doc. Ing. Jiří Kléma, Ph.D.

Free text is unstructured information difficult to be searched. In medical texts, it is a frequent task to find named entities (diseases, risk factors, genes, proteins) and assess their relationships and interactions (correlation, causal relationship, interaction denial). This topic assumes namely sequential analysis that understands text as a sequence of tokens (words). At the same time, a representation that will facilitate efficient sequential search shall be proposed. Additional information can be found at ida.felk.cvut.cz/klema.

Data mining is an open problem namely for the sake of scale and heterogeneity of the mined data, both interdomain and intradomain. Incremental procedures frequently get applied, Extensive but often noisy measurements originating from industrial devices, business processes or live organisms are automatically matched with the existing knowledge, the knowledge regularizes the measurements and measurements get used to update the knowledge. The topic aims at proposal of dedicated efficient algorithms for knowledge-driven data mining.

Knowledge discovery is an open problem namely for the sake of scale and heterogeneity of the mined data, both interdomain and intradomain. Incremental procedures frequently get applied, Extensive but often noisy measurements originating from industrial devices, business processes or live organisms are automatically matched with the existing knowledge, the knowledge regularizes the measurements and measurements get used to update the knowledge. The topic aims at proposal of dedicated efficient algorithms for knowledge-driven data mining.

Many biological and industrial processes generate sequential or time series data. Their analysis, modelling and prediction represent an important research topic. In here, we assume namely genomics as the field of application (DNA, amino acids). The tasks can be the search for homological sequences, segmentation or in contrast sequence assembly from short reads. Nevertheless, the topic is open towards industrial projects (fault prediction in an industrial production line, etc). Additional information can be found at ida.felk.cvut.cz/klema.

Many biological and industrial processes generate sequential or time series data. Their analysis, modelling and prediction represent an important research topic. In here, we assume namely genomics as the field of application (DNA, amino acids). The tasks can be the search for homological sequences, segmentation or in contrast sequence assembly from short reads. Nevertheless, the topic is open towards industrial projects (fault prediction in an industrial production line, etc).