Marketing Applications of Bayesian Nonparametric Methods
Marketing Applications of Bayesian Nonparametric Methods

Author: Yuhao Fan

Publisher:

Published: 2021

Total Pages: 0

ISBN-13:

DOWNLOAD EBOOK

I explore the application of Bayesian statistical modelling, and in particular Bayesian nonparametric methods in marketing research. I apply Bayesian nonparametric methods in both chapters of my dissertation to model two types of customer dynamics.In the first chapter, I investigate the impact of implementing a free cancellation program on customer behavior and firm profits in a hostel booking setting. While many firms have recently introduced free cancellation programs, the impact of such programs on customer behavior and firm profits remains unclear. I investigate this question empirically, using data from a hostel booking platform that recently introduced a free cancellation program. To understand the program's impact on a myriad of aspects of customer behavior, including booking timing, spend amount, and propensity to cancel, while also accounting for latent attrition and customer heterogeneity, I build a hierarchical, Bayesian nonparametric model of behavior, leveraging Gaussian process change points to capture the effect of the free cancellation program on booking dynamics, and a Dirichlet process mixture specification for customer heterogeneity. These nonparametric components of the model allow us to make minimal assumptions about important aspects of booking behavior, while uncovering rich insights about the time-varying impact of the program, and the heterogeneity of customers. Our results suggest that the free cancellation program led customers to book more frequently, book earlier, spend more, and cancel more of their trips. Crucially, the increase in bookings generally outweighed the increase in cancellations in long term, resulting in an increase in average customer lifetime value.In the second chapter, I apply Bayesian nonparametric methods, in particular, Multi-output Gaussian Process, to model the cross-category dynamics of customers' preference parameters in brand choice models. I show that the proposed model allows us to transfer information about customers' preference parameters within and across categories, and that modelling the cross-category dynamics of customers' preference parameters improves model fit and prediction accuracy. Moreover, leveraging information across categories gives us more reliable estimates of price elasticities. Together, these two chapters illustrate the power of Bayesian methods to gain deep insights into dynamic marketing problems.

Bayesian Nonparametric Methods in Marketing
Bayesian Nonparametric Methods in Marketing

Author: Saisandeep Reddy Satyavolu

Publisher:

Published: 2016

Total Pages: 126

ISBN-13:

DOWNLOAD EBOOK

The proliferation of available data in marketing has placed an emphasis on the applicability of extant marketing models to big data. To tackle this problem, methods from machine learning have been increasingly applied by the marketing community. This line of research is a subset of research in marketing that is becoming interdisciplinary. A number of marketing researchers have successfuly adopted methods from other seemingly unrelated fields in their research. In that vein, this thesis examines the applicability of Bayesian Nonparametric methods (from the field of machine learning) to marketing. The first chapter of this thesis provides a very brief survey of marketing research papers that have enhanced pure marketing models using methods from machine learning. The second chapter describes the Dirichlet Process, a key component of Bayesian Nonparametric analysis and provides two synthetic data applications. Going forward, we study the applicability of Bayesian Nonparametric methods to model Heterogeneity across multiple markets. Bayesian Nonparametric methods have been used in marketing and economics literature to model heterogeneity in discrete choice models, but past applications have only been limited to data from a single market. So as to compare heterogeneity in consumer preferences across multiple markets, we use the Hierarchical Dirichlet Process (HDP) which lets multiple "groups" of data "share statistical strength". Heterogeneity across multiple markets is modeled using the HDP in two different contexts (B2C and B2B) in this thesis. Our work shows that the HDP provides a convenient "middle ground" to other extreme modeling options, which are (1) ignore heterogeneity of preferences across markets and (2) model each market separately. Another aspect of the HDP is the ease with which it can be incorporated into models of discrete choice. The models developed and estimated in this thesis are also helpful for the marketing manager. In the B2C application, the results of the model provide the manager with a practical way of tailoring targeting activities towards consumers with varying preferences. Finally, in the B2B application, we find that based on the Stage of the selling process, some marketing activities play a larger role than others in converting sales leads into clients. These results provide a data driven basis for the manager to appropriately allocate marketing dollars to activities based on the selling process.

