In line with the visual line of approach applied at GRIS, we support analysts conducting Data Mining approaches in a visual way.
Information Visualization helps not only to visualize Data Mining results, but can also be used to make the entire Data Mining process more transparent, referring to as “open the black box”. Visual Analytics particularly brings together the complementary strengths of humans and machines. Visual Analytics achieve a tight coupling of Data Mining (and Machine Learning) models with visual-interactive interfaces.
Similarly, Interactive Machine Learning combines visualization with algorithmic models for learning and analyzing data characteristics, including techniques from Data Mining.
With these research endeavors, Data Mining not only becomes a visual but also an interactive process, with the ability to have the human-in-the-loop. As such, our perspective on Data Mining is in line with innovative approaches to address challenges related to Data Science.
Data Mining explained
Data Mining is the systematic application of statistical methods on large data sets to detect patterns such as clusters, outliers, trends, or relations between data instances. These patterns build the basis for the generation of knowledge from data. Data Mining can support confirmatory or exploratory data analysis. The first supports validating/rejecting given hypotheses, whereas the latter supports ill-defined information needs seeking for novel insights from data that may contribute to the hypotheses- generation process.
With the size of today’s data sets (“big data”) algorithmic support in the data mining process is indispensable. In addition, visualization as become an integral part of the data mining process: not only for the visualization of results, but also as a means to support the entire data mining process starting from raw and “dirty” data.