# 2.1 Machine learning lecture 2 course notes  (Page 6/6)

 Page 6 / 6
$\begin{array}{ccc}\hfill {\Phi }_{35000|y=1}& =& \frac{{\sum }_{i=1}^{m}1\left\{{x}_{35000}^{\left(i\right)}=1\wedge {y}^{\left(i\right)}=1\right\}}{{\sum }_{i=1}^{m}1\left\{{y}^{\left(i\right)}=1\right\}}=0\hfill \\ \hfill {\Phi }_{35000|y=0}& =& \frac{{\sum }_{i=1}^{m}1\left\{{x}_{35000}^{\left(i\right)}=1\wedge {y}^{\left(i\right)}=0\right\}}{{\sum }_{i=1}^{m}1\left\{{y}^{\left(i\right)}=0\right\}}=0\hfill \end{array}$

I.e., because it has never seen “nips” before in either spam or non-spam training examples, it thinks the probability of seeing it ineither type of email is zero. Hence, when trying to decide if one of these messages containing “nips” is spam, it calculates theclass posterior probabilities, and obtains

$\begin{array}{ccc}\hfill p\left(y=1|x\right)& =& \frac{{\prod }_{i=1}^{n}p\left({x}_{i}|y=1\right)p\left(y=1\right)}{{\prod }_{i=1}^{n}p\left({x}_{i}|y=1\right)p\left(y=1\right)+{\prod }_{i=1}^{n}p\left({x}_{i}|y=0\right)p\left(y=0\right)}\hfill \\ & =& \frac{0}{0}.\hfill \end{array}$

This is because each of the terms “ ${\prod }_{i=1}^{n}p\left({x}_{i}|y\right)$ ” includes a term $p\left({x}_{35000}|y\right)=0$ that is multiplied into it. Hence, our algorithm obtains $0/0$ , and doesn't know how to make a prediction.

Stating the problem more broadly, it is statistically a bad idea to estimate the probability of some event to be zero just because you haven't seen it before in your finite training set. Take the problem of estimatingthe mean of a multinomial random variable $z$ taking values in $\left\{1,...,k\right\}$ . We can parameterize our multinomial with ${\Phi }_{i}=p\left(z=i\right)$ . Given a set of $m$ independent observations $\left\{{z}^{\left(1\right)},...,{z}^{\left(m\right)}\right\}$ , the maximum likelihood estimates are given by

${\Phi }_{j}=\frac{{\sum }_{i=1}^{m}1\left\{{z}^{\left(i\right)}=j\right\}}{m}.$

As we saw previously, if we were to use these maximum likelihood estimates, then some of the ${\Phi }_{j}$ 's might end up as zero, which was a problem.To avoid this, we can use Laplace smoothing , which replaces the above estimate with

${\Phi }_{j}=\frac{{\sum }_{i=1}^{m}1\left\{{z}^{\left(i\right)}=j\right\}+1}{m+k}.$

Here, we've added 1 to the numerator, and $k$ to the denominator. Note that ${\sum }_{j=1}^{k}{\Phi }_{j}=1$ still holds (check this yourself!), which is a desirable property since the ${\Phi }_{j}$ 's are estimates for probabilities that we know must sum to 1. Also, ${\Phi }_{j}\ne 0$ for all values of $j$ , solving our problem of probabilities being estimated as zero. Under certain (arguably quite strong) conditions, it can be shown that the Laplace smoothing actually gives the optimal estimator of the ${\Phi }_{j}$ 's.

Returning to our Naive Bayes classifier, with Laplace smoothing, we therefore obtain the following estimates of the parameters:

$\begin{array}{ccc}\hfill {\Phi }_{j|y=1}& =& \frac{{\sum }_{i=1}^{m}1\left\{{x}_{j}^{\left(i\right)}=1\wedge {y}^{\left(i\right)}=1\right\}+1}{{\sum }_{i=1}^{m}1\left\{{y}^{\left(i\right)}=1\right\}+2}\hfill \\ \hfill {\Phi }_{j|y=0}& =& \frac{{\sum }_{i=1}^{m}1\left\{{x}_{j}^{\left(i\right)}=1\wedge {y}^{\left(i\right)}=0\right\}+1}{{\sum }_{i=1}^{m}1\left\{{y}^{\left(i\right)}=0\right\}+2}\hfill \end{array}$

(In practice, it usually doesn't matter much whether we apply Laplace smoothing to ${\Phi }_{y}$ or not, since we will typically have a fair fraction each of spam and non-spam messages, so ${\Phi }_{y}$ will be a reasonable estimate of $p\left(y=1\right)$ and will be quite far from 0 anyway.)

## Event models for text classification

To close off our discussion of generative learning algorithms, let's talk about one more model that is specifically for text classification. While Naive Bayes as we've presentedit will work well for many classification problems, for text classification, there is a related model that does even better.

In the specific context of text classification, Naive Bayes as presented uses the what's called the multi-variate Bernoulli event model . In this model, we assumed that the way an email is generated is that first it is randomly determined (according to the classpriors $p\left(y\right)$ ) whether a spammer or non-spammer will send you your next message. Then, the person sending the email runs through the dictionary, deciding whether to include each word $i$ in that email independently and according to the probabilities $p\left({x}_{i}=1|y\right)={\Phi }_{i|y}$ . Thus, the probability of a message was given by $p\left(y\right){\prod }_{i=1}^{n}p\left({x}_{i}|y\right)$ .

Here's a different model, called the multinomial event model . To describe this model, we will use a different notation and set of features for representing emails. We let ${x}_{i}$ denote the identity of the $i$ -th word in the email. Thus, ${x}_{i}$ is now an integer taking values in $\left\{1,...,|V|\right\}$ , where $|V|$ is the size of our vocabulary (dictionary). An email of $n$ words is now represented by a vector $\left({x}_{1},{x}_{2},...,{x}_{n}\right)$ of length $n$ ; note that $n$ can vary for different documents. For instance, if an email starts with “A NIPS ...,” then ${x}_{1}=1$ (“a” is the first word in the dictionary), and ${x}_{2}=35000$ (if “nips” is the 35000th word in the dictionary).

In the multinomial event model, we assume that the way an email is generated is via a random process in which spam/non-spam is first determined (according to $p\left(y\right)$ ) as before. Then, the sender of the email writes the email by first generating ${x}_{1}$ from some multinomial distribution over words ( $p\left({x}_{1}|y\right)$ ). Next, the second word ${x}_{2}$ is chosen independently of ${x}_{1}$ but from the same multinomial distribution, and similarly for ${x}_{3}$ , ${x}_{4}$ , and so on, until all $n$ words of the email have been generated. Thus, the overall probability of a message is given by $p\left(y\right){\prod }_{i=1}^{n}p\left({x}_{i}|y\right)$ . Note that this formula looks like the one we had earlier for the probability of a message under themulti-variate Bernoulli event model, but that the terms in the formula now mean very different things. In particular ${x}_{i}|y$ is now a multinomial, rather than a Bernoulli distribution.

The parameters for our new model are ${\Phi }_{y}=p\left(y\right)$ as before, ${\Phi }_{k|y=1}=p\left({x}_{j}=k|y=1\right)$ (for any $j$ ) and ${\Phi }_{i|y=0}=p\left({x}_{j}=k|y=0\right)$ . Note that we have assumed that $p\left({x}_{j}|y\right)$ is the same for all values of $j$ (i.e., that the distribution according to which a word is generated does not depend on its position $j$ within the email).

If we are given a training set $\left\{\left({x}^{\left(i\right)},{y}^{\left(i\right)}\right);i=1,...,m\right\}$ where ${x}^{\left(i\right)}=\left({x}_{1}^{\left(i\right)},{x}_{2}^{\left(i\right)},...,{x}_{{n}_{i}}^{\left(i\right)}\right)$ (here, ${n}_{i}$ is the number of words in the $i$ -training example), the likelihood of the data is given by

$\begin{array}{ccc}\hfill \mathcal{L}\left(\Phi ,{\Phi }_{k|y=0},{\Phi }_{k|y=1}\right)& =& \prod _{i=1}^{m}p\left({x}^{\left(i\right)},{y}^{\left(i\right)}\right)\hfill \\ & =& \prod _{i=1}^{m}\left(\prod _{j=1}^{{n}_{i}},p,\left({x}_{j}^{\left(i\right)}|y;{\Phi }_{k|y=0},{\Phi }_{k|y=1}\right)\right)p\left({y}^{\left(i\right)};{\Phi }_{y}\right).\hfill \end{array}$

Maximizing this yields the maximum likelihood estimates of the parameters:

$\begin{array}{ccc}\hfill {\Phi }_{k|y=1}& =& \frac{{\sum }_{i=1}^{m}{\sum }_{j=1}^{{n}_{i}}1\left\{{x}_{j}^{\left(i\right)}=k\wedge {y}^{\left(i\right)}=1\right\}}{{\sum }_{i=1}^{m}1\left\{{y}^{\left(i\right)}=1\right\}{n}_{i}}\hfill \\ \hfill {\Phi }_{k|y=0}& =& \frac{{\sum }_{i=1}^{m}{\sum }_{j=1}^{{n}_{i}}1\left\{{x}_{j}^{\left(i\right)}=k\wedge {y}^{\left(i\right)}=0\right\}}{{\sum }_{i=1}^{m}1\left\{{y}^{\left(i\right)}=0\right\}{n}_{i}}\hfill \\ \hfill {\Phi }_{y}& =& \frac{{\sum }_{i=1}^{m}1\left\{{y}^{\left(i\right)}=1\right\}}{m}.\hfill \end{array}$

If we were to apply Laplace smoothing (which needed in practice for good performance) when estimating ${\Phi }_{k|y=0}$ and ${\Phi }_{k|y=1}$ , we add 1 to the numerators and $|V|$ to the denominators, and obtain:

$\begin{array}{ccc}\hfill {\Phi }_{k|y=1}& =& \frac{{\sum }_{i=1}^{m}{\sum }_{j=1}^{{n}_{i}}1\left\{{x}_{j}^{\left(i\right)}=k\wedge {y}^{\left(i\right)}=1\right\}+1}{{\sum }_{i=1}^{m}1\left\{{y}^{\left(i\right)}=1\right\}{n}_{i}+|V|}\hfill \\ \hfill {\Phi }_{k|y=0}& =& \frac{{\sum }_{i=1}^{m}{\sum }_{j=1}^{{n}_{i}}1\left\{{x}_{j}^{\left(i\right)}=k\wedge {y}^{\left(i\right)}=0\right\}+1}{{\sum }_{i=1}^{m}1\left\{{y}^{\left(i\right)}=0\right\}{n}_{i}+|V|}.\hfill \end{array}$

While not necessarily the very best classification algorithm, the Naive Bayes classifier often works surprisingly well. It is often also a very good “first thing to try,”given its simplicity and ease of implementation.

what does nano mean?
nano basically means 10^(-9). nanometer is a unit to measure length.
Bharti
do you think it's worthwhile in the long term to study the effects and possibilities of nanotechnology on viral treatment?
absolutely yes
Daniel
how to know photocatalytic properties of tio2 nanoparticles...what to do now
it is a goid question and i want to know the answer as well
Maciej
Abigail
for teaching engĺish at school how nano technology help us
Anassong
Do somebody tell me a best nano engineering book for beginners?
what is fullerene does it is used to make bukky balls
are you nano engineer ?
s.
fullerene is a bucky ball aka Carbon 60 molecule. It was name by the architect Fuller. He design the geodesic dome. it resembles a soccer ball.
Tarell
what is the actual application of fullerenes nowadays?
Damian
That is a great question Damian. best way to answer that question is to Google it. there are hundreds of applications for buck minister fullerenes, from medical to aerospace. you can also find plenty of research papers that will give you great detail on the potential applications of fullerenes.
Tarell
what is the Synthesis, properties,and applications of carbon nano chemistry
Mostly, they use nano carbon for electronics and for materials to be strengthened.
Virgil
is Bucky paper clear?
CYNTHIA
so some one know about replacing silicon atom with phosphorous in semiconductors device?
Yeah, it is a pain to say the least. You basically have to heat the substarte up to around 1000 degrees celcius then pass phosphene gas over top of it, which is explosive and toxic by the way, under very low pressure.
Harper
Do you know which machine is used to that process?
s.
how to fabricate graphene ink ?
for screen printed electrodes ?
SUYASH
What is lattice structure?
of graphene you mean?
Ebrahim
or in general
Ebrahim
in general
s.
Graphene has a hexagonal structure
tahir
On having this app for quite a bit time, Haven't realised there's a chat room in it.
Cied
what is biological synthesis of nanoparticles
what's the easiest and fastest way to the synthesize AgNP?
China
Cied
types of nano material
I start with an easy one. carbon nanotubes woven into a long filament like a string
Porter
many many of nanotubes
Porter
what is the k.e before it land
Yasmin
what is the function of carbon nanotubes?
Cesar
I'm interested in nanotube
Uday
what is nanomaterials​ and their applications of sensors.
what is nano technology
what is system testing?
preparation of nanomaterial
how did you get the value of 2000N.What calculations are needed to arrive at it
Privacy Information Security Software Version 1.1a
Good
Got questions? Join the online conversation and get instant answers!