Search code examples
machine-learningwekaarff

weka explorer incorrect output after loading csv file


Any idea why am I getting this output?

I am using below dataset. http://archive.ics.uci.edu/ml/machine-learning-databases/auto-mpg/auto-mpg.data


Solution

  • Colors are a representation of the distribution of carname in each column of the histogram.

    Apparently, it behave like this because of the number of values in this attribute, which is a nominal attribute (I see it above the histogram) with 305 distinct values: there would be too many distinct colors, and each color would represent only one instance. Weka fails to represent all those colors (I suppose it doesn't have so many colors saved).

    By the way, as Anony-Mousse said in his comment, ignore the colors: carname should be a string attribute, not a nominal one, and if you convert it to string attribute, colors will disappear, because watching its distribution doesn't make sense.