WGU D801 Machine Learning for Computer Scientists Comprehensive Study Guide | Algorithms,
Python, Model Evaluation & Key Formulas | 2026 Updated.
1. Machine Learning Foundations
1.1 What Is Machine Learning?
Machine learning is a field of artificial intelligence in which computer systems learn patterns
from data and use those patterns to make predictions or decisions.
Instead of programming every rule explicitly, the developer:
1. Collects data.
2. Selects relevant features.
3. Chooses an algorithm.
4. Trains a model.
5. Evaluates performance.
6. Deploys and monitors the model.
A model is a mathematical representation learned from training data.
1.2 Traditional Programming vs. Machine Learning
Traditional programming:
[
\text{Data} + \text{Rules} \rightarrow \text{Answers}
]
Machine learning training:
[
\text{Data} + \text{Answers} \rightarrow \text{Learned Model}
]
Machine learning prediction:
[
\text{New Data} + \text{Learned Model} \rightarrow \text{Predictions}
]
Use traditional programming when rules are clear and stable. Use machine learning when rules
are too numerous, complex, or dependent on patterns in historical data.
,2. Major Types of Machine Learning
2.1 Supervised Learning
Supervised learning uses labeled data. Every training example includes input features and a
known target.
[
(X,y)
]
Where:
(X) is the feature matrix.
(y) is the target vector.
Classification
Classification predicts a category.
Examples:
Spam or not spam
Fraudulent or legitimate
Disease present or absent
Product category
Customer churn class
Regression
Regression predicts a continuous numerical value.
Examples:
House price
Temperature
Monthly sales
Delivery time
Energy consumption
2.2 Unsupervised Learning
,Unsupervised learning uses data without target labels. Its purpose is to discover structure or
patterns.
Common tasks include:
Clustering
Dimensionality reduction
Density estimation
Anomaly detection
Association analysis
2.3 Semi-Supervised Learning
Semi-supervised learning combines:
A small amount of labeled data
A larger amount of unlabeled data
It is useful when obtaining labels is expensive but raw data is abundant.
2.4 Reinforcement Learning
An agent interacts with an environment and learns from rewards or penalties.
Important concepts:
Agent
Environment
State
Action
Reward
Policy
Value function
Exploration
Exploitation
The objective is generally to maximize cumulative reward.
3. Core Terminology
Term Meaning
Observation One row or instance in a dataset
, Term Meaning
Feature An input variable used for prediction
Target The value the model must predict
Label Known target associated with an observation
Model Learned mathematical relationship
Algorithm Procedure used to train the model
Parameter Value learned during training
Hyperparameter Configuration selected before training
Training Process of learning model parameters
Inference Using a trained model to make predictions
Epoch One full pass through the training data
Batch Subset processed during one training step
Loss function Measures prediction error during training
Metric Measures model performance
Generalization Performance on previously unseen data
Parameters vs. Hyperparameters
Parameters are learned from data:
Regression coefficients
Neural-network weights
Decision-tree split values
Hyperparameters are configured by the practitioner:
Learning rate
Number of trees
Maximum tree depth
Number of clusters
Regularization strength
(k) in k-nearest neighbors
4. The Machine-Learning Workflow
A sound workflow follows these stages:
1. Define the business or research problem.
2. Identify the prediction target.
3. Collect relevant data.
Python, Model Evaluation & Key Formulas | 2026 Updated.
1. Machine Learning Foundations
1.1 What Is Machine Learning?
Machine learning is a field of artificial intelligence in which computer systems learn patterns
from data and use those patterns to make predictions or decisions.
Instead of programming every rule explicitly, the developer:
1. Collects data.
2. Selects relevant features.
3. Chooses an algorithm.
4. Trains a model.
5. Evaluates performance.
6. Deploys and monitors the model.
A model is a mathematical representation learned from training data.
1.2 Traditional Programming vs. Machine Learning
Traditional programming:
[
\text{Data} + \text{Rules} \rightarrow \text{Answers}
]
Machine learning training:
[
\text{Data} + \text{Answers} \rightarrow \text{Learned Model}
]
Machine learning prediction:
[
\text{New Data} + \text{Learned Model} \rightarrow \text{Predictions}
]
Use traditional programming when rules are clear and stable. Use machine learning when rules
are too numerous, complex, or dependent on patterns in historical data.
,2. Major Types of Machine Learning
2.1 Supervised Learning
Supervised learning uses labeled data. Every training example includes input features and a
known target.
[
(X,y)
]
Where:
(X) is the feature matrix.
(y) is the target vector.
Classification
Classification predicts a category.
Examples:
Spam or not spam
Fraudulent or legitimate
Disease present or absent
Product category
Customer churn class
Regression
Regression predicts a continuous numerical value.
Examples:
House price
Temperature
Monthly sales
Delivery time
Energy consumption
2.2 Unsupervised Learning
,Unsupervised learning uses data without target labels. Its purpose is to discover structure or
patterns.
Common tasks include:
Clustering
Dimensionality reduction
Density estimation
Anomaly detection
Association analysis
2.3 Semi-Supervised Learning
Semi-supervised learning combines:
A small amount of labeled data
A larger amount of unlabeled data
It is useful when obtaining labels is expensive but raw data is abundant.
2.4 Reinforcement Learning
An agent interacts with an environment and learns from rewards or penalties.
Important concepts:
Agent
Environment
State
Action
Reward
Policy
Value function
Exploration
Exploitation
The objective is generally to maximize cumulative reward.
3. Core Terminology
Term Meaning
Observation One row or instance in a dataset
, Term Meaning
Feature An input variable used for prediction
Target The value the model must predict
Label Known target associated with an observation
Model Learned mathematical relationship
Algorithm Procedure used to train the model
Parameter Value learned during training
Hyperparameter Configuration selected before training
Training Process of learning model parameters
Inference Using a trained model to make predictions
Epoch One full pass through the training data
Batch Subset processed during one training step
Loss function Measures prediction error during training
Metric Measures model performance
Generalization Performance on previously unseen data
Parameters vs. Hyperparameters
Parameters are learned from data:
Regression coefficients
Neural-network weights
Decision-tree split values
Hyperparameters are configured by the practitioner:
Learning rate
Number of trees
Maximum tree depth
Number of clusters
Regularization strength
(k) in k-nearest neighbors
4. The Machine-Learning Workflow
A sound workflow follows these stages:
1. Define the business or research problem.
2. Identify the prediction target.
3. Collect relevant data.