Machine Learning

Machine learning (ML) is a subset of artificial intelligence that involves the development of algorithms and statistical models enabling computers to perform tasks without explicit instructions. These systems learn from data, identifying patterns and making decisions based on the information they process.

Machine learning models can be trained using various methods, including supervised learning, where the model learns from labelled data, and unsupervised learning, where the model identifies patterns in unlabelled data.

Supervised v unsupervised models

Supervised machine learning

Supervised approaches involve training a model on a labelled dataset, where each input is paired with the correct output. The model learns to map inputs to outputs by identifying patterns in the data, and its performance is evaluated based on its ability nbv canto predict the correct labels for new, unseen data. This approach is commonly used for tasks such as classification, where the goal is to assign inputs to predefined categories (e.g., spam detection in emails), and regression, where the aim is to predict continuous values (e.g., house price prediction).

All these approaches need a set of correctly-labelled data to train and test the model. This data has to be split into a training set and a testing set, and the model adjusted and retrained until the results on the test set reach an acceptable level of accuracy.

Unsupervised machine learning

Unsupervised approaches deal with unlabelled data, meaning the model is not provided with explicit instructions on what to learn. Instead, it seeks to identify inherent structures and patterns within the data.

Common techniques include clustering, where the model groups similar data points together (e.g., customer segmentation), and dimensionality reduction (e.g. PCA), which simplifies data by reducing the number of features while preserving important information.

Unsupervised learning is particularly useful for exploratory data analysis and discovering hidden patterns without prior knowledge of the data's structure.

Key Machine Learning algorithms

Linear Regression : This algorithm models the relationship between a dependent variable and one or more independent variables using a linear equation. It is commonly used for predicting numerical values, such as housing prices or sales forecasts.

Logistic Regression : This technique is used for binary classification problems, modelling the probability of a binary outcome using a logistic function. Typical use-cases include spam detection and medical diagnosis.

Cluster Analysis : Cluster analysis is a type of unsupervised machine learning used to group similar data points together based on their features.

Decision Trees : Decision trees split data into branches based on feature values, leading to a decision at each node. They are often used in classification tasks, such as customer segmentation and fraud detection.

Random Forests : An ensemble method that combines multiple decision trees to improve accuracy and prevent overfitting. It is used in applications like predicting loan defaults and disease outbreaks.

Support Vector Machines : Support Vector Machines (SVM) classify data by finding the optimal hyperplane that separates different classes. They are effective in image recognition and text categorisation.

K-Nearest Neighbours : K-Nearest Neighbours (KNN) classifies data points based on the majority class of their nearest neighbours. It is used in recommendation systems and pattern recognition.

Principal Component Analysis : Principal Component Analysis (PCA) reduces the dimensionality of data while preserving its variance, simplifying complex datasets. It is commonly used in data visualisation and noise reduction.

Neural Networks : Inspired by the human brain, neural networks consist of interconnected nodes that process data in layers. They are used in deep learning applications like speech recognition and autonomous driving.

Gradient Boosting Machines (GBM) : GBM builds models sequentially, correcting errors of previous models to improve performance. It is used in predictive modelling tasks such as credit scoring and customer churn prediction.