I have 5 variables nu, gamma, %train error, %test error and %validate error. I have nearly a set of nearly 1000 values for the % errors for a set of nu and gamma.
I wish to plot (line preferably) the %train error for a set of nu- as x-axis and gamma as y-axis.
Similarly would like to do the same of other 2 errors.
I am trying this in Python.
I have designed a 3D plot for the same but I only wish to display the error values on the plot and not the nu and gamma values.
i suggest you use heat-map, you can use whatever you like as x and y(nu & gamma in your case) values and use colours and their intensities to show the third value(%train erro), please have look at these
hope it helped!
Thanks a lot. I had a look at both the links you provided. But they did not prove very helpful. As, I find difficulty in interpreting the heatmaps.
Is there no way of plotting a scatter for 3 errors but wrt nu as x-axis and gamma as y-axis? Though I can understand I am trying to visualize 3 infact 5 variables in a 2D graph. Though for different errors we can provide different colors and markers for differentiation, so if I can do it for 3 (nu, gamma and train error) then probably I can do it for all 5 too.
Hi Renu,
I would first define the two classes... one for a point and another for a polyline such as like this:
import math
class Points:
def __init__(self, x=0.0, y=0.0):
self.x,self.y = x, y
class Polyline:
def __init__(self, points =[] ):
self.points = points
def getLength(self):
i = 0
length = 0.0
while i < len(self.points)-1:
length += math.sqrt((self.points[i+1].x -self.points[i].x)**2 + (self.points[i+1].y - self.points[i].y)**2 )
i += 1
return length
Define the points and/or lines you'd like to use. Then, you can draw them using the Tkinter or graphics packages.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Programming Languages
- Python