Resumo em Inglês:

Due to the imprecise nature of biological experiments, biological data is often characterized by the presence of redundant and noisy data. This may be due to errors that occurred during data collection, such as contaminations in laboratorial samples. It is the case of gene expression data, where the equipments and tools currently used frequently produce noisy biological data. Machine Learning algorithms have been successfully used in gene expression data analysis. Although many Machine Learning algorithms can deal with noise, detecting and removing noisy instances from the training data set can help the induction of the target hypothesis. This paper evaluates the use of distance-based pre-processing techniques for noise detection in gene expression data classification problems. This evaluation analyzes the effectiveness of the techniques investigated in removing noisy data, measured by the accuracy obtained by different Machine Learning classifiers over the pre-processed data.

Resumo em Inglês:

The quest for interoperability is one of the main driving forces behind international organizations such as OGC and W3C. In parallel, a trend in systems design and development is to break down GIS functionalities into modules that can be composed in an ad hoc manner. This component-driven approach increases flexibility and extensibility. For scientists whose research involves geospatial analysis, however, such initiatives mean more than interoperability and flexibility. These efforts are progressively shielding these users from having to deal with problems such as data representation formats, communication protocols or pre-processing algorithms. Once scientists are allowed to abstract from lower level concerns, they can shift their focus to the design and implementation of the computational models they are interested in. This paper analyzes how interoperability and componentization efforts have this underestimated impact on the design and development perspective. This discussion is illustrated by the description of the design and implementation of WebMAPS, a geospatial information system to support agricultural planning and monitoring. By taking advantage of new results in the above areas, the experience with WebMAPS presents a road map to leverage system design and development by the seamless composition of distributed data sources and processing solutions.

Resumo em Inglês:

Semantic portals are characterized for storing and structuring content according to specific domain ontologies. This content is represented through ontological languages, which enable not only adding semantic value to information treatment, but also inferring new knowledge from it. Publication in a semantic portal is typically done by instantiating its ontology, and this is often performed manually or through the use of specific forms. However, in order to keep portals constantly up-to-date, it is necessary to provide means for a more dynamic publication, integrating the portal content with information retrieved from different ontology-based sites on the same or on complementary domains. Reusing information from different ontologies requires specific and efficient mechanisms to align them, taking into account syntactical and semantical conflicts. This paper proposes an extension of the Crosi Mapping System, a matching mechanism which calculates similarities between ontologies. Some of its original algorithms have been enriched with additional functionality. This extension, named e-CMS, has been evaluated using the OAEI ontology alignment benchmark, and results show an increase of 69% in alignment precision when compared to the CMS original version. In order to illustrate its use, the e-CMS strategy was applied to SiGePoS, a System for Generating Semantic Portals. The semantic module, one of the system components, implements the alignment mechanism between ontologies, which is performed by the e-CMS.

Resumo em Inglês:

Mobile phones are a widespread platform for ICT applications because they are highly pervasive in contemporary society. Hence, we can think of mobile gaming as a serious candidate to being a prominent form of entertainment in the near future. However, most games (for computers, console and mobile devices) make extensive use of the visual medium, which tends to exclude visually-impaired users from the play. While mobile gaming could potentially reach many visually-impaired users, who are very familiar with this technology, currently there seems to be only very few alternatives for this community. In an attempt to explore new interactive possibilities for such users, this work presents an initial study on non-visual interfaces for mobile phone games. It is based on Semiotic Engineering principles, emphasizing communication through aural, tactile and gestural signs, and deliberately excluding visual information. Results include a number of issues that can be incorporated to a wider research agenda on mobile gaming accessibility, both for the visually-impaired and sighted.