ColumnMetadata
Models & Types · AiDotNet.NeuralNetworks.SyntheticData
Describes the metadata for a single column in a tabular dataset, including its name, data type, categories (for categorical columns), and summary statistics.
For Beginners
Think of this as a “column profile” that describes everything the generator needs to know about one column in your table:
- Name: A human-readable label (e.g., “Age”, “Income”)
- DataType: Whether it’s continuous, discrete, or categorical
- Categories: For categorical columns, the list of possible values
- Statistics: Min, max, mean, and standard deviation (computed during fitting)
Example:
How It Works
Column metadata is used by synthetic data generators to understand the structure of each
column and apply the appropriate preprocessing (e.g., VGM normalization for continuous,
one-hot encoding for categorical). Statistics are populated during the Fit step.
Constructors
| Constructor | Summary |
|---|---|
ColumnMetadata(string name, ColumnDataType dataType, IEnumerable<string> categories, int columnIndex) |
Initializes a new instance of the ColumnMetadata class. |
Properties
| Property | Summary |
|---|---|
ColumnDataType DataType { get; set; } |
Gets or sets the data type of the column. |
IReadOnlyList<string> Categories { get; set; } |
Gets or sets the list of category values for categorical columns. |
bool IsCategorical { get; } |
Gets whether this column is categorical. |
bool IsNumerical { get; } |
Gets whether this column is numerical (continuous or discrete). |
double Max { get; set; } |
Gets or sets the maximum observed value for numerical columns. |
double Mean { get; set; } |
Gets or sets the mean (average) value for numerical columns. |
double Min { get; set; } |
Gets or sets the minimum observed value for numerical columns. |
double Std { get; set; } |
Gets or sets the standard deviation for numerical columns. |
int ColumnIndex { get; set; } |
Gets or sets the index of this column in the original data matrix. |
int NumCategories { get; } |
Gets the number of categories for categorical columns. |
string Name { get; set; } |
Gets or sets the name of the column. |
Methods
| Method | Summary |
|---|---|
ColumnMetadata Clone() |
Creates a deep copy of this column metadata. |