But the area/perimeter ratio appear to be approximately equally spaced integers!? What is on the vertical axis? Is that just a classification code? Is the classification perfect (the colours and vertical placement line up perfectly)?
Thanks for clarifying this. Like Andrew, I thought the graph was output from the ANN classifying the different genomes (are they "genomes"?) with perfect classification!
Another way to summarize this kind of 1D statistic across categories is a bar graph of means with a variance "error bar", or a "box plot".
I like the area/perimeter plot but I don't know what it means... Can you explain more?
ReplyDeleteIt is the ratio of Area divided by Perimeter (for all data I had copied for those cells).
ReplyDeleteI figured although the size changed the shape was pretty much identical, so the ratio's should be similar or close together for each class.
But the area/perimeter ratio appear to be approximately equally spaced integers!? What is on the vertical axis? Is that just a classification code? Is the classification perfect (the colours and vertical placement line up perfectly)?
ReplyDeleteI rounded them to integers to have a better "number line" format.
ReplyDeleteThe vertical placement is just so they weren't overlapped on top of each other, to make it easier to distinguish.
So the classification isn't perfect, was just to make it easier to read.
Thanks for clarifying this. Like Andrew, I thought the graph was output from the ANN classifying the different genomes (are they "genomes"?) with perfect classification!
ReplyDeleteAnother way to summarize this kind of 1D statistic across categories is a bar graph of means with a variance "error bar", or a "box plot".
Did you have any luck with the Cintiq tablet?