Skip to main content

Machine Learning

Artificial Intelligence (AI) is the process of programming or developing a system so computers is able to do tasks that typically require human intelligence. Machine Learning (ML) is a subfield of AI that specifically makes them learn by themselves from experience or data. Learning from data means that the computers will analyze pattern in data using statistical techniques, this pattern will be used to make predictions or decisions.

Machine learning is divided by 3 based on how we teach or train the program (often called model):

  • Supervised Learning: This approach involves training the model on labeled data, which means the dataset used has known input and output values. The goal is to enable the model to recognize patterns and make predictions for new input instances where the corresponding output is unknown.

  • Unsupervised Learning: On the other hand, this approach involves training model on unlabeled data. The dataset consist of input data without corresponding output. The objective is to discover patterns, structures, or relationships in the data without explicit guidance. These are typically used to identify clusters or similarity between each data in the dataset.

  • Reinforcement Learning: Reinforcement learning is a machine learning paradigm where we have an entity or system (often called agent) that interacts with an environment. The system will explore different actions in the environment, there will be rewards or penalties for each actions done depending on whether we want them to do it or not. Reinforcement learning is based on trial and error by utilizing rules and reward, the objective is to get as many reward as possible.

This notes will include explanation about traditional ML algorithms (implementation not included).

Main Source:

All pages​