Over a million developers have joined DZone.

Classification From Scratch, Part 8 of 8: Linear Discrimination

DZone's Guide to

Classification From Scratch, Part 8 of 8: Linear Discrimination

We continue our article series on using R for creating regression and classification models from scratch by looking at Bayes/Naive Classifiers and Linear Regression.

· Big Data Zone ·
Free Resource

Hortonworks Sandbox for HDP and HDF is your chance to get started on learning, developing, testing and trying out new features. Each download comes preconfigured with interactive tutorials, sample data and developments from the Apache community.

This is the eight post of our series on classification from scratch. The latest one was on SVM, and today, I want to get back to some very old stuff and also a linear separation of the space, using Fisher’s linear discriminant analysis.

Bayes (Naive) Classifier

Consider the following naive classification rule:

Image title


Image title

In the case where y takes two values, that will be standard {0,1} here, one can rewrite the later as

Image title

and the set

Image title

is called the decision boundary.

Assume that




then explicit expressions can be derived

Image title

where ry2 is the Manalahobis distance

Image title

Let δbe defined as  

Image title

the decision boundary of this classifier is

{x such that δ0(x)=δ1(x)} 

which is quadratic in x. This is the quadratic discriminant analysis. This can be visualized as below.

The decision boundary is here:

But that can’t be the linear discriminant analysis, right? I mean, the frontier is not linear… Actually, in Fisher’s seminal paper, it was assumed that Σ0=Σ1.

In that case, actually,

Image title

and the decision frontier is now linear in x. This is the linear discriminant analysis. This can be visualized as below:

Here the two samples have the same variance matrix and the frontier is:

Link With the Logistic Regression

Assume as previously that 





Image title

is equal to

Image title

which is linear in 

Image title

Hence, when each group has Gaussian distributions with an identical variance matrix, then LDA and the logistic regression lead to the same classification rule.

Observe furthermore that the slope is proportional to Σ−1[μ1−μ0], as stated in Fisher’s article. But to obtain such a relationship, he observes that the ratio of between and within variances (in the two groups) was

Image title

which is maximal when ω is proportional to Σ−1>[μ1−μ0], when Σ0=Σ1.

Homebrew Linear Discriminant Analysis

To compute vector ω

m0 = apply(myocarde[myocarde$PRONO=="0",1:7],2,mean)
m1 = apply(myocarde[myocarde$PRONO=="1",1:7],2,mean)
Sigma = var(myocarde[,1:7])
omega = solve(Sigma)%*%(m1-m0)
FRCAR -0.012909708542
INCAR  1.088582058796
INSYS -0.019390084344
PRDIA -0.025817110020
PAPUL  0.020441287970
PVENT -0.038298291091
REPUL -0.001371677757

For the constant – in the equation ω^x+b=0 – if we have equiprobable probabilities, use

b = (t(m1)%*%solve(Sigma)%*%m1-t(m0)%*%solve(Sigma)%*%m0)/2

Application (on the Small Dataset)

In order to visualize what’s going on, consider the small dataset, with only two covariates:

x = c(.4,.55,.65,.9,.1,.35,.5,.15,.2,.85)
y = c(.85,.95,.8,.87,.5,.55,.5,.2,.1,.3)
z = c(1,1,1,1,1,0,0,1,0,0)
df = data.frame(x1=x,x2=y,y=as.factor(z))
m0 = apply(df[df$y=="0",1:2],2,mean)
m1 = apply(df[df$y=="1",1:2],2,mean)
Sigma = var(df[,1:2])
omega = solve(Sigma)%*%(m1-m0)
x1 -2.640613174
x2  4.858705676

Using R regular functions, we get:

fit_lda = lda(y ~x1+x2 , data=df)

Coefficients of linear discriminants:
x1 -2.588389554
x2  4.762614663

which is the same coefficient as the one we got with our own code. For the constant, use:

b = (t(m1)%*%solve(Sigma)%*%m1-t(m0)%*%solve(Sigma)%*%m0)/2

If we plot it, we get the red straight line


As we can see (with the blue points), our red line intersects the middle of the segment of the two barycenters:


Of course, we can also use an R function:

predlda = function(x,y) predict(fit_lda, data.frame(x1=x,x2=y))$class==1
contour(vu,vu,vv,add=TRUE,lwd=2,levels = .5)

One can also consider the quadratic discriminant analysis since it might be difficult to argue that Σ0=Σ1.

fit_qda = qda(y ~x1+x2 , data=df)

The separation curve is here:

predqda=function(x,y) predict(fit_qda, data.frame(x1=x,x2=y))$class==1
contour(vu,vu,vv,add=TRUE,lwd=2,levels = .5)

Hortonworks Community Connection (HCC) is an online collaboration destination for developers, DevOps, customers and partners to get answers to questions, collaborate on technical articles and share code examples from GitHub.  Join the discussion.

data classification ,big data ,r ,bayes classifier

Published at DZone with permission of

Opinions expressed by DZone contributors are their own.

{{ parent.title || parent.header.title}}

{{ parent.tldr }}

{{ parent.urlSource.name }}