Main Ongoing Projects
Semantic Annotation of Web Data
This project aims to better exploit valuable information that can be found in vast amounts of Web data (e.g., Web pages, digital library contents, social media). These data are usually unstructured or semi structured and lack well defined semantics. The goals of our research group in this project are: (i) selection and adaptation of tools and methods for semantic annotation of Web data with linked data and controlled vocabularies; (ii) improve semantic annotation results via ensembling of distinct methods, use of contexts for disambiguation and semantic expansion, and gradual automatic accumulation of experience about the annotation process; (iii) derive new annotations from existing ones based on contexts and acquired experience; and (iv) employ the resulting semantically annotated data to better realize application tasks such as semantic search, information analysis (e.g., in data warehouses), sentiment analysis and recommendation.
This subproject aims to explore the recognition and extraction of relevant words particularly in a mass of tweets collected by using the Twitter API. The specific goals for this subproject are: (i) selection and adaptation of tools for automatic morphosyntactic annotation; (ii) experimentation with these tools for identifying relevant words with particular morphosyntactic classes such as verbs and nouns; (iii) link these words to synsets with specific meanings described in controlled vocabularies.
PLATINUM - A Platform for NoSQL Databases Integrated Search
Start: 03/2016 Host: INE/UFSC
The goal of this research project is to propose a software platform that enables the systematic development of techniques aimed at the merger, integration and search of data present on NoSQL data sources in the cloud. In this way, search integrated data in a certain knowledge domain will be possible, preventing users and applications know the particular layout of each data source of interest, formulate specific queries to each of them and manually integrate the results.
SEEK - Semantic Enrichment of Trajectory Knowledge Discovery
The goal of the project is to investigate methods to extract meaningful knowledge from large amounts of movement data, by defining techniques for an advanced semantic-rich knowledge discovery process. The particular goals of INE/UFSC in this project are: (i) develop conceptual models for semantic enrichment and analysis of movement data; (ii) develop methods for knowledge extraction and semantic enrichment of movement data; (iii) contribute to the development of movement data warehouses.
The goal of this project is to develop a Geographical Information System (GIS) for Santa Catarina State. This system will support the collaborative gathering, annotation, management, interchange, and handling of geographical information about the state. It will include high resolution photographic data of the whole territory and several information layers, including terrain, hydrology, and the official layers CONCAR, INDE, and ANA. The database will be extensible in the intentional and the extensional levels. New layers can be defined by users of authorized organizations to collect new data or publish visions of their available data for other organizations or the public in general. These data, associated metadata, and the data exchange means supported by SIG@SC will be in conformance with OGC standards, as well as some national, state, and institutional standards. The intention is to make SIG@SC a tool to leverage governmental and community actions with facilitated access to quality integrated information.
Software Platform to Foster and Continually Update Good Inspection Practices in Hemotherapy Services
The goal of this project is to develop software and contents to improve the inspection instruments for monitoring sanity risk of the Brazilian services of hemotherapy. The inspection instrument itself (modules, section, and items of inspection) and the guidelines for making proper inspections are all represented as data and knowledge. They can be edited and enriched with sophisticated tools and application on the Web, to support inspections with state-of-the art procedures, on the Web or by using mobile devices. The collected information about inspections performed across the country, feed a decision support system for Anvisa and the the Brazilian Ministry of Health. The system is being developed to be easily customized for inspecting a variety of things. Hemotherapy services and restaurants are the first case studies.
Some Previous Projects
Context and Contents Similarity for Efficient Retrieval of Complex Data
This research project developed and started to test an approach to efﬁciently execute conjunctive queries on big complex data together with contextual conventional data. The database is horizontally fragmented according to criteria frequently used in query predicates. The collection of fragments is indexed to efﬁciently ﬁnd the fragment(s) whose contents satisfy some query predicate(s). The contents of each fragment are then indexed as well, to support efﬁcient ﬁltering of the fragment data according to other query predicate(s) conjunctively connected to the former.
Semantic search with context information
This project resulted in the creation of a prototype called Praesto, a semantic search system that keeps track of the ontological context of the user as he/she poses keyword-based queries and browses the returned results. The context expresses the particular user's view and preferences for different denotations of keywords, which are described in an underlying ontology. It enables disambiguation and semantic extension of queries, in order to automatically generate search results that are semantically related to the user's preferences.