**Data science** is a technique that applies all these particular parts and techniques (Machine learning and deep learning) and uses some mathematical tools like statistics, probability, linear algebra, calculus, etc. to solve a business problem or build a product.

Artificial Intelligence, Machine Learning, and Deep learning are toolkits that can be used to solve that business problem or build a product.

Artificial Intelligence is a technique that allows machines to act like humans by replicating their behavior and nature. Artificial Intelligence is the broader umbrella under which Machine Learning and Deep Learning come.

**Machine Learning** is a subset of Artificial Intelligence, It provides us with statistical tools to explore, analyze and understand the data, we have three different approaches in Machine Learning

**1. Supervised learning** –

We will have some labeled/parsed data, using this data we can do some predictions. E.g.: if we have height and weight, we need to classify if a person is obese or fit. We will create the model and train the model as we have past data and we know what will be the output of this particular data

**2. Unsupervised learning**

We don’t have labeled output, and will not know what is the output. we will solve the clustering technique, based on some similarities of the data, it will group the data using Euclidean distance. The three popular clustering algorithms- K-means, hierarchical clustering, and dbscan clustering are used for unsupervised machine learning

**3. Reinforcement learning/ Semi-supervised**

Some parts of the data will be labeled and later on some parts of the data will not be marked, so the machine model learns slowly from past data and will learn after the new data will be coming up. The algorithm learns to react to the environment.

**Deep learning**, is a subset of Machine learning.

Scientists thought can we make machines learn like how we with the help of the human brain actually try to learn things, that is the main idea behind **deep learning**, In deep learning, we create architecture called Multi neural network architecture

The different networks in deep learning

**1. Artificial Neural Networks (ANN)** -are used where in most problems statements, data are present in numbers

**2. Convolution Neural Networks (CNN)** – are used where input is in form of images.

**3. Recurrent Neural Networks (RNN)** -are used where input is in form of time series.

Using concepts of Machine learning and deep learning, we derive Artificial Intelligence applications.E.g.:- self-driving cars, recommendation engines, etc.

Data Science is the practice of using all the above tools to solve a business problem or build an AI product. You may not know all the tools in the Data Science universe, but when to apply what tools can help.

Please subscribe to the weekly newsletter for such great information every Wednesday.

### Published by

Lead Data Scientist | Analytics Leader | Retail & Customer Domain | Ex- Accenture, HP, Dell | Mentor