Bayesian Non- and Semi-parametric Methods and Applications
Bayesian Non- and Semi-parametric Methods and Applications

Author: Peter Rossi

Publisher: Princeton University Press

Published: 2014-04-27

Total Pages: 219

ISBN-13: 1400850304

DOWNLOAD EBOOK

This book reviews and develops Bayesian non-parametric and semi-parametric methods for applications in microeconometrics and quantitative marketing. Most econometric models used in microeconomics and marketing applications involve arbitrary distributional assumptions. As more data becomes available, a natural desire to provide methods that relax these assumptions arises. Peter Rossi advocates a Bayesian approach in which specific distributional assumptions are replaced with more flexible distributions based on mixtures of normals. The Bayesian approach can use either a large but fixed number of normal components in the mixture or an infinite number bounded only by the sample size. By using flexible distributional approximations instead of fixed parametric models, the Bayesian approach can reap the advantages of an efficient method that models all of the structure in the data while retaining desirable smoothing properties. Non-Bayesian non-parametric methods often require additional ad hoc rules to avoid "overfitting," in which resulting density approximates are nonsmooth. With proper priors, the Bayesian approach largely avoids overfitting, while retaining flexibility. This book provides methods for assessing informative priors that require only simple data normalizations. The book also applies the mixture of the normals approximation method to a number of important models in microeconometrics and marketing, including the non-parametric and semi-parametric regression models, instrumental variables problems, and models of heterogeneity. In addition, the author has written a free online software package in R, "bayesm," which implements all of the non-parametric models discussed in the book.

Bayesian Non- and Semi-parametric Methods and Applications
Bayesian Non- and Semi-parametric Methods and Applications

Author: Peter Rossi

Publisher: Princeton University Press

Published: 2014-04-27

Total Pages: 218

ISBN-13: 0691145326

DOWNLOAD EBOOK

This book reviews and develops Bayesian non-parametric and semi-parametric methods for applications in microeconometrics and quantitative marketing. Most econometric models used in microeconomics and marketing applications involve arbitrary distributional assumptions. As more data becomes available, a natural desire to provide methods that relax these assumptions arises. Peter Rossi advocates a Bayesian approach in which specific distributional assumptions are replaced with more flexible distributions based on mixtures of normals. The Bayesian approach can use either a large but fixed number of normal components in the mixture or an infinite number bounded only by the sample size. By using flexible distributional approximations instead of fixed parametric models, the Bayesian approach can reap the advantages of an efficient method that models all of the structure in the data while retaining desirable smoothing properties. Non-Bayesian non-parametric methods often require additional ad hoc rules to avoid "overfitting," in which resulting density approximates are nonsmooth. With proper priors, the Bayesian approach largely avoids overfitting, while retaining flexibility. This book provides methods for assessing informative priors that require only simple data normalizations. The book also applies the mixture of the normals approximation method to a number of important models in microeconometrics and marketing, including the non-parametric and semi-parametric regression models, instrumental variables problems, and models of heterogeneity. In addition, the author has written a free online software package in R, "bayesm," which implements all of the non-parametric models discussed in the book.

Bayesian Statistics and Marketing
Bayesian Statistics and Marketing

Author: Peter E. Rossi

Publisher: John Wiley & Sons

Published: 2024-09-10

Total Pages: 405

ISBN-13: 1394219113

DOWNLOAD EBOOK

Fine-tune your marketing research with this cutting-edge statistical toolkit Bayesian Statistics and Marketing illustrates the potential for applying a Bayesian approach to some of the most challenging and important problems in marketing. Analyzing household and consumer data, predicting product performance, and custom-targeting campaigns are only a few of the areas in which Bayesian approaches promise revolutionary results. This book provides a comprehensive, accessible overview of this subject essential for any statistically informed marketing researcher or practitioner. Economists and other social scientists will find a comprehensive treatment of many Bayesian methods that are central to the problems in social science more generally. This includes a practical approach to computationally challenging problems in random coefficient models, non-parametrics, and the problems of endogeneity. Readers of the second edition of Bayesian Statistics and Marketing will also find: Discussion of Bayesian methods in text analysis and Machine Learning Updates throughout reflecting the latest research and applications Discussion of modern statistical software, including an introduction to the R package bayesm, which implements all models incorporated here Extensive case studies throughout to link theory and practice Bayesian Statistics and Marketing is ideal for advanced students and researchers in marketing, business, and economics departments, as well as for any statistically savvy marketing practitioner.

