|
on Discrete Choice Models |
| By: | B. LARIVIÈRE; D. VAN DEN POEL |
| Abstract: | This study addresses an important issue for both managers and researchers: whether it is advantageous for financial services providers to invest in youth marketing. More specifically, the effectiveness of these investments is evaluated in terms of retention proneness once youngsters enter the lifecycle event of “moving in together”. The study identifies eight constructs of youth marketing and contrasts their impact against the best deal when youngsters decide to move in together and consequently experience the need to buy their first collectivized financial products, such as a joint account or a mortgage for their new home. Furthermore, the influence of the partner, prior patronage behaviour, customer demographics and psychographic variables are tested for. The findings of the study reveal that (i) individuals are likely to change their banking behaviour during crucial lifetime events such as moving in together, (ii) not all youth marketing investments are equally effective, while (iii) the best deal components (e.g. convenience, price conditions, etc.) have a major impact. |
| Keywords: | Marketing; Banking; Strategic planning; Ordered logit analysis. |
| Date: | 2007–01 |
| URL: | http://d.repec.org/n?u=RePEc:rug:rugwps:07/433&r=dcm |
| By: | A. PRINZIE; D. VAN DEN POEL |
| Abstract: | Several supervised learning algorithms are suited to classify instances into a multiclass value space. MultiNomial Logit (MNL) is recognized as a robust classifier and is commonly applied within the CRM (Customer Relationship Management) domain. Unfortunately, to date, it is unable to handle huge feature spaces typical of CRM applications. Hence, the analyst is forced to immerse himself into feature selection. Surprisingly, in sharp contrast with binary logit, current software packages lack any feature selection algorithm for MultiNomial Logit. Conversely, Random Forests, another algorithm learning multi class problems, is just like MNL robust but unlike MNL it easily handles high-dimensional feature spaces. This paper investigates the potential of applying the Random Forests principles to the MNL framework. We propose the Random MultiNomial Logit (RMNL), i.e. a random forest of MNLs, and compare its predictive performance to that of a) MNL with expert feature selection, b) Random Forests of classification trees. We illustrate the Random MultiNomial Logit on a cross-sell CRM problem within the home-appliances industry. The results indicate a substantial increase in model accuracy of the RMNL model to that of the MNL model with expert feature selection. |
| Keywords: | multiclass classifier design and evaluation, feature evaluation and selection, data mining methods and algorithms, customer relationship management (CRM) |
| Date: | 2007–01 |
| URL: | http://d.repec.org/n?u=RePEc:rug:rugwps:07/435&r=dcm |
| By: | Susumu Imai (Queen's University); Neelam Jain (Northern Illinois University); Andrew Ching (University of Toronto) |
| Abstract: | We propose a new methodology for structural estimation of dynamic discrete choice models. We combine the Dynamic Programming (DP) solution algorithm with the Bayesian Markov Chain Monte Carlo algorithm into a single algorithm that solves the DP problem and estimates the parameters simultaneously. As a result, the computational burden of estimating a dynamic model becomes comparable to that of a static model. Another feature of our algorithm is that even though per solution-estimation iteration, the number of grid points on the state variable is small, the number of effective grid points increases with the number of estimation iterations. This is how we help ease the "Curse of Dimensionality". We simulate and estimate several versions of a simple model of entry and exit to illustrate our methodology. We also prove that under standard conditions, the parameters converge in probability to the true posterior distribution, regardless of the starting values. |
| Keywords: | Bayesian Estimation, Dynamic Discrete Choice Model, Dynamic Programming, Markov Chain Monte Carlo, Bayesian Dynamic Programming Estimation |
| JEL: | C51 C61 C63 L00 |
| Date: | 2006–12 |
| URL: | http://d.repec.org/n?u=RePEc:qed:wpaper:1118&r=dcm |
| By: | Arne Risa Hole (Centre for Health Economics, University of York) |
| Abstract: | This paper examines the distribution of preferences in a sample of patients who responded to a discrete choice experiment on the choice of general practitioner appointments. In addition to standard logit, mixed and latent class logit models are used to analyse the data from the choice experiment. It is found that there is significant preference heterogeneity for all the attributes in the experiment and that both the mixed and latent class models lead to significant improvements in fit compared to the standard logit model. Moreover, the distribution of preferences implied by the preferred mixed and latent class models is similar for many attributes. |
| Keywords: | discrete choice experiment; mixed logit; latent class logit |
| JEL: | I10 C25 |
| Date: | 2007–01 |
| URL: | http://d.repec.org/n?u=RePEc:chy:respap:rp22&r=dcm |