Scribd
Upload
44 views

Overview of Parallel Coordinates, Visualizing Neural Network and Visualization of Trees

Parallel coordinates is a method for visualizing multivariate data by plotting each variable along its own axis and data points as polylines across the axes. It allows for high-dimensional data to be visualized on a two-dimensional plane. Visualizing neural networks can be done using link weights represented by line thickness or a Hinton diagram where link weights are shown as square sizes and colors. Trees are often visualized as linear tree diagrams with nodes and lines, and boxes at branch points show information like record counts and error estimates. Zooming functionality allows viewing the entire tree or sections.

Uploaded by

NahidaNisar
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
44 views

Overview of Parallel Coordinates, Visualizing Neural Network and Visualization of Trees

Parallel coordinates is a method for visualizing multivariate data by plotting each variable along its own axis and data points as polylines across the axes. It allows for high-dimensional data to be visualized on a two-dimensional plane. Visualizing neural networks can be done using link weights represented by line thickness or a Hinton diagram where link weights are shown as square sizes and colors. Trees are often visualized as linear tree diagrams with nodes and lines, and boxes at branch points show information like record counts and error estimates. Zooming functionality allows viewing the entire tree or sections.

Uploaded by

NahidaNisar
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 9

Overview of Parallel Coordinates,

Visualizing Neural Network and


Visualization of Trees.
NEED OF PARALLEL COORDINATES

when we try to map three or more than three
dimensions into the two dimensions of the computer
some data is obscured by other data values because
one dimension is drawn either into or out from the
screen.
PARALLEL COORDINATES:
Parallel coordinates is a common way of visualizing high-
dimensional geometry and analyzing multivariate data.

In case of parallel coordinate we draw all the dimensions parallel to
each other instead of drawing them perpendicular.


100 $100,000 3(green)
50
0
$50,000
2(brown)
0 0
1(blue)
Age Income Eyes
EXAMPLE:

Mapping of eye color onto numeric values-
unknown=0,blue=1,brown=2,green=3.
Then a customer record could be described by three
values such as 28,40,2 which means a 28 yrs old man
making $40,000 per yr and having brown eyes.
VISUALIZING NEURAL NETWORKS
Neural network is an interconnected
group of nodes( node: corresponds to
neuron in human brain.)

Visualization of neural network is done
by:
1. Using link weight
2. Using Hinton diagram


In case of link weight the neural network
is drawn out with nodes and lines and
denoting the strength of the link by
width of the line representing it.
HINTON DIAGRAM:
In case of Hinton diagram each link
weight is represented by a square
and the size of the square
represents the magnitude of weight.

In Hinton diagram color of the
square represents the sign of the
link weight.

Each node plays a role for final
output and some of them have
more influence as compared to
other.



VISUALIZATION OF TREES
Linear tree



Trees are often visualized as 2D linear tree diagrams
with labeled dots and connecting lines. we simply
draw out the tree and splitting decisions and at each
branch point in the tree, we draw a box which
contain relevant information at that sub division
level.such information include:

1.The no of records.

2.The proportion of each predictive
value(density,histogram).

3.Some estimate of the error.
CONTINUED.
Trees can be large in size. So ways of zooming in and zooming
out are provided ,so that he can see entire tree or just one
section.

You might also like