Setiawan Hadi

Sumber : https://hackr.io

The words data science and machine learning are often used in conjunction, however, if you are planning to build a career in one of these, it is important to know the differences between machine learning and data science.

Before doing so, we need to understand a few important terms that are related but different.

AI (Artificial intelligence) – AI or machine intelligence refers to the intelligent decisions made by machines at par with their human counterparts. It is a study where we enable machines to learn through experience and make it intelligent enough to perform human-like tasks. In my article about AI vs ML, I have listed the differences between AI and Machine learning. For this article, let me give you a simple definition of machine learning.

Machine learning – Think of ML as a subset of AI. Same way as humans learn with experience, machines can learn with data (experience) rather than just following simple instructions. This is called as machine learning. Machine learning uses 3 types of algorithms – supervised, unsupervised and reinforced.

Deep learning – Deep learning is a part of Machine learning, which is based on artificial neural networks (think of neural networks similar to our own human brain). Unlike machine learning, deep learning uses multiple layers and structures algorithms such that an artificial neural network is created that learns and makes decisions on its own!

Big Data – Humongous sets of data that can be computationally analyzed to understand and process trends, patterns and human behavior.

Data Science – How is all the big data analyzed? Fine, the machine learns on its own through machine learning algorithms – but how? Who gives the necessary inputs to a machine for creating algorithms and models? No points for guessing that it is data science. Data Science is a uses different methods, algorithms, processes, and systems to extract, analyze and get insights from data.

Check out our exciting data science tutorials here.

If we were to see the relationship between all the above in a simple diagram, this is how it would look like this picture.

Artificial Intelligence (AI)

Artificial Intelligence includes both Machine learning and Data science which are correlated. Thus, data science is also a part (the most popular and most important one) of AI.

As we see above, Data science and machine learning are closely related and provide useful insights and generate the necessary trends or ‘experience’. In both, we use supervised methods of learning i.e. learning from huge data sets.

How are both correlated?

Data Science is a broader field of study that uses algorithms and models of machine learning to analyse and process data. Apart from learning, data science also involves data integration, visualization, data engineering, deployment and business decisions.

Data Science vs Machine learning

So, what’s the difference?
On one hand, data science focuses on data visualization and a better presentation, whereas machine learning focuses more on the learning algorithms and learning from real-time data and experience.

Always remember – data is the main focus for data science and learning is the main focus for machine learning and that is where the difference lies.

To appreciate this difference more, let us take a use case and see how both data science and machine learning can be used to achieve the results we want –

Let us say you want to purchase a phone on xyz.com. This is the first time you are visiting xyz.com and you are browsing through phones of all ranges. You use various filters to narrow down your preferences and out of the results you get, you choose 4-5 of the phones and compare those. Once you select a phone model, you will see a recommendation below the product – for a similar product in a lesser price or with more features, or related accessories for the phone you have chosen and so on. How does the website recommend you these things? It has no history about you!

That’s through the data from millions of other people who may have tried to purchase the same phone, and searched/bought other accessories along. This makes the system automatically recommend the same to you.

The entire process of collection of data from the users, cleaning and filtering out the required data for evaluation, evaluation of the filtered data for building patterns, finding similar trends and building a model for a recommendation of the same thing to other users and finally the optimization – is data science.

Where is machine learning in all this? Well, how do you build a model? Through machine learning algorithms. Based on the data collected and trends generated, the machine understands that these are the accessories that are usually bought by other users with a particular phone. Hence, it suggests you the same thing based on what it has ‘experienced’ before.

Data Science (DS)

The modeling (second last) step is the most critical step because that is what improves the overall business and makes the machine understand human behavior. If the right machine learning model is applied, it could mean more progressive learning for the machine as well as success for the business model.

This step is called as the data modeling step – which is essentially the machine learning phase of the data science lifecycle.

Data modeling – how does machine learning work?
There are different types of machine learning algorithms, the most common being clustering, matrix factorization, content-based, recommendations, collaborative filtering and so on. Machine learning involves the 5 basic steps –

The huge set of data that we receive in the first step is split into the training set and testing set and the model is built and test using the training set. A significant portion of data is used for training purposes so that different conditions of input and output can be achieved and the model built is closest to the required result (recommendation, human behavior, trends, etc…).

Once built, the model is tested for efficiency and accuracy using the test data so that it can be cross-validated.

As we can see, Machine Learning comes into picture only during the data modeling phase of the Data Science lifecycle. Data Science thus contains machine learning.

With machine learning, the machine can generate complex mathematical algorithms that need not be programmed by a human, and further can improvise and improve the programs all by itself. When compared to the traditional statistical analysis techniques, machine learning evolves as a better way of extraction and processing the most complex sets of big data, thereby making data science easier and less chaotic.

Furthermore, machines tend to be more accurate and have a better memory than humans, they can learn and produce accurate results based on experiences. We get fast algorithms and data-driven models without the errors that are possible by humans.