Scribd
Upload
135 views12 pages

Linear Discriminant Analysis

Yes, you are right that there may still be some overlap even after applying LDA, reducing the accuracy of the model. A few solutions in that case: 1. Collect more training data to better learn the class distributions and increase separation. 2. Try other classification algorithms like logistic regression, decision trees, SVM etc on the LDA projected features to further separate the classes. 3. Increase the number of features - LDA projection to higher dimensions may give better separation. 4. Consider combining LDA with other dimensionality reduction techniques like PCA before applying classification. So in summary, when there is overlap after LDA - collect more data, try ensemble models, increase feature/dimension space are some approaches to further improve classification. The

Uploaded by

MohamedAbdelrazek
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
135 views12 pages

Linear Discriminant Analysis

Yes, you are right that there may still be some overlap even after applying LDA, reducing the accuracy of the model. A few solutions in that case: 1. Collect more training data to better learn the class distributions and increase separation. 2. Try other classification algorithms like logistic regression, decision trees, SVM etc on the LDA projected features to further separate the classes. 3. Increase the number of features - LDA projection to higher dimensions may give better separation. 4. Consider combining LDA with other dimensionality reduction techniques like PCA before applying classification. So in summary, when there is overlap after LDA - collect more data, try ensemble models, increase feature/dimension space are some approaches to further improve classification. The

Uploaded by

MohamedAbdelrazek
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 12

Basic

LDA
• Linear Discriminant Analysis or most commonly
known as ‘LDA’ is one of the most interesting machine
learning techniques till date.
• The idea was first coined by “Dr. Ronald Fisher” to
classify binary classes using ‘Fisher’s linear
discriminant‘ and later on it was generalized for
multiple classes as well.
• In case of a binary class problem, LDA acts as a
classifier like “logistic regression” and when it comes
to multi classes it acts as a dimensional reduction
technique like “PCA”.
• In this article I will not be talking much about how to
use LDA for a multi class problem, instead, let’s focus
on binary classes.
• However, let’s take a peak into how “LDA” and “PCA”
are different yet function the same.
Using Linear Discriminant Analysis on Binary Classification

• Step 1: Create a matrix containing the mean of each of the features (X) pertaining to a
particular class say ‘Ci’.
• Step 2: Create a scatter matrix (Si) for each class using the following formula.

• Step 3: Calculate the projection vector(W).


Steps
• Step 4: The last and final step is to project these data points in this vector (W).Here
‘Y’ represents the new projected data points which is a scalar quantity.
Explanation-
• Let’s look at the following figure of a two feature binary
classification problem.You can clearly see how the data
points are projected along the vector(the line represents
W) after applying LDA.
• I have given you the basic difference between “LDA” and
“PCA“, also I have given all the necessary steps for LDA but
the question remains, How is LDA doing what it’s doing?
• I haven’t given you the intuition which led ‘Dr.Fisher‘ to
formulate this beautiful technique. If you recall “logistic
regression” what it does is that it takes a number of input
features and represents in terms of probabilities.Well, LDA
does the same except it takes the input feature
vector(don’t get confused it means the set of features) and
represents them in terms of scalar quantity.Yes, you read
that right it converts vectors of features to scalars.Now,
these scalars can be used to separate the two classes and
very soon we will go through an example to get a better
idea of the same.But first, let’s look at the following figure.
Linear Discriminant Analysis for Binary Classification
• import numpy as np
• import pandas as pd
• import matplotlib.pyplot as plt
• from pandas import Series,DataFrame
• %matplotlib inline

• First of all lets prepare a data set

• data = DataFrame()
• X1 = Series(np.array([4,2,2,3,4,9,6,9,8,10]))
• X2 = Series(np.array([1,4,3,6,4,10,8,5,7,8]))
• data['X1']=X1
• data['X2']=X2
• data['class']=np.array(['class1','class1','class1','class1','class1','clas
s2','class2','class2','class2','class2'])
Lets plot the above data and see how it looks
Code-
• plt.figure(figsize=(10,7))
• plt.scatter(data.ix[data['class']=='class1',0],dat
a.ix[data['class']=='class1',1],color='r',marker='
*',label='class1')
• plt.scatter(data.ix[data['class']=='class2',0],dat
a.ix[data['class']=='class2',1],color='g',marker='
o',label='class2')
• plt.legend(loc='best')

• Scatter Plot
Lets plot the above data and see how it looks
Using Step 2 to find the scatter mattrices of each
Using Step 1 to calculate the mean matrix
class and also the within class scatter matrix
• s1 = np.dot((np.array(data.ix[data['class']=='class1',:-1])-
mu1).T,np.array(data.ix[data['class']=='class1',:-1])-mu1)
• mu1 = • s1#scatter matrix for class 1
np.array([np.mean(data.ix[data['class']=='class1',
0]),np.mean(data.ix[data['class']=='class1',1])]) • array([[ 4. , -2. ],

• mu2 = • [ -2. , 13.2]])


np.array([np.mean(data.ix[data['class']=='class2', • s2 = np.dot((np.array(data.ix[data['class']=='class2',:-1])-
0]),np.mean(data.ix[data['class']=='class2',1])]) mu2).T,np.array(data.ix[data['class']=='class2',:-1])-mu2)

• print(mu1,mu2) • s2#scatter matrix for class 2

• [ 3. 3.6] [ 8.4 7.6] • array([[ 9.2, -0.2],


• [ -0.2, 13.2]])
• sw = s1 + s2
• sw#within class scatter matrix
• array([[ 13.2, -2.2],
• [ -2.2, 26.4]])
Lets plot the above data and see how it looks

Using Step 3 to find the projection vector Using Step 4 to find the scalar quantities for each set
W of input features
• W = np.dot(np.linalg.inv(sw),(mu1-
mu2).reshape(2,1)) • f=[]

• W • for i in range(len(data.ix[:,:-1])):

• array([[-0.44046095], • f.append(np.dot(W,np.array(data.ix[i,:-
1],dtype='float64').reshape(2,1)))
• [-0.18822023]])
• data['projection']=np.array(f)
• W = -W.T[0] #converting it to positive as it is not
going to create much of a difference
• W
• array([ 0.44046095, 0.18822023])
The final data along with the projected scalar quantities looks
like this
Lets plot and see how well they are separated
Code-
• plt.scatter(data.ix[data['class']=='class1',3],np.a
rray([1,1,1,1,1]),color='r',marker =
'*',label='class1')
• plt.scatter(data.ix[data['class']=='class2',3],np.a
rray([1,1,1,1,1]),color='g',marker =
'o',label='class2')
• plt.legend(loc='best')
• <matplotlib.legend.Legend at 0x2090caa7588>

• Scatter Plot
Questions-

• Using the above plot of the projected scalar quantities we can easily separate the
two classes i.e any class with a scalar value less than, say ‘3’ belongs to class 1
and values greater than ‘4’ belongs to class 2.How easily LDA solved our problem
of binary classification, but I wish life was that simple.
• There might be some cases that even after applying LDA there is overlapping of
the two classes.What to do then? Our model’s accuracy will degrade.Is there a
solution?

You might also like