Monday, 26 November 2018

Handwritten Digit Recognition Project in Machine Learning

The dataset which we are going to use is made up of 1797 8𝗑8 images. Each image is of a hand-written digit.

In order to utilize an 8𝗑8 figure, we have to first transform it into a feature vector with length 64.

We will be using Super Vector Machine (SVM).

Plot 8𝗑8 image

# Import dataset
from om sklearn.datasets import load_digits

# Import the sklearn for SVM
from sklearn import svm

digits = load_digits( )

# Each datapoint is a 8𝗑8 image of digit
# Plot the image
plt.gray( )
plt.matshow(digits.image[20])
plt.show


Print the 8𝗑8 array which represents each pixel

print digits.images [20]
' ' '

Output :-
[ [  0.    0.     3.    13.    11.    7.     0.    0.]
  [  0.    0.     11.   16.    16.  16.     2.   0.]
  [  0.    4.     16.    9.      1.    14.    2.   0.]
  [  0.    4.     16.    0.      0.    16.    2.   0.]
  [  0.    0.     16.    1.      0.     12.   8.   0.]
  [  0.    0.     15.    9.      0.     13.   6.   0.]
  [  0.    0.      9.    14.     9.     14.   1.   0.]
  [  0.    0.      2.    12.     13.    4.    0.   0.]
' ' '

Train and Test the model

clf = svm.SVC( )
#  Train the model
clf.fit (digits.data[:-1], digits.target[ :-1])

# Test the model
  prediction = clf.predict(digits.data[20:21])

print  "Predicted Digit  ->" , prediction
' ' '

Output :
Predicted Digit  ->  [0]
' ' '

No comments:

Post a Comment

Popular Posts