Some machine learning Algorithms are
- Supervised ML Algorithm
- Unsupervised ML Algorithm
- Recommender systems
- Reinforcement Learning
Supervised ML Algorithm
This algo basically learns from input to output labels
Just see how these different Algorithms take inputs and gives which outputs
All of these models have training models and you train them with data.
Lets learn from an example
But you can do some complex works too , to get most correct data.
If you create a curve, you can see the price changes and this time it shows even more from the past time
Ultimately, you will get a value but mainly this is how in supervise model, machine learns what should be the data. We will know more correctly later, which process you should pick . But for this type of problem, we will use Regression .
We will know about this terms in the later part of the blog
Now, lets learn another way called Classification
Mainly we will try to predict breast tumors based on sizes.
from a tumor, we can detect if it is malignant or not
Now depending on the size of the tumor, it can be cancer (Represented by 1) or not (0)
Depending on the output, we can draw points
Now, we later found that there are 2 types of malignant . So, we pointed them out and classified them in this manner
So, classification algorithms predict categories and they not only predict numbers. It can be non numeric like malignant or benign for this Breast cancer prediction.
Now if we also provide age with tumor size, the graph can be like this
The doctor can easily predict the age & tumor size from the graph
Also, from this graph, we can make a curved line to differ malignant patients & not benign patients
In summary: Analyzes data and creates a label. For example, finds if a patient is diabetes patient or not.
Unsupervised ML Algorithm
To find the patterns in data, we use unsupervised or Algorithm. For example, in supervised learning we tried to find if a person has cancer or not using tumor size etc . But in unsupervised learning, we will just find a pattern between them. We don't want to hear if a patient has cancer or not by providing the tumor size.
Basically it divides the data into clusters
Each row stores gene information of each person and each column has genes which control certain behaviour
So, this is how we can cluster.
Moreover, we can have another example. We can have a clusters of customers.
Therefore , we can say
In summary: creates different clusters by finding patterns