In this section we will discuss three key concepts in machine learning – data model and learning. How they are related and their functions.
In previous sections we have discussed a few other key concepts, such as Python and scikit-learn, the role of python libraries in machine learning and many more.
However, we have not discussed the key concept on which machine learning is based on.
Firstly, let us define machine learning. It is nothing but designing algorithms.
Take a close look at the concept algorithms. We have learned that algorithms represent a set of rules or instructions that a computer translates to machine code and according to those rules, it performs.
However, machine learning algorithms are of different kinds. They extract valuable information from data in an automated way.
When we say automated way, it means there is a kind of self-regulation. Now the speciality of these algorithms lies in the concept of this self-regulation.
Why?
Because machine learning algorithms don’t act on a single datasets but it applies its general purpose methodologies to many datasets.
What do these algorithms extract?
Some meaningful predictions, visualizations, and many more that we will see in the future.
Now how these three core concepts data model and learning is related to machine learning?
Let’s see to it.
The role of data in machine learning
Firstly, data is at the heart of machine learning.
Right? It’s because machine learning is by default data driven.
A machine learns from datasets and trains itself to a more accurate prediction.
For example we can take a look at the data table. A two dimensional grid of data which we lovingly call a table.
To generate such a generous data table we can import a library seaborn which we will use more and more in future.
import seaborn as sns
# this will pick up the head part
iris_head = sns.load_dataset('iris')
iris_head.head()
When we run the code it displays the following data table.
The data table comes from famous Iris datasets.
We have used the seaborn library.
Since we have used the head() function, we can also use the tail() function to see what kind of flower varieties it can produce.
# this will pick up the tail part
iris_tail = sns.load_dataset('iris')
iris_tail.tail()It produces the following data table which is distinguishable in its nature.
We can see the sepal and petal width, length are different.
Certainly Iris Setosa and Iris Virginica are two different kinds of flowers although they almost look alike.
We get the point.
If you are a complete beginner your journey to learn TensorFlow might start from here.
For the TensorFlow beginners we have a dedicated category – TensorFlow for Beginners.
But besides that, you may need to learn several other machine learning and data science libraries.
As a result, you may check these categories as well – NumPy, Pandas, Matplotlib.
However, without learning Python, you cannot learn the usages of these libraries. Why? Because they all use Python as the Programming language.
Therefore please learn Python at the very beginning and start learning TensorFlow.
And, finally please check our Mathematics, Discrete Mathematics and Data Structures categories specially. We have tried to discuss from basic to intermediate level so that you can pick up the core ideas of TensorFlow.
Without data we cannot think of a machine learning model which is another key concept.
Our machine learning algorithms have successfully extracted meaningful patterns.
The role of model in machine learning
Based on the datasets we have seen above, we can build a generalized model that can learn from data.
If the performance of the model improved we will conclude that this model will be learning better in future.
Why? Because it performs better and better when more difficult tasks are given to it.
There are a lot of models available. We need not worry about it at present.
All we need to do is understand the core concepts of machine learning.
The role of learning in machine learning
When algorithms perform good and a machine learns to extract meaningful patterns from datasets, we say that there is a ‘learning’ involved.
Basically the learning acts upon the models.
How it occurs we will see in future.
So stay tuned and cheer up.
Leave a Reply