Dierker Lisa, Rose Jennifer, Tan Xianming, Li Runze
Psychology Department, Wesleyan University, 207 High Street, Middletown, CT 06459, USA.
J Prim Prev. 2010 Dec;31(5-6):333-48. doi: 10.1007/s10935-010-0224-6.
This paper describes and compares a selection of available modeling techniques for identifying homogeneous population subgroups in the interest of informing targeted substance use intervention. We present a nontechnical review of the common and unique features of three methods: (a) trajectory analysis, (b) functional hierarchical linear modeling (FHLM), and (c) decision tree methods. Differences among the techniques are described, including required data features, strengths and limitations in terms of the flexibility with which outcomes and predictors can be modeled, and the potential of each technique for helping to inform the selection of targets and timing of substance intervention programs.
本文描述并比较了一系列现有的建模技术，这些技术用于识别同质人群亚组，以指导有针对性的物质使用干预。我们对三种方法的共同特征和独特特征进行了非技术性综述：(a)轨迹分析，(b)功能分层线性建模(FHLM)，以及(c)决策树方法。描述了这些技术之间的差异，包括所需的数据特征、在对结果和预测变量进行建模的灵活性方面的优势和局限性，以及每种技术在帮助指导物质干预项目的目标选择和时机方面的潜力。
