What Is Machine Learning?
Machine learning (ML) is a branch of artificial intelligence (AI) that focuses on developing computer systems capable of learning from data, identifying patterns, and making decisions or predictions with minimal human intervention. Instead of being explicitly programmed with step-by-step instructions for every task, a machine learning system is designed to improve its performance automatically as it is exposed to more data and experience.
At its core, machine learning involves training algorithms—sets of rules or models—on large datasets. These algorithms learn to recognize patterns or relationships within the data, enabling them to make predictions or classifications when presented with new, unseen data. For example, a machine learning model trained on thousands of labeled images of cats and dogs can learn to distinguish between the two, even when shown new images it has never encountered before.
Types of Machine Learning
Machine learning can be broadly categorized into three main types, each defined by how the algorithm learns from data:
• Supervised Learning: The model is trained on labeled data, meaning each training example is paired with the correct answer. The algorithm learns to map inputs to outputs and can then predict outcomes for new, unseen data. Common applications include image recognition and spam detection.
• Unsupervised Learning: The model works with unlabeled data and tries to find hidden patterns or groupings within the data. Clustering customers by purchasing behavior or reducing the complexity of data are typical uses.
• Reinforcement Learning: The model learns by interacting with an environment, receiving feedback in the form of rewards or penalties. Its goal is to maximize the cumulative reward over time. This approach is often used in robotics, game playing, and autonomous vehicles.
ML a-z
A
Term
Definition
B
Bagging
AKA, Bootstrap aggregating. This is an ensemble learning technique that is used to reduce variance in a noisy dataset to improve the model’s accuracy.Bootstrapping
Bootstrapping is a sampling technique used to create diverse samples. Different subsets of training data are created by selecting data points at random and repeating them. This means you are able to select the same instance multiple times causing a sample to occur more often.Bootstrap aggregating
See Bagging.
C
Term
Definition
D
Term
Definition
E
Ensemble learning
Ensemble learning combines predictions from multiple models in order to improve predictions.
F
Feature
A feature is an input variable that is used to make the prediction.Feature engineering
The process of determining which features, or input variables, will be useful in training a model, then turning (converting) those features into usable data for the model.Feature set
Refers to the group of features (input variables) you use to train your model. For example, zip code, property size, and property condition might be a feature set used to predict the price of a house.
G
Term
Definition
H
Term
Definition
I
Term
Definition
J
Term
Definition
K
Term
Definition
L
Label
In supervised learning, a label is the answer, result, or prediction made from modelling features.
M
Term
Definition
N
Term
Definition
O
Refers to a model that so closely matches the training data, it fails to make the correct prediction on new data. It happens when the model includes so much noise in the training data to the extent that it negatively impacts performance of the model on new data.
P
Term
Definition
Q
Term
Definition
R
Term
Definition
S
Term
Definition
T
Term
Definition
U
Underfitting refers to a model that is not good enough to model the data nor generalize new new data. It is of course, not a suitable model.
V
Term
Definition
W
Term
Definition
X
Term
Definition
Y
Term
Definition
Z
Term
Definition
