If you are curious about Machine Learning but are tired of long and Boring articles, not to worry. In this article, you will be briefed about Machine Learning. The algorithms are also used in Data Science. If the roaming myths threaten you about the field do give it a read here.
In this article, you will go through the following Contents:
- What is Machine Learning
- Understanding PET
- Supervised Learning
- Understanding Supervised Learning
- Supervised Learning, Not Programming
- Regression
- Linear Regression
- Multiple Regression
- Polynomial Regression
- Classification
- Binary
- Multiple
- Examples of Classification and Regression
- Unsupervised Learning
What is Machine Learning?
Machine Learning is the use and development of computer systems that can learn and adapt without following explicit instructions, by using algorithms and statistical models to analyze and draw inferences from patterns in data.
In simpler terms, A computer program is said to learn from experience ‘E’ for some task ‘T’ and some performance measure ‘P’, if it’s performance on next task ‘T’, as measured by ‘P’ improves with Experience ‘E’.
Understanding P.E.T.
If our email program watches which kind of email we usually mark as spam (such as brand deals) or which type of videos we usually watch on YouTube (such as gaming videos), the algorithm will automatically mark those emails as spam and more gaming videos will appear on our YouTube feed.
Summarizing, it first observes our behavior, learns what we do, and performs the task accordingly. For further improvement, it will learn from our response too.
In this case,
â– Performance P is us spamming the respective emails, and searching gaming videos on YouTube.
â– Experience E is it learning the type of emails we mark as spam and the type of videos we watch on YouTube
â– Task T is the algorithm itself marking emails as spam and showing more gaming videos in the YouTube feed.
Supervised Learning
Supervised learning is the machine learning task of learning a function that maps an input to an output based on example input-output pairs. Each example provided consists of input and its desired output. The supervised learning algorithm will analyze the given examples and will produce an inferred function, through which the output of future inputs can be derived.
Understanding Supervised Learning
For example, we have a fruit basket. Now we have to categorize the fruits according to their categories. These specifications will be given as example data.
| Size & Color | Shape Description | Fruit Name(Response variable) |
| Small and Green | Round to oval; in a bunch | Grape |
| Big and Red | Round shape with a depression on top | Apple |
| Big and Yellow | Long Curving Shape | Banana |
The last column is the output column and the rest are description columns. After the given example the machine learning algorithm will compare the test data with each specification and will result with the closest match.
Supervised learning! Not Programming…..
The above example looks like it was programmed with if() and else() statements for the categorization of fruits, but it is different from programming.
The supervised learning algorithm will give the closest output. For example, the specification of an Apple is RED, BIG, ROUND, and a DEPRESSION ON TOP. If an apple has a yellow patch or a little stem on top, a Supervised learning algorithm will still classify it as an apple but a program will be unable to identify it unless provided with a bunch of exceptions.
This is the reason why supervised learning is used as it shortens the amount of data to be provided.
Read more on Supervised Learning
Regression
It is used to predict single value output using training data set. The output value is always called the dependent variable while the inputs are the independent variable. We have different types of regression in Supervised Learning. For example,
Linear Regression
Here we have only one input which is used for calculating the output
Y = X + 3
Multiple Regression
Here we have more than one inputs which are used for calculating the output
Y = X0 + 7X1 + 5
Polynomial Regression
Here the relation between inputs and outputs is not linear
Y = X2 – 2
Y = X31 – 2X2 + 8
Classification
It predicts results in discrete categories. There are two main types of classifications.
Binary classification
The algorithm tries to group objects into only two different groups
Multiple classifications
The algorithm tries to group objects into more than two different groups
Examples of Classification and Regression
Classification-Binary
- Guilty or Innocent
- Pass or Fail
- Spam or no-spam
- Legal or illegal
Classification-Multiple
- Grades – A* to E
- Rating 1 to 5
- Difficulty- Easy, Medium, Hard, Insane
Regression
- The trend of Crime Rate
- House Price
- Pass/fail ratio
Unsupervised Learning
It allows us to approach problems with little or no idea what our results should look like. We can derive structures from data and divide structures into different clusters
How does Clustering work?
The thing to understand is there is no example data given in Unsupervised learning. The algorithm compares the data with itself and divides it into groups having the most similarities.
Such as the food basket was given, this time it won’t be given specifications on what each thing looks like. The algorithm will compare each fruit to another and will divide each respective fruit into different clusters.
Read more on Unsupervised Learning
There are many free courses online, form where you can start your career in Machine Learning from the beginning. A few of them are as follows.
- Kaggle
- Introduction to Machine Learning
- Intermidiate machine Learning
- Feature Engineering
- Machine Learning Explanability
- Coursera
Apart from these, there are a lot of courses on these websites and others too.
