Minimum spanning tree with missing data


How are missing data treated when calculating a Minimum Spanning Tree? Is it tolerated or used as a zero value?


In principle, a Minimum Spanning Tree cannot deal with missing values and they are replaced by a zero value in practice. In case of MLST or equivalent techniques, this means that missing values become an additional state "zero". This approach is justified as long as the relative number of missing values is low.


Applicable for: 
Version 3.5 - 7.6


Share this: