Variable selection for market basket analysis

Dippold, Katrin; Hruschka, Harald

doi:10.1007/s00180-012-0315-3

Cited by 13 publications

(12 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Chib et al (2002) nd that the neglect of heterogeneity induces overestimated cross-category eects. In contrast, Dippold and Hruschka (2010) demonstrate that augmenting the number of included product categories increases the percentage of pairwise independent categories. Some publications with logit models incorporate a higher number of product categories.…”

Section: Literature Review On Multicategory Purchase Incidence Modelsmentioning

confidence: 77%

“…These models disregard heterogeneity in purchase behavior and in reactions to marketing measures. Hruschka (1991) and Hruschka et al (1999) as well as Dippold and Hruschka (2010) neither integrate heterogeneity in the model structure nor allow for customer-specic covariates. Likewise, Russell and Petersen (2000) do not account for latent heterogeneity.…”

Section: Literature Review On Multicategory Purchase Incidence Modelsmentioning

confidence: 99%

“…Boztu § and Reutterer (2008) circumvent the estimation of a complex MVL model with all 65 product categories in their data set by estimating a single MVL model for every prototype that contains only the 4 or 5 most important or distinguished (p.305) categories thus limiting the computational burden exogenously. For a data set with 30 product categories, Dippold and Hruschka (2010) apply Bayesian variable selection techniques that endogenously determine which category pairs are independent.…”

Section: Literature Review On Multicategory Purchase Incidence Modelsmentioning

confidence: 99%

“…In order to handle the -with number of categories quadratically growing -amount of coecients, we apply a Bayesian variable selection technique that limits the enormous number of intercategory relationships to the relevant interaction eects as in Dippold and Hruschka (2010).…”

Section: Literature Review On Multicategory Purchase Incidence Modelsmentioning

confidence: 99%

“…Hruschka et al (1999) show that purchases are independent for 95 percent of all category pairs in their data with over 70 categories. Dippold and Hruschka (2010) nd 66 percent insignicant intercategory eects in a basket data set with 30 product categories.…”

Section: Model 21 Finite Mixture Multivariate Logit Model (Fm-mvl)mentioning

confidence: 99%

See 4 more Smart Citations

A Model of Heterogeneous Multicategory Choice for Market Basket Analysis

Dippold

Hruschka

2013

Review of Marketing Science

Self Cite

View full text Add to dashboard Cite

Based on market basket data, multicategory purchase incidence models analyze demand interdependencies between product categories. We propose a nite mixture multivariate logit model to derive segment-specic intercategory eects of market basket purchase. Under the assumption that only a fraction of intercategory eects are signicant, we exclude irrelevant eects by variable selection. This leads to a detailed description of consumers' shopping behavior that varies over segments not only w.r.t. parameters' values but also w.r.t. included interaction eects. We nd that a homogeneous model would overestimate the intensity of interaction between product categories.

show abstract

Section: Literature Review On Multicategory Purchase Incidence Modelsmentioning

confidence: 77%

Section: Literature Review On Multicategory Purchase Incidence Modelsmentioning

confidence: 99%

Section: Literature Review On Multicategory Purchase Incidence Modelsmentioning

confidence: 99%

Section: Literature Review On Multicategory Purchase Incidence Modelsmentioning

confidence: 99%

Section: Model 21 Finite Mixture Multivariate Logit Model (Fm-mvl)mentioning

confidence: 99%

See 3 more Smart Citations

A Model of Heterogeneous Multicategory Choice for Market Basket Analysis

Dippold

Hruschka

2013

Review of Marketing Science

Self Cite

View full text Add to dashboard Cite

show abstract

Comparing Small‐ and Large‐Scale Models of Multicategory Buying Behavior

Hruschka

2013

Journal of Forecasting

Self Cite

View full text Add to dashboard Cite

Focusing on the interdependence of product categories we analyze multicategory buying decisions of households by a finite mixture of multivariate Tobit-2 models with two response variables: purchase incidence and expenditure. Mixture components can be interpreted as household segments. Correlations for purchases of different categories turn out to be much more important than correlations among expenditures as well as correlations among purchases and expenditures of different categories. About 18% of all pairwise purchase correlations are significant. We compare the best-performing large-scale model with 28 categories to four small-scale models each with seven categories. In our empirical study the large-scale model clearly attains a better forecasting performance. The small-scale models provide several biased correlations and miss about 50% of the significant correlations which the large scale model detects.

show abstract