Scholarlink

In this article, we present an approach for estimating the parameters of probabilistic choice models for individuals or groups when the amount of data available for each decision-making unit (DMU) is sparse. The approach hinges on the assumption that these parameters vary across DMUs and decisions made by the same DMU in a systematic manner. Furthermore, it is assumed that the systematic variation is dictated by a set of personal or situational variables that are known and observable. We demonstrate that these assumptions allow one to aggregate data across DMUs and yet estimate an idiosyncratic set of parameters for each DMU. A significant aspect of the output of the approach is the ability to assess the relative impact of the personal and situational variables on each DMU’s parameter values. As an illustration, the approach is applied to a weighted probability model of group choice proposed by Choffray and Lilien (1980).

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Estimating Probabilistic Choice Models from Sparse Data: A Method and an Application to Groups

  • 章节

  • 发表日期：

    2026年

查看全文
  • Joel H. Steckel,
  • Donald R. Lehmann,
  • Kim P. Corfman

摘要

In this article, we present an approach for estimating the parameters of probabilistic choice models for individuals or groups when the amount of data available for each decision-making unit (DMU) is sparse. The approach hinges on the assumption that these parameters vary across DMUs and decisions made by the same DMU in a systematic manner. Furthermore, it is assumed that the systematic variation is dictated by a set of personal or situational variables that are known and observable. We demonstrate that these assumptions allow one to aggregate data across DMUs and yet estimate an idiosyncratic set of parameters for each DMU. A significant aspect of the output of the approach is the ability to assess the relative impact of the personal and situational variables on each DMU’s parameter values. As an illustration, the approach is applied to a weighted probability model of group choice proposed by Choffray and Lilien (1980).