An Interactive Interface for Multi-dimensional Data Stream Analysis

Analysis of a token density metric for concern detection in Matlab sources using UbiSOM

“…Figure shows a U‐Matrix for the metrics data fed to the UbiSOM, from the example used in Section 6.2 to illustrate and summarize the analyses. In general, well‐separated regions (i.e., the blue “lakes” of Marques et al, ) in the U‐Matrix represent clearly separated patterns in the data fed to the UbiSOM. Individual lakes are semi‐automatically labelled using a derivation of the method described in Matos et al ().…”

“…As previously explained, the average values in both the U‐Matrix and the CPs are represented as a JET colour map (Marques et al, ), with red representing high values, blue representing low values (down to zero), and green representing intermediate values. In addition (and consistently), green areas with a blue hue represent values lower than pure green, whereas “dirty” green areas with a red hue represent values higher than pure green.…”

“…We previously developed multiSOM (Marques, Silva, & Santos, ), an interactive visualization tool that uses the UbiSOM as the underlying data mining algorithm. For the purposes of the analysis reported in this article, the previously available tool received a few improvements, including commands providing new visualisation functionalities.…”

“…In general, well-separated regions (i.e., the blue "lakes" of Marques et al, 2016) in the U-Matrix represent clearly separated patterns in the data fed to the UbiSOM. Individual lakes are semi-automatically labelled using a derivation of the method described in Matos et al (2015).…”

“…As previously explained, the average values in both the U-Matrix and the CPs are represented as a JET colour map (Marques et al, 2016), with red representing high values, blue representing low values (down to zero), and green representing intermediate values.…”