3 Replies Latest reply: May 29, 2016 12:28 PM by Marco Wedel RSS

    Decisiontree visualisation with scatter plot

    Petra Zimmer

      Hello everyone,

       

      I'm trying to visulize my Decision Tree. I would prefer something like this: https://farm3.staticflickr.com/2942/15194160150_c4e1d356bd_z.jpg

       

      I have many, many fields like this %StationYearKey;station_longitude_deg;station_latitude_deg;station_altitude;Pop_Dichte;Yield (100 kg/ha);Harvested production (1000 t);Area (cultivation/harvested/production) (1000 ha);Main area (1000 ha);milk;goats;sheeps;pigs;buffaloes;cows and bovine;Wetland;Water;Bare land;Grassland;Shrubland;Woodland;Cropland;Artificial land;O3_class

       

      but perhaps my final decision tree uses only 10 attributes.

       

      I search for a easy way to get from my model with looks like this:

      Learned classification tree model:

      DecisionTreeModel classifier of depth 6 with 103 nodes

        If (feature 0 <= 0.072434549052584)

         If (feature 7 <= 32.81)

          If (feature 4 <= 448.24)

           If (feature 6 <= 18.19)

            If (feature 6 <= 0.0)

             If (feature 5 <= 1364.1)

              Predict: 1.0

             Else (feature 5 > 1364.1)

              Predict: 0.0

            Else (feature 6 > 0.0)

             Predict: 2.0

           Else (feature 6 > 18.19)

            Predict: 0.0

          Else (feature 4 > 448.24)

           If (feature 7 <= 15.66)

            Predict: 2.0

           Else (feature 7 > 15.66)

            If (feature 4 <= 1729.98)

             Predict: 1.0

            Else (feature 4 > 1729.98)

             If (feature 3 <= 842.7)

              Predict: 1.0

             Else (feature 3 > 842.7)

              Predict: 2.0

      ....

       

      to a nice and easy visualisation, because I have to do it at least 5 times.

       

      Hoping for inspiration for data modelling and visualisation tips:)