Skip to main content

Types of Machine Learning

Using computers to analyze data and find patterns in large datasets is achieved with Machine Learning. However, there are different ways that data can be observed.

     Classification - Looking at elements in a dataset and determining what category the element belongs to, based on preset parameters.
Visual representation of elements divided by line which was created by a function, created by Classification
     Imagine that I worked in a post office and I had to sort mail into different categories, Packages and letters. The only information that I am given is the weight of the mail. Based on the weight of the incoming mail, a model can be built so that mail can be classified as a package or a letter.

     Clustering - Grouping elements in a dataset into categories, based on attributes of the elements.
Visual representation of elements that have been classified by Clustering
An example of this would be if I had a data set of what students used to read textbooks. The program could first look at the given data set and separate the electronic methods from the non-electronic. Then we would have 2 groups which we could label as: Physical Books and Electronic devices. We would have the program continue and it could separate the Electronic devices by screen size and we would now have 3 groups: Physical books, Phones, and Laptops.
This differs from Classification in that the data is not trained since this is an unsupervised Machine Learning method.

     Regression - Creating a function to predict what a value will be, when given certain parameters.
Visual representation of line created by a function, created by Regression
Let’s say I wanted to buy a car. I could create a model which looks at the year of the car, gas mileage and other various information and produce an estimated price of the car.

These are the most common types of machine learning algorithms. I'll be delving into them in more detail in my later posts

Popular posts from this blog

Compiled vs Interpreted

Do you program? Do you write code? Depending on what you're programming or trying to create, you'll need a programming language and depending on the task at hand the decision will need to be made of using an interpreted or compiled language.

What's the difference?

Think about it like this, when you read a book, you read it line by line. An interpreted language does the same kind of thing. It reads the code in a file line by line. Immediately after that line is read, the program reading the code (the interpreter) runs that line of code then moves on to the next one. This is typically done for platforms such as the web.

A compiled language is slightly different. A compiled language would read the entire file of code, and translate it to another format (ex. machine code or another language). This is done by a compiler. Once the code is in another format it can be run. These types of languages are usually used for native desktop or mobile applications.

There is a mix of interprete…

Comment your code (well)

From basic beginners to experienced experts, everyone is either a poor commenter or knows someone who is a poor commenter. If you need a hair cut, try understanding an open source uncommented project. You will tear out all of your hair from frustration. 
Not writing comments is bad, but so is writing bad comments. If your comment is so vague that if I'm reading through your code and it seems completely random, or makes me have to remind myself what kind of project that I'm looking at, maybe it shouldn't be there.
There are 5 types of comments that I've seen: No comment - There's nothing there (Bad)Vague comment - It just doesn't make sense in the context (Bad)//This counts Unnecessary comment - It just doesn't need to be there (Bad)//this line adds 1 and 2 together Funny commentHelpful/Descriptive comment - Brief yet descriptive about the method/function. Placed at a point in the program that could be potentially confusing. A helpful comment might also describe …

So what is Machine Learning?

It’s mentioned all the time in the news and companies are constantly touting how enhanced their applications are with learning machines. But what does that really mean?
To put it simply, its teaching a machine how to do something. There’s more than one way to do this but most models follow the same method of “Practice makes perfect”.
For example, if I wanted to learn how to draw, I would start with the basics and do simple things. I would practice the motions and gestures that I wanted to get the desired effect. The more I practice, the better (in most cases) I would get. My skill would improve over time. However I could be learning the wrong way and actually be making my drawing ability worse.
Teaching machines is the same way. A machine is given a data set (ex. a series of images it wants to learn how to draw) and it runs a loop that practices its skill at finding patterns amongst the data set (ex. a flower petal is more curves than straight lines, while a sidewalk is mostly straig…