While the notion of geometric intrinsic dimension has been studied for complete formal contexts, its behavior in the presence of incomplete data was not investigated until now. In this paper, we investigate the geometric intrinsic dimension of incomplete formal contexts, where the object-attribute relationship is partially unknown or undefined. Specifically, we introduce a theoretical upper bound on the intrinsic dimension of the completion of such contexts, derived from so-called stable certain concepts. These are defined by well-known patterns of incompleteness from the literature. We validate our theoretical results through computational experiments and discuss their implications. We also show that certain concepts can be a universal tool for analyzing incomplete formal concepts (The Source code is available under https://www.codeberg.org/thille/ic-gid ).

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

Incomplete Formal Contexts and Their Intrinsic Dimension

  • Tobias Hille,
  • Tom Hanika

摘要

While the notion of geometric intrinsic dimension has been studied for complete formal contexts, its behavior in the presence of incomplete data was not investigated until now. In this paper, we investigate the geometric intrinsic dimension of incomplete formal contexts, where the object-attribute relationship is partially unknown or undefined. Specifically, we introduce a theoretical upper bound on the intrinsic dimension of the completion of such contexts, derived from so-called stable certain concepts. These are defined by well-known patterns of incompleteness from the literature. We validate our theoretical results through computational experiments and discuss their implications. We also show that certain concepts can be a universal tool for analyzing incomplete formal concepts (The Source code is available under https://www.codeberg.org/thille/ic-gid ).