Simply put, Machine Learning is about training a machine (model) to predict output for a given input.
Let us try to understand the math behind the prediction part.
Can you predict a missing number in the pattern below?
(1, 10) (2, 20) (3, 30) (4, 40) (5, ?)
That was quite simple. Try another one...
(0, 5) (2, 3) (6, -1) (4, ?)
This one isn't that obvious. But there's a simple way. Let us treat these as
(x,y) co-ordinates and try to plot a graph.
Now, were you able to find the missing number? Yes, it is 1 (the red dot).
Can we do it without a graph? The answer is a linear equation.
A linear equation is an algebraic equation where each term has an exponent of 1 and when this equation is graphed, it always results in a straight line. This is the reason why it is named a 'linear' equation.
The linear equation is given as -
y = mx + b
m = slope (know more)
b = y-intercept (the point at which the line intersects the Y-axis) (know more)
Now, to find the value of
y using the linear equation, for a given
x we will need values of the slope
m and the y-intercept
Consider the given points.
(x, y) (0, 5) (2, 3) (6, -1)
Fortunately, we have a point with
x = 0 so finding
b, the y-intercept is easy.
Using the first point
(x, y) = (0, 5) :
y = m * x + b 5 = m * 0 + b 5 = 0 + b 5 = b
b = 5.
b as 5 and the second point in the table
(x, y) = (2, 3), let us find
m, the slope.
y = m * x + b 3 = m * 2 + 5 3 - 5 = m * 2 -2 = m * 2 -2 / 2 = m -1 = m
m = -1.
Let us use
m = -1 and
b = 5 to find the missing number.
(x, y) (4, ?) y = m * x + b y = -1 * 4 + 5 y = -4 + 5 y = 1
We were finally able to find the missing number, 1.
In the next one, ML: Simple Model, we'll try to train a model to predict the value for the same example above.