Glossary

Key Terms and Definitions

Machine Learning Terms

Algorithm: A set of rules or instructions for solving a problem or making predictions.
Cross-validation: A technique for evaluating model performance by training and testing on different subsets of data.
Feature: An input variable used to make predictions (also called predictor or independent variable).
Feature Engineering: The process of creating new features from raw data to improve model performance.
Gradient Boosting: An ensemble method that builds trees sequentially, with each tree correcting errors from previous ones.
Hyperparameter: A setting that controls the learning process (e.g., number of trees, learning rate).
Overfitting: When a model learns the training data too well, including noise, and performs poorly on new data.
Random Forest: An ensemble method that combines many decision trees trained on random subsets of data.
Regularization: Techniques to prevent overfitting by penalizing model complexity (e.g., L1/Lasso, L2/Ridge).
Target: The output variable being predicted (also called dependent variable or label).
Train/Test Split: Dividing data into separate sets for training and evaluating the model.
Underfitting: When a model is too simple to capture patterns in the data.

Agent: An autonomous entity in a simulation that follows rules and interacts with its environment.
Agent-Based Model (ABM): A simulation approach where system behavior emerges from individual agent interactions.
Deterministic Model: A model that produces the same output given the same inputs (no randomness).
Emergence: Complex system-level patterns arising from simple individual-level rules.
Monte Carlo Simulation: Using random sampling to understand system behavior and uncertainty.
Stochastic Model: A model that includes random elements, producing different outputs each run.
System Dynamics (SD): A modeling approach focusing on feedback loops and flows between system components.

Buffer Stock: Extra inventory held to protect against uncertainty (also called safety stock).
Lead Time: The time between placing an order and receiving the goods.
Stockout: When inventory reaches zero and demand cannot be met.
Wastage: Loss of products due to expiration, damage, or other factors.
Reorder Point: The inventory level at which a new order should be placed.
Service Level: The probability of meeting demand without a stockout.

Categorical Variable: A variable with discrete categories (e.g., region, facility type).
Data Wrangling: The process of cleaning and transforming raw data for analysis.
Encoding: Converting categorical variables to numerical format for modeling.
Imputation: Filling in missing values using statistical methods.
Normalization/Scaling: Transforming variables to a common scale.
Outlier: An observation that differs significantly from other observations.

DataFrame: A two-dimensional data structure in pandas (like a spreadsheet).
Function: A reusable block of code that performs a specific task.
Library/Package: A collection of pre-written code for specific tasks (e.g., pandas, scikit-learn).
Loop: Code that repeats a set of instructions multiple times.
Variable: A named storage location for data in a program.