Practical Nonparametric and Semiparametric Bayesian Statistics
Practical Nonparametric and Semiparametric Bayesian Statistics

Author: Dipak D. Dey

Publisher: Springer Science & Business Media

Published: 2012-12-06

Total Pages: 376

ISBN-13: 1461217326

DOWNLOAD EBOOK

A compilation of original articles by Bayesian experts, this volume presents perspectives on recent developments on nonparametric and semiparametric methods in Bayesian statistics. The articles discuss how to conceptualize and develop Bayesian models using rich classes of nonparametric and semiparametric methods, how to use modern computational tools to summarize inferences, and how to apply these methodologies through the analysis of case studies.

Advanced Methods for Modeling Markets
Advanced Methods for Modeling Markets

Author: Peter S. H. Leeflang

Publisher: Springer

Published: 2017-08-29

Total Pages: 725

ISBN-13: 3319534696

DOWNLOAD EBOOK

This volume presents advanced techniques to modeling markets, with a wide spectrum of topics, including advanced individual demand models, time series analysis, state space models, spatial models, structural models, mediation, models that specify competition and diffusion models. It is intended as a follow-on and companion to Modeling Markets (2015), in which the authors presented the basics of modeling markets along the classical steps of the model building process: specification, data collection, estimation, validation and implementation. This volume builds on the concepts presented in Modeling Markets with an emphasis on advanced methods that are used to specify, estimate and validate marketing models, including structural equation models, partial least squares, mixture models, and hidden Markov models, as well as generalized methods of moments, Bayesian analysis, non/semi-parametric estimation and endogeneity issues. Specific attention is given to big data. The market environment is changing rapidly and constantly. Models that provide information about the sensitivity of market behavior to marketing activities such as advertising, pricing, promotions and distribution are now routinely used by managers for the identification of changes in marketing programs that can improve brand performance. In today’s environment of information overload, the challenge is to make sense of the data that is being provided globally, in real time, from thousands of sources. Although marketing models are now widely accepted, the quality of the marketing decisions is critically dependent upon the quality of the models on which those decisions are based. This volume provides an authoritative and comprehensive review, with each chapter including: · an introduction to the method/methodology · a numerical example/application in marketing · references to other marketing applications · suggestions about software. Featuring contributions from top authors in the field, this volume will explore current and future aspects of modeling markets, providing relevant and timely research and techniques to scientists, researchers, students, academics and practitioners in marketing, management and economics.

Bayesian Nonparametrics via Neural Networks
Bayesian Nonparametrics via Neural Networks

Author: Herbert K. H. Lee

Publisher: SIAM

Published: 2004-01-01

Total Pages: 106

ISBN-13: 9780898718423

DOWNLOAD EBOOK

Bayesian Nonparametrics via Neural Networks is the first book to focus on neural networks in the context of nonparametric regression and classification, working within the Bayesian paradigm. Its goal is to demystify neural networks, putting them firmly in a statistical context rather than treating them as a black box. This approach is in contrast to existing books, which tend to treat neural networks as a machine learning algorithm instead of a statistical model. Once this underlying statistical model is recognized, other standard statistical techniques can be applied to improve the model. The Bayesian approach allows better accounting for uncertainty. This book covers uncertainty in model choice and methods to deal with this issue, exploring a number of ideas from statistics and machine learning. A detailed discussion on the choice of prior and new noninformative priors is included, along with a substantial literature review. Written for statisticians using statistical terminology, Bayesian Nonparametrics via Neural Networks will lead statisticians to an increased understanding of the neural network model and its applicability to real-world problems.