Structural Feature Based Anomaly Detection for Packed Executable Identification - Computational Intelligence in Security for Information Systems - page 252

control rules but it is important for him that data are in the right place, protected
by unauthorized accesses and can be easily retrieved when it is needed. So the
data confidentiality is strongly related to the structure of data itself and to the
physical support. Although producing data without following structural rules
can be faster, as time goes by, it can result very dicult to maintain consistency
of not structured data as well as to query and process not structured data. At
this aim we are exploiting the adoption of semantic techniques to analyze texts
and automatically extract relevant information, concepts and complex relations;
the final goal is to be able to format not structured information in a struc-
tured way and with automatic or human-assisted tools. Once structured data
and resources, it is possible to associate to each of them different protection
mechanisms. In previous works we proposed a methodology for the classification
and the protection of sensitive data, able to retrieve and associate the security
rules to the resource to protect. It is based on semantic approaches for relevant
concepts identification in textual data by means of lexical-statistical techniques
[6,7]. The main idea of our work is to design a reconfigurable framework for
documents processing that accepts in inputs a collection of heterogeneous data,
including textual and multimedia, belonging to specialist domains and provides
semiautomatic procedures for structuring data and extracting information of in-
terest. Each domain has in common the activity of knowledge extraction from
texts that includes different kinds of text analysis methodologies. The state of
the art in this field is related to techniques of NLP and to cross-disciplinary
perspectives including Statistical Linguistics whose objective is the study and
the analysis of natural language through computational tools and models. The
framework that we propose is based on transformation rules that are strongly
related to the application domain, we were able to locate a set of tools, tech-
niques and methodologies that can be adopted in different contexts and can be
easily personalized and tuned. Once structured, the ”monolithic” documents can
be seen as a set of resources that can be separately accessible; we will represent
them as XML files and we can easily enforce fine-grained access control policies
with available security enforcer modules as XACML [12]. We will illustrate the
applicability of the framework with a real case study to structure and protect
e-health documents.
The reminder of the paper is structured as follows: in Section 2 we will illus-
trate the document processing framework for data transformation; in Section 3
we propose a detailed example of the framework implementation for the security
domain, in Section 4 a simple case study will be presented to put in evidence
how the framework can be used for medical record structuring and protection.
Finally, in Section 5 some conclusion and future work are drawn.
2 A Model for Document Processing Framework
In several contexts as medical, juridical, humanistic, knowledge management deal-
ing with acquiring, maintaining, and accessing knowledge within data can improve
public and private services providers. Diculties arise when the informations are