Machine learning is the part of artificial intelligence (AI) that provides the system able to automatically learn and improve from experience without being explicitly programmed.
Machine learning focuses on the development of computer programs that can access data and use it to learn for themselves.
Difference between Traditional Programming vs Machine Learning Programming:
Programming is telling a computer what to do with the given set of instructions that we call as input. It basically makes the dumbest computers which make the same mistake again and again with the same set of instructions.
On the other hand, Machine Learning is making a machine to learn, it involves making a computer to learn from given set of instructions, in this case, the computer won’t repeat its mistakes but learn from them as humans.
Both have different applications like we use basic programming to make simple applications, but when it comes to making such programs where user interest needs to be taken care of, or study of a varying parameter is necessary like in a recommend-er system or a stock predictor, we use machine learning.
The process of learning begins with observations or data, such as examples, direct experience, or instruction, in order to look for patterns in data and make better decisions in the future based on the examples that we provide. The basic premise of machine learning is to build algorithms that can receive input data and use statistical analysis to predict an output while updating outputs as new data becomes available.
Machine Learning Types :
Machine learning is divided into further three parts:
1.) Supervised Machine learning
2.) Unsupervised
3.) Reinforcement Learning
Supervised Learning:
Supervised Learning is very important in machine learning because it contains a target attribute. Now you can see in the table, two features and one target/label attribute is given.
| Weight | Height | Gender |
|---|---|---|
| 75 | 175 | Male |
| 58 | 140 | Female |
| 86 | 185 | Male |
| 63 | 150 | Female |
| 89 | 180 | ? Is it belong to male class or female |
According to the table, If any person is 75 kg weights and 175 cm that means it belongs to the male class. In this similar manner, the weight is 58 kg and the height is 140 cm that means it belongs to the female class.
So this type we have given that dataset to the supervised machine learning model, so the model can learn the dataset and give the results according to it.
Now for testing purposes, we will give the other data (for example 89 kg and 185 cm ) which is not present in the training data to check the model is able to predict the gender or not.
So we have already provided the Target/Label attribute in the dataset and teach the model about like a supervisor. So that’s why it is known as the supervised machine learning algorithm.
The Label/ target attribute is that you are trying to predict, or you can also be called as an Outcome Measurement in Supervised Learning.
So, In supervised learning, a dataset with labels is provided. Here, We will train the machine with the data, the machine learns to make predictions. In this process, we train the machine with the data and the machine learns to make good predictions accordingly.
2.) Unsupervised Learning
Unsupervised Learning means there is no any type of supervisor, There is no any type of target attribute is given as you can see on the given table. Here only features are present and there is no any type of target attribute.
In unsupervised learning, a dataset without labels is provided that means there is no gender attribute is available as you can see on the table below:
| Weight | Height | |
|---|---|---|
| 75 | 175 | |
| 58 | 140 | |
| 86 | 185 | |
| 63 | 150 | |
| 89 | 180 |
There is no supervisor, Here model learning by itself and improved from the previous output and the Target values are missing, So patterns are used for predictions.
As you can see here, we have applied a pure orange and red apple to the unsupervised machine learning model.
Then we will apply the test yellow-orange to the model and we want the model to predict for us.
Is this test data is orange or apple? So the unsupervised learning model creates a two cluster of given input and matches the test data with them.
If the test data matches with the orange then the result belongs to the Orange class otherwise it belongs to the apple class. So this is the classification problem.
In unsupervised learning, a pattern matching has been done of each created cluster. If the test data matches any pattern then it will belong to the particular class.
3.) Reinforcement Learning:
Reinforcement Learning is the third type and advanced machine learning methodology depends on “Reward” that’s why it is called Reward-based learning for every Positive action and “Penalty” for every negative action.
The machine learns from its previous mistakes and tries to make the model more perfect.
The practical approaches of Reinforcement Learning in the following fields:
- Manufacturing
- Finance Sector
- Stock Market
- Inventory Management
I tried to cover the basics of machine learning and different types of machine learning like Supervised, Unsupervised and Reinforcement Learning and their applications.