Reducing Snapshots to Points: A Visual Analytics Approach to Dynamic Network Exploration by Elzen et. al. Best Paper of VAST 2015.
Binarization, 1 is assigned to each cell with a value Mi, jk > 0 and 0 otherwise.
Min-max normalization scales the data to a fixed range [0, 1]. Min-max normalization leads to smaller standard deviations, thus
suppressing the effect of outliers.Z-normalization transforms the data to have zero mean and unit variance by subtracting the mean and dividing by standard deviation.
For a linear dimensionality reduction method, such as PCA,
z-normalization is the preferred choice, since we are interested in
the components that maximize the variance.