Enum DataComplexity

Namespace

AiDotNet.Enums

Assembly

AiDotNet.dll

Represents the level of complexity in a dataset, which helps determine appropriate model selection and preprocessing.

public enum DataComplexity

Fields

Complex = 2

Indicates data with intricate patterns, many features, significant noise, or complex dependencies.

Complex data typically has these characteristics: - Highly non-linear patterns - Many features (typically 50+) - Significant feature interactions - High dimensionality - Substantial noise or outliers - Significant missing data - Severe class imbalance (for classification problems) - Temporal or spatial dependencies

Examples include:

• Natural language processing tasks
• Image or video recognition
• Financial market prediction
• Genomic data analysis
• Complex sensor data with multiple sources

Complex data often requires deep learning models, sophisticated feature engineering, or specialized algorithms designed for the specific domain.

Moderate = 1

Indicates data with somewhat complex patterns, a moderate number of features, and some noise.

Moderate complexity data typically has these characteristics: - Somewhat non-linear patterns - Moderate number of features (typically 10-50) - Some feature interactions - Moderate amount of noise - Some missing data - Slight class imbalance (for classification problems)

Examples include:

• Customer purchase prediction with demographic and behavioral data
• Housing price prediction with multiple factors
• Seasonal time series with multiple patterns

Moderate complexity data often requires ensemble methods, shallow neural networks, or more sophisticated algorithms.

Simple = 0

Indicates data with clear patterns, few features, and minimal noise.

Simple data typically has these characteristics: - Linear or easily separable patterns - Few relevant features (typically less than 10) - Clear relationships between inputs and outputs - Minimal noise or outliers - Little to no missing data - Balanced classes (for classification problems)

Examples include:

• Height vs. weight relationships
• Basic customer segmentation with clear groups
• Simple time series with obvious trends

Simple data often works well with linear models, decision trees, or basic statistical methods.

Remarks

For Beginners: Data complexity refers to how difficult it is for a machine learning model to find patterns in your data.

Think of it like solving puzzles:

• Simple data is like a basic jigsaw puzzle with few, large pieces
• Complex data is like an advanced puzzle with many tiny pieces and subtle patterns

Understanding your data's complexity helps you choose the right model:

• Simple data often works well with basic models
• Complex data usually requires more sophisticated models

This enum helps you categorize your data to make better decisions about which algorithms to use.