Looks interesting and different / orthogonal to other ML approaches.
http://arxiv.org/abs/1204.1995
During the early 1980s, the mathematical methodology of Formal Concept Analysis (FCA)
emerged within the community of set and order theorists, algebraists and discrete math-
ematicians. The aim was to find a new, concrete and meaningful approach to the un-
derstanding of complete lattices (ordered sets such that for every subset the supremum
and infimum exist). The following discovery proved fruitful: Every complete lattice is
representable as a hierarchy of concepts, which were conceived as sets of objects sharing
a maximal set of attributes. This paved the way for using the field of lattice theory for a
transparent and complete representation of very different types of knowledge.
Originally FCA was inspired by the educationalist Hartmut von Hentig [99] and his
program of restructuring sciences aiming at interdisciplinary collaboration and democratic
control. The philosophical background traces back to Charles S. Peirce (1839 - 1914), who
condensed some of his main ideas to the pragmatic maxim:
Consider what effects, that might conceivably have practical bearings, we con-
ceive the objects of our conception to have. Then, our conception of these
effects is the whole of our conception of the object. [78, 5.402]
In that tradition, FCA aims at unfolding the observable, elementary properties defining
the objects subsumed by scientific concepts. If applied to temporal transitions, effects
of specific combinations of state attributes can be modelled and predicted in a clear and
concise manner. Thus, FCA seems to be appropriate to describe causality – and the limits
of its understanding.
At present, FCA is a well developed mathematical theory and there are practical appli-
cations in various fields such as data and text mining, knowledge management, semantic
web, software engineering or economics [36]. The main application of this thesis is re-
lated to molecular and systems biology. Due to the rapid accumulation of data about
molecular inter-relationships, there is an increasing demand for approaches to analyse the
resulting regulatory network models (for a short introduction see Section 7.1, an example
is represented in Figure 8.4). Therefore, we developed a formal representation of pro-
cesses, especially biological processes. The purpose was to construct knowledge bases of
rules expressing temporal dependencies within gene regulatory (or signal transduction and
metabolic) networks.
