what is machine learning
what is machine learning

What is Machine Learning:

Machine Learning is a sub-type of Artificial Intelligence. Machine Learning often refers to as predictive modeling. In simple words, machine learning is a programmed model/algorithm that receives input data and predicts some output based on the data. When new data is fed to the algorithm, they learn and improve their performance over time. Furthermore, Recommender systems are common examples of machine learning in today’s world. Other popular machine learning uses are fraud detection, spam filtering, chatbots, business automation, etc.

Why is machine learning important?

Machine Learning is playing a vital role in this era. ML enterprise companies a view of trends in the market based on the behavior of customers. Machine learning gives an idea of future products based on the data and behaviors of current market trends. Furthermore, many leading companies like Google, Amazon, Facebook, Uber make machine learning a central and integral part of their business.

Types of Machine Learning:

machine learning types
machine-learning types

Machine Learning has three types given below.

  1. Supervised Learning
  2. Unsupervised Learning
  3. Reinforcement Learning

Machine Learning types are defined because of their different ways of training machine learning algorithms. However, It is necessary to understand the types of machine learning because when given a task we may know which type/environment we have to use for this particular problem.

1)  Supervised Learning:

In simple words, we train a model based on some inputs that predict some output best based on labeled data. It is the most popular type of machine learning. It is easiest to understand and implement. The supervised machine learning model is shown in a flow chart below.

Supervised Learning
Supervised Learning

In supervised learning, an operator gives a known dataset to a machine learning model/algorithm which includes the desired inputs and the outputs. And the machine learning algorithm has to find the relationship between inputs and outputs. However, On the other hand, the operator knows the correct label of the input, the machine learning algorithm finds relationships in data, learns observation, and makes predictions on data. While the operator corrects the algorithm’s prediction and the process goes on and on until the machine learning algorithm achieves a maximum or high level of accuracy.

There are two types of Supervised Learning.

  • Classification
  • Regression


In classification tasks, the machine learning algorithm has to classify the new input based on the training model.

For Example:

  • When classifying the emails as ‘spam’ and ‘not spam’ the ML model looks at the existing observation data and classifies the new email
  • If you have a group of vegetables and fruits. We can train our model based on that dataset. After training our model we will try that model to a new instance (fruit or vegetable) and our model predicts the class based on their


Regression allows you to estimate how a dependent variable changes when an independent variable changes. Linear Regression, is used to find/estimate the relationship between two quantitative variables. For Example, Regression is used to predict the price of plots in a specific city or a value of a stock.

2)  Unsupervised Learning:

Unsupervised learning is almost the opposite of supervised learning. In this machine learning type, we don’t have any output variable. However, In this method, we fed a lot of data to our machine learning algorithm to understand the properties and relationships among data. From the data, the ML algorithm tries to structure that data. The algorithm learns to group or cluster the data in a way that humans or other algorithms make some sense of the organized data.

There are two types of Unsupervised Learning.

  • Clustering
  • Dimension Reduction


Clustering is used to find similar groups in a dataset. In data, we don’t have any labels. The ML algorithm has to understand the data and make a cluster of the data based on their nature. Unlike classification, we have an output class. This technique is very useful for segmenting the data into several clusters. The clustering chart flow is shown below.


For Example, Clustering maybe use to find the same tweets based on the nature of the content. A clustering algorithm maybe use to find a group of photos with similar cars.

Dimension Reduction:

Unsupervised learning is also use for the dimension reduction of the data. In simple words, it is use to reduce the complexity of the data. This ML algorithm is use to reduce the number of variables considered to find exact information. Dimensionality reduction is needed because it reduces the time and the storage required. It also becomes easy to visualize the data when reduced to a very low dimension like 2D or 3D.

3) Reinforcement learning:

Reinforcement learning is a completely different method as compared to supervised learning and unsupervised learning. Furthermore, Reinforcement learning is behavior-driven. This type of ML is inspired by neuroscience and psychology. It directly takes inspiration from humans how humans learn from data in their lives.

Sometimes the output is not the same as the required output so a system gives feedback on that output. Learning based on that feedback is called Reinforcement Learning. The reinforcement learning diagram is showing below.

Furthermore, It is often used for gaming, robotics, and navigation. This type of learning has only three main parts: The decision-maker (Agent), everything that interacts with the agent (Environment), and what agents are doing (Action). The only goal of reinforcement learning is to learn the best policy.

For Example:

  • Applications in self-driving
  • Industry automation with Reinforcement
Reinforcement learning
Reinforcement learning

Applications of Machine Learning:

Machine learning is use in situations where we need continuous improvement. This nature of ML solutions is the most selling point. ML can be use as medium-skilled labor as a substitute.

For Example, In large B2C (Business to Consumer) companies customer service representatives are replace by the natural language processing ML chatbots. These chatbots read customer queries and give them solutions accordingly. The same is the case with Social Media companies and large e-commerce companies like Amazon, AliExpress, eBay, Facebook, Netflix. They used human recommender systems that are based on the human’s likes and dislikes and show them content accordingly.

Amazon uses ML

Amazon uses an ML algorithm to show the products according to the user search. Facebook shows news feeds according to the likes of the user. In the market, there are plenty of ML applications that benefit the business in their needs. Virtual Personal Assistants like Siri, Alexa are using in everyday life. We can activate it on mobile and ask them to fix the alarm, schedule a meeting, book an appointment, etc. They learn from the use of the tool and improve its performance.

ML is use in the medical field as well. It is use to detect cancer and other deadly diseases. Early diagnosis of disease helps doctors to cure it rapidly. Medical Imaging is very complex to detect and also has a chance of human error as well. Machine learning helps to reduce human errors and detect and classify the disease using the medical images of the patients.

Online Transportation Networks like Uber, Cream are using ML to estimate their fare based on demand and weather conditions. Video Surveillance is also a popular system nowadays. It detects the unusual behavior of a person and gives an alert to the human attendant. It also providing fraud detection. PayPal is using ML for the detection of legal and illegal transactions and protection against money laundering. PayPal uses an algorithm that compares millions of transactions and differentiates it between legitimate and illegitimate transactions.


Understanding AI and ML are compulsory for everyone working in the Computer Science field. Corporates are now adopting artificial intelligence due to its ease of use in the field. This field is basically a combination of Statics, Computer Science, and logical thinking. Companies offer various positions like Data Analytics, Machine Learning Engineer, Data Scientist, etc.

ML is very useful in different fields like Medical, Geographic Information systems (GIS), Forecasting, Robotics, etc. Much work has already been done in Medical Imaging to detect and classify different cancers and their types. ML helps in the early detection of disease, which helps cure it in the early stages. ML algorithms are also using in the forecasting department. It gets data on humidity, temperature, and other factors and helps in the early detection of the weather.

You may also like to read: Artificial Intelligence in Marketing


Please enter your comment!
Please enter your name here

three × 5 =