First, we gather and transform data, next step is related to data science when we convert data from one format to another. Then the gathered data is used to make predictions and derive patterns and trends by using ML techniques.

Unsupervised and supervised ML techniques are used to extract data and make predictions. This is when data analysis is used, a process that sits under data science. AI is then using your data to perform some actions, based on machine automation and human decisions.

• AI (Artificial Intelligence) – The broadest field; any system that mimics human intelligence.

• Data Science – Encompasses statistics, analysis, ML, AI, and data processing to extract insights.

• Machine Learning (ML) – A subset of AI that allows systems to learn patterns from data without explicit programming.

• Deep Learning (DL) – A subset of ML using multi-layered neural networks for complex tasks like image recognition and NLP..

Difference between ML and traditional programming approach

In the conventional approach to programming, developers manually define the logic that the computer follows. You provide:

• Explicit instructions (rules) such as conditions, loops, and functions

• Input data

The system then processes the input using these rules to generate an output.

Machine Learning flips this process. Instead of coding the rules yourself, you provide:

• Examples of inputs and their correct outputs (labelled data)

The ML system then automatically learns the underlying patterns or rules from the data and builds a model. This model can then predict outputs for new, unseen inputs.

Example:

To classify emails as spam or not, you supply the system with a large dataset of emails and their labels (spam or not). The algorithm finds patterns in the text and learns how to classify future emails without needing hardcoded rules.

Traditional programming is ideal when the problem-solving steps can be clearly described. On the other hand, machine learning excels in situations where the solution relies on detecting patterns in large datasets—especially when those patterns are too intricate to define manually.