Machine Learning: Data Pre-processing

Data Quality

A measure of how well data meets a specific need, affecting the accuracy and reliability of predictions.

Data Pre-processing

The process of transforming raw data that contains inconsistencies into a format that improves model performance.

Accuracy

The degree to which data correctly represents an object in a real-world context.

Completeness

A measure of how many missing values a dataset contains.

Consistency

The absence of difference between instances stored in multiple locations.

Timeliness

The degree to which data represents reality at a given point or period in time.

Uniqueness

The absence of duplicate instances within a dataset.

Validity

How well data conforms to a specified format.

Data Cleaning

The process of removing missing values, duplicates, and incorrectly formatted data.

Data Integration

The process of combining data from different sources into a unified view.

Data Reduction

The process of reducing the dimensionality of a dataset, simplifying the data.

Data Transformation

The process of converting features into a format suitable for specific models or algorithms.

Missing Values

Occur when a feature has no recorded value for an instance in a dataset.

Missing Completely at Random (MCAR)

Values that have the same probability of being missing for all cases.

Missing at Random (MAR)

Values that have the same probability of being missing for specific observable cases.

Missing Not at Random (MNAR)