Weka Tutorial 01: ARFF 101 (Data Preprocessing) Video Lecture | Weka Tutorial - Data & Analytics

hello everyone this is the very first
tutorial to getting started with wake up
machine learning workbench today we will
see how you are going to prepare your
data for analyzing inside wake up so the
data file wake up work suite is called
an AR FF file or in short I tell our
file so an hour file actually has three
sections the first section is called
the relation section and every section
begins with at the rate sign and this is
the syntax of our fund this is how
wakeup deals with an hour file in order
to understand the sections so it starts
with at the rate sign and you can see at
the very first line of this weather dot
R file that at the rate relation weather
so this states that this this part is
actually the relation section of our
file and the weather this name is you
can put any other name in order to
define your relation this is just to
understand what this our file is talking
about so this our file is actually
talking about weather and that's why we
have named the relation as weather this
is not actually important for Vega for
now getting a name for your relation you
can put any other names for your
relation the second section is attribute
section this section is actually holding
the data definition of your our file of
your data and this section is called
attribute and this also starts with at
the rate sign you can see at the second
line of this weather dot our file is
that at the rate attribute and in some
names and some other things so in order
to start us an attribute section you
have to put at the rate sign and you
have to put attribute and for every
feature you are collecting from your
data you have to define them in this
attribute section so in this attribute
section you can also see outlook this is
actually the
name of the feature and there goes some
curly braces and the second line of
these are file and inside the curly
brace you can see there are three values
sunny overcast and rainy so what this
means is that our outlook feature can
have three values you're restricting the
number of values this particular feature
can have throughout this our file and
our second attribute is temperature and
temperatures data type is actually real
so it can take eighty five point five or
85 or 90 point three five or anything
that is a real value and our target to
build is also a real value which is
called humidity our fourth attribute is
called windy and the windy can take
windy add feature can take two values
either true or false and you have to
take a look at the capital is
capitalized letters of true and false
because this is important this is I'm
going to mention in some later parts of
this tutorial
and our last attribute is called clay
and this is actually a class attribute
that you are defining your class it
areas positive instance or negative
instance if this is a positive instance
if you can play then it's going to be
carrying a value yes otherwise no so in
this attribute section you're going to
put your features names and feature
values also at the very range of those
values and the datatypes generally there
are two data types for our files one is
nominal data type and two is the numeric
data type nominal data type or nominal
types of features can take values that
are not actually numeric that are
actually can be string or date sorry
data is also numeric type of data type
that's my mistake nominal data types are
actually string then
types so outlook is actually a nominal
data type here for numeric data type you
are going to carry some numeric values
for those data types those features like
temperature and humidity here because
those will contain some numeric values
not nominal values so in general in
feature section or attribute section in
our file you can have either nominal
types of data type nominal types of
features or numeric types of features so
the third section of is the data section
and it also starts with at the rate sign
at the rate data so it states waketh at
okay I'm going to start populating the
our file with my actual data and you can
see that you have to maintain the order
of the features in your data section
just like the way you have ordered your
features at attribute section and the
first one is sunny so the outlook is
sunny temperature was 85 maybe it's in
Fahrenheit humidity is 85 it's in
percentage and the value of windy is
false so what the way you read it is
that your outlook was sunny your
temperature was 85 degree Fahrenheit
your humidity was 85% and it was not
windy so what are you going to do are
you going to play or not the class value
for the first data here is no so that
means that your data says that if it is
sunny and 85 degrees centigrade five
degree fine height and 85% of humidity
then it's not windy then you're not
going to play outside
similarly the tard data in our data
section says that if it is overcast
outlook outside with edit with an 83
degree fahrenheit temperature and 86% of
humidity but without any wind if that is
the situation then you can play so in
this way you have to populate your AR
file with your data you have collected
from your training sets or
sets so this is the very first step you
have to consider in order to use wake up
making learning or data mining workbench
to analyze your data to mining data or
to learn some machine learning models
all to test your learned models on some
test sets