These will essentially be notes throughout the course of learning Machine Learning. These notes will start out with the basics, and will progress to the actual computation later on.
What is Machine Learning?
A computer program is said to learn from experience(E) with respect to some class of tasks(T) and performance measure(P), if its performance at tasks in (T), as measured by (P), improves with experience (E).

A great example would be checkers.
E
Experience of playing (X) amount of games of checkers.
T
Task of playing checkers.
P
The probability that our algorithm will win the next game.

Generally speaking Machine Learning can be assigned to one of two broad classifications:

Supervised & Unsupervised learning.



Lesson 1
--------
Supervised Learning

The basics

Our problem:
Predicting housing prices.
X = Price
Y = Size in ft ^ 2

In supervised learning we give a dataset to the algorithm. In our basic problem, we will provide the algorithm with a dataset. By providing this dataset, we are providing the “right answers”, essentially telling the program the “proper” price(X) for the size of the house(Y). This is a problem of regression, where we are attempting to predict a continuous valued output(X) based on a set of data provided (X) : (Y).

In layman’s terms, we are guiding the algorithm through the process of learning. By providing a solid set of data, we are telling the algorithm that these are the typical prices(X) based on the amount of area(Y). The algorithm will then try and predict an output(X), based on this set of data.

Terminology:
Classification problem:
A problem in which we attempt to predict a discrete valued output. In supervised learning, the algorithm will make this prediction based on the data it is given.

Closing question:
You’re running a company, and you want to develop learning algorithms to address each of two problems.

Problem 1: You have a large inventory of identical items. You want to predict how many of these items will sell over the next three months.
Problem 2: You’d like software to examine individual customer accounts, and for each account decide if it has been hacked/compromised.
How would you treat these problems?
I would treat 1.) as a regression problems, because it is a problem that is attempting to predict a valued output. Two would be a classification problem. While two is attempting to predict a discrete output based off of a variety of factors.
(Answered correctly)

What I got is that regression is attempting to predict a valued output, aka a number of some sort. While classification problems attempt to predict an overall outcome, I.E whether a tumour is malignant or benign.

Regression -
We could use facial recognition to predict someone’s age based off of variables like skin tone, eye tone, wrinkles, and so forth.
Classification -
We could use a data set of malignant / benign tumours to given an output between 1 and 0, 1 being malignant, and 0 being benign.

Lesson 2
--------
Unsupervised learning