Glossary¶

Use this page to record terms and ideas that help you understand professional analytics projects.

This project covers feature engineering and selection: preparing raw data so a machine learning model can learn from it effectively.

Pro-tip: Expand the VS Code Outline view (below the navigator on the right) to see this file organization at-a-glance.

Data and Outputs¶

raw data¶

Raw data is the original input data used by the project. It should usually be kept unchanged so the analysis can be repeated from the original source.

processed data¶

Processed data is data that has been cleaned, filtered, transformed, summarized, or prepared for later use.

artifact¶

An artifact is a file created by running a project. Examples include logs, charts, reports, and exported data files.

Feature Engineering¶

feature engineering¶

Feature engineering is the process of creating, transforming, or selecting input columns to improve model performance. Good features often matter more than the choice of model.

missing values¶

Missing values are cells in a dataset with no recorded entry. Common strategies include removing rows with missing values, filling them with the mean or median, or using a placeholder category.

imputation¶

Imputation is the process of filling in missing values with estimated ones. The choice of imputation strategy is an analyst decision.

encoding¶

Encoding is the process of converting categorical text values into numbers so a model can work with them. One-hot encoding creates a separate binary column for each category. Ordinal encoding assigns an integer to each category.

one-hot encoding¶

One-hot encoding converts a categorical column into multiple binary columns, one per category. Each row has a 1 in the column for its category and 0 in all others.

scaling¶

Scaling transforms numeric features so they fall within a similar range. This prevents features with large values from dominating those with small values.

standardization¶

Standardization rescales a feature so it has a mean of 0 and a standard deviation of 1. It is a common scaling method when features have different units or ranges.

normalization¶

Normalization rescales a feature so all values fall between 0 and 1. It is useful when the model is sensitive to absolute value ranges.

feature selection¶

Feature selection is the process of choosing which input columns to keep and which to remove before training a model. Removing irrelevant or redundant features can improve accuracy and reduce complexity.

correlation¶

Correlation measures how closely two variables move together. High correlation between a feature and the target suggests the feature may be useful for prediction. High correlation between two features suggests one may be redundant.

outlier¶

An outlier is a data point that is very different from the rest. Outliers can distort model training and may need to be removed or capped.