Integrating Statistical Methods for Characterizing Causal Influences on Planner Behavior over Time
نویسندگان
چکیده
Given a complex planner and or environment, it can be di cult to determine why it behaves as it does. Statistical causal modeling techniques allow us to develop models of behavior, but they tend to be limited in what they can model: either continuing, repetitive in uences or causal in uences without cycles, but not both as appear in most planning environments. This paper describes how two statistical modelling techniques can be combined to suggest speci c hypotheses about how the environment and the planner's design causally in uence the planner's behavior over many examples of interacting in its environment and to construct models of those in uences. One technique, dependency detection, is designed to identify relationships (dependencies) between particular failures, the methods that repair them and the occurrence of failures downstream. Another method, path analysis, builds causal models of correlational data. Dependency detection operates over a series of events, and path analysis models within a temporal snapshot. We explain the integration of the techniques and demonstrate it on data from the Phoenix planner. This research was supported by ARPA-AFOSR contract F30602-93-C-0100. The US Government is authorized to reproduce and distribute reprints for governmental purposes notwithstanding any copyright notation hereon.
منابع مشابه
Integrating Statistical Methods for Characterizing Causal Innuences on Planner Behavior over Time
متن کامل
کاربرد مدل ارتقای سلامت در مطالعه رفتار فعالیت جسمانی دانش آموزان شهر سنندج
Background and Aim: There is evidence that physical activity declines during adolescence in Western countries. However, this pattern has not yet been shown in Iranian youth. The purpose of this study was to detect changes in physical activity behaviors, psychological factors (including perceived benefits, barriers, and self-efficacy), and interpersonal influences associated with exercise behav...
متن کاملIndirect Causes in Dynamic Bayesian Networks Revisited
Modeling causal dependencies often demands cycles at a coarse-grained temporal scale. If Bayesian networks are to be used for modeling uncertainties, cycles are eliminated with dynamic Bayesian networks, spreading indirect dependencies over time and enforcing an infinitesimal resolution of time. Without a “causal design,” i.e., without anticipating indirect influences appropriately in time, we ...
متن کاملTopographic and meteorological influences on spatial scaling of heavy convective rainfall in mountainous regions
Characterizing the spatial scaling and dynamics of convective precipitation in mountainous terrain and the development of downscaling methods to transfer precipitation fields from one scale to another is the overall motivation for this research. Substantial progress has been made on characterizing the space-time organization of midwestern convective systems and tropical rainfall, which has led ...
متن کاملMethods for Finding Innuences on Program Failure 1.1 Previous Approaches to Solution 2 Dependency Detection
This paper describes two approaches for detecting patterns of detrimental program behavior, called dependencies , over long periods of time; these dependencies indicate cases where previous events innuence the occurrence of later failure. This research extends a previous approach that was limited to temporally adjacent events. The two approaches, heuristic search and local search, are demonstra...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 1994