Titulo Estágio
Semantic-based Platform for Knowledge Extraction in the European Epilepsy Database
Área Tecnológica
Informática Médica
Local do Estágio
DEI
Enquadramento
An exciting application of the seizure prediction technology is its potential use in therapeutic epilepsy devices to trigger intervention before a crisis can occur. One important step for this goal is the development of a platform for knowledge extraction using semantic mining for data of epileptic patients stored in huge (web) data bases. The rationale behind is a search engine capable of integrating Semantic, Linguist and Contextual Epilepsy Knowledge to improve the robustness and precision of the search process. The system should be able not only to answer queries based on simple keywords (type of epilepsy, type of signals available, e.g. EEG, epilepsy ontology, etc.) but also questions such as: What is the relationship between a given type of epilepsy and a selected group of features? In the positive case, with which probability does it occur? Which classifier (among the previously ones stored) is the best for a certain patient? The inference engine raises several challenging algorithmic issues.
The prototype to be developed should be able to analyze epileptic patient records, generate a response to the Health-personnel using the system, and a report to help him/her to take appropriate decisions before crisis occurrence (e.g. ministration of drug, etc.). The analysis should include: topic extraction, summarization, identification of patient crucial data, etc.. The prototype should be implemented and tested.
Objetivo
In this project, the goal is to develop a platform for knowledge extraction using semantic mining in epileptic data bases. One important component is the construction of a large epileptic patient record containing health status information such as: EEG, ECG, 24 features per channel (w/ large number of channels), clinical information (type of epilepsy, drug medication, etc.), classifiers algorithm (SVM , MLNN, RBF, NN, etc.) . A second component is the integration with the Epilepsy Ontology (already available). This ontology provides a common understanding of the Epilepsy domain that can be shared (and communicated) between people and/or machines. It makes an explicit conceptualization that describes the semantics of the epilepsy data with the intuition of reasoning about that data. A third component is a natural language processing module. There are frameworks available for human language processing composed by several components and plug-ins that cover a large portion of the natural language processing tasks. Finally, we aim at the development of an inference engine for knowledge extraction from all the available (structured) data.
For this project the prospective student should develop a system able to:
- Load patients and data in a Epileptic Multimodal Data Base including all the needed data for mining,
- Analyze Epileptic Patient records,
- Generate Responses to Health-personnel.
- Generate Reports for Decision Making
The student will develop a prototype by following the background available in the team, specifically the Ontology already available. Furthermore, he/she will perform a large experimental analysis by following an appropriate methodology according to the literature (and technologies) available.
This work will be performed within the FP7 European project EPILEPSIAE- Evolving Platform for Improving Living Expectation of Patients a leaded by ACG/CIUSC, Portugal (see http://www.epilepsiae.eu). There will be funding available within the research group that is the leading partner of this project, depending on the curriculum and disponibility of the candidate.
Plano de Trabalhos - Semestre 1
1 Month: Literature review: Multimodal (and heterogeneous) data bases, Semantic Mining, Inference Engines
2 Months: Completion of the database with features and classifiers already available
0,5 Month: Prototype Specification and Requirements
0,5 Month: Thesis Proposal Writing
Plano de Trabalhos - Semestre 2
1 Month: Prototype Development
1 Month: Prototype Testing: Knowledge Extraction Platform with Semantic Mining for Epileptic Patients Records
1 Month- Prototype Experimentation and Results analysis
1 Month: Thesis writing
Condições
The ACGLab and its computing resources are available for performing experiments. The Epilepsy database is available to be completed and fullfilled
Observações
The student will work in a very highly motivated research group with wide international visibility. He/she will strengthen his abilities on the design and analysis of algorithms and data structures for problem solving and will develop advanced programming skills. This type of work is appropriate to students that want to follow a professional career in advanced database and data mining companies, and also for students aiming an academic career in design and analysis of algorithms and in theoretical aspects of Computer Science.
Students with excellent record on Algorithms and Data Structures, Data Bases,Software Engineering, Web Semantica, Adaptive Computation, are highly preferable. Good experience in Java and Oracle.
Orientador
Prof. Doutora Bernardete Martins Ribeiro
bribeiro@dei.uc.pt 📩