Data Mining for High Productivity Information System

When dealing with large amounts of data it might be useful to either predict some trends or to figure out patterns in existing data.
Investigate existing data mining algorithms – we would like to use stat-of-the-art version of really basic algorithms for classifiers, decision trees and clustering. Other options are also possible when suitable.
Analyze the possibilities usage of data mining algorithms in an information system.
Design a way how to bring the power of data mining algorithms available to basic users without any deep technical knowledge. The users should be able to use the algorithms to for example predict some trends in data or to figure out patterns in their existing data.
The existing data are unlikely to be compatible with the input types required for data mining algorithms. The data will be fulltext, ordinal or floating point numbers, dates etc. The proposed system should automatically try to pre-process these types of data in the best possible way.
Demonstrate the proposed design by implementing a sample application leveraging open-source platform Lumeer.
It should be possible to integrate the resulting application into Lumeer later. This requires the backend to be based on Lumeer backed or to use some REST mock service. The frontend should be developed in Angular 2.0/4.

You must be logged in to perform this action!

Martin Večeřa

Location: Brno