Too often, parametric models though they are easy to understand and easy to work with, fail to give a fair representation of what is happening in the real world that are very human and chaotic. As opposed, non-parametric models may be better representations but do not lend themselves well to logical analysis. A semiparametric model allows you to have the best of both worlds A) understandable and can be manipulated B) also offer a fair representation of the messiness that is involved in real life. There is high value in examining the theory in real life case scenarios.
In the marketing literature there does not seem to be a widely accepted answer to the question of whether, when purchasing brands in certain product category consumers react differently to a reduction in price and an increase in discount deals. Some previous studies suggest, prices and deals have a linear effect on a brand’s indirect utility. However, it is possible to estimate the utility of price and discount nonparametrically and Instead of a linear structure on this function, it may require only that it be decreasing in price and increasing in deal amount. This specification allows us for a general pattern of interaction effects between prices and deals to influence buying actions and is also applicable on estimating brand choice models. Meaning, they recommend we use a semiparametric approach, in which the distribution of the stochastic components of brand utilities is specified parametrically, and a fully nonparametric approach in which this distribution is left unspecified. It is necessary to carry out a number of analyses on panel datasets for across product categories before coming to a conclusion.
Semiparametric models seem to be useful when it comes to market share analysis and brand choice analysis with both definite and indefinite components. With the prospect of even complex age of international players interacting with each other, and consumers looking for brand advocacy in buying, marketing understanding consumer behaviour is crucial. I assume this can be an addition to a number of researches conducted in your seminar and I would love to take part. Thank you very much for taking the time to read and I look forward to hearing from you soon on this matter.
*”Semiparametric Bayesian Estimation for Marginal Parametric Potential Outcome Modeling: Application to Causal Inference”, Journal of the American Statistical Association, 2013, 108, 1189-1204.