The Parallel DBMS group
at UH focuses on developing scalable and parallel algorithms to analyze
large data sets with machine learning and statistical models (e.g. clustering,
regression, dimensionality reduction, variable/feature selection, time series),
cubes (ad-hoc queries, decision support systems) and graphs
(paths, cliques, vertex neighborhood).
In the past we worked on sequential data mining algorithms and
parallel row DBMSs with applications in corporate data warehouses and
Our approach is
radically different from most database research since
our techniques are developed on
how to integrate algorithms with parallel DBMSs having column and
array based storage. Our work also includes
R language and streams.
Parallel algorithms (machine learning, graphs, cubes).
Analytics inside parallel DBMSs and Hadoop (MapReduce, Spark).
Eliminating RAM and parallel processing limitations from math packages (R and Matlab).
Query optimization: recursive queries, cubes, skylines, pivoting.
Semi-structured data: text, web pages, documents, ontologies.
Software engineering: ER modeling, data quality, debugging source code.
Applications: medicine, bioinformatics, corporate data warehouses, physics, network monitoring.
Dr. Carlos Ordonez
Department of Computer Science
University of Houston
Houston TX, 77204
firstname AT central DOT uh DOT edu