X hits on this document

PDF document

Vedecká grantová agentúra Ministerstva školstva SR a Slovenskej ... - page 4 / 14





4 / 14

Information Resources in Organizations and their Processing

Appropriate information resources processing in an organization is the most important part in the process of introducing knowledge oriented technologies into organizations. The processing depends first of all on the type of information resource, the knowledge application model as well as on the problem, which the resource processing results should help to resolve. Fundamental information resources in organizations currently include documents, electronic communication and databases. Most of these resources are able to be transformed to text that can be followed by application of methods for text processing in an effect to acquire necessary information and knowledge.

Over the world information resource processing is the focus of many research institutions; however, this research is far from being finished. The English language processing is handled quite well, however in our conditions it is important to deal with information resource processing in the Slovak language that is still at the beginning. Another open challenge is appropriate use of electronic communication which does not have to serve solely as an information and knowledge resource but also for creation models of tasks, actors and resources within an organization.

Information retrieval from text documents that influences knowledge system development has a great importance nowadays. Therefore, the number of electronic documents is permanently growing. From the year 1950, when the electronic documents in digital libraries were placed, methods using the document representation by Boolean model were introduced [5]. This was the first attempt to create an information retrieval system. Then emerged further document representations (described below), which enabled creation of new more effective information retrieval systems. The neural networks were also applied in these systems [17, 18, 22, 24].

The information retrieval system from text documents in a natural language on the basis of user queries enables to retrieve relevant documents from the document collection. For the information retrieval system it is needed to define the query definition manner (words, pattern forms and structured queries [5], the document representation manner and the relevant document retrieval system. The documents may be represented by indexation, inverted index, semantic nets and ontology, on the basis of intelligent agents and semantic web [7, 26, 27] etc.

Information retrieval models can be divided into three main groups [5], namely boolean and fuzzy models [5, 4, 15], vector models [6, 12] and probabilistic models [3, 20]. Currently the information retrieval research inclines to document representation via ontologies [26, 27], in particular in the semantic web [7].

Recently the research field of information retrieval has begun using neural networks [11, 17, 24]. They are applied anywhere, where there are no strict data given (different modifications of words), which complicate the information retrieval from the text documents and where is a need to solve the classification or prediction problem of processed, stored and retrieved collection of documents. These are supervised [17, 24] and unsupervised [13, 18, 23] feed-forward neural networks, recurrent neural networks and hierarchical neural networks [22].

On the basis of foundations of given research field it can be mentioned that the neural networks have their place in the text document retrieval systems, because they create the prerequisites to simplify methods for the language analysis of the text and they provide effective tools to access the documents. They also found their place in hybrid approaches of representation and processing of the documents by neural networks and ontology [11, 22].

Every published method was used for the English language, where the words in the documents were processed by Porter’s algorithm [21]. For the Slovak language such algorithm is very complicated, because the Slovak language is very flexible and inflexive endings. Papers discussing the problem of the information retrieval in the Slovak language were published [9], while they use a linguistic approach, which is too complicated. Therefore, to address the partial problems in the Slovak language analysis, the neural networks can be used.

Availability, Management and Sharing of Information Resources

In order to use information and knowledge in an organization properly, it is necessary to ensure availability of information for users as well as interconnection with existing systems. A way of interconnection with existing and external systems is the use of web services seeing intensive research in the area of their extension to semantic web services [38] and service oriented architectures [10]. This can consequently ensure sharing of necessary information and providing of services among organization departments as well as external organizations (suppliers and customers). The right use of information and knowledge must be backed by systems for sharing knowledge among system users (found for example in projects Pellucid; DECOR that prepared context sensitive knowledge form archived documents; or SmartGov the stored and provided knowledge in context [32]) on an organizational level as well as support of cooperation among them on the basis of e-Collaboration solutions [33] for instance in combination with solutions based on electronic communication enabling active knowledge presentation on the basis of the user context.

Another significant part is knowledge and information maintenance in an organizational memory. The knowledge maintenance is tied with presentation of formalized knowledge to users. A knowledge maintenance interface can be built on approaches based on object trees (ontology editors such as Protégé [34] or OilEd). This approach is mostly understandable to experts who understand the ontology. Other methods can be graph-based since most representations of formalized knowledge such as DAML, OWL [26] or RDF [27] use a graph structure. Examples are UML, Ontoviz Tab [35] in the Protégé editor or Cartoo Technologies [36] products. For knowledge maintenance interfaces also solutions based on XSL [37] transformation can be used. The knowledge maintenance should also support user feedback. Mechanisms such as reviewing, evaluating, voting or ranking on presented knowledge need to be incorporated.


Document info
Document views28
Page views28
Page last viewedSat Oct 22 20:33:21 UTC 2016