Parameter Object

A parameter is a variable that can affect a single experiment trial. SigOpt will suggest values for your parameters in order to optimize the relevant objective. double and int parameters represent numeric values. categorical parameters take on one of a fixed list of values (similar to an enum in many programming languages).

Fields

KeyTypeValue

bounds

A range of values that a parameter can occupy. Only applicable for parameters of type double or int.

categorical_values

A list of categorical values. Only applicable for parameters of type categorical.

conditions

default_value

string or number

Used to replace missing parameter assignments in parameters added after initial experiment creation.

grid

array<double>

Only applicable for parameters of type double and int. The discrete numerical values that a parameter can take. When grid is provided, bounds should not be set. Refer to the Grid Values documentation for more information.

name

string

A user-specified name for this parameter.

precision

int

Only applicable for parameters of type double. The number of decimal places when displaying on the web dashboard.

prior

Only applicable for parameters of type double. The prior distribution of the parameter. Refer to the Prior Beliefs documentation for more information.

transformation

string

Only applicable to parameters of type double. The scale of a parameter. Refer to the Parameter transformation documentation for more information.

type

string

The type of this parameter. Must be one of double, int, or categorical.

Examples

Double Parameter

{
  "bounds": {
    "max": 1,
    "min": 0.001,
    "object": "bounds"
  },
  "categorical_values": null,
  "conditions": {},
  "default_value": null,
  "name": "gamma",
  "object": "parameter",
  "precision": null,
  "prior": null,
  "transformation": null,
  "type": "double"
}

Integer Parameter

{
  "bounds": {
    "max": 5,
    "min": 1,
    "object": "bounds"
  },
  "categorical_values": null,
  "conditions": {},
  "default_value": null,
  "name": "degree",
  "object": "parameter",
  "precision": null,
  "prior": null,
  "transformation": null,
  "type": "int"
}

Categorical Parameter

{
  "bounds": null,
  "categorical_values": [
    {
      "enum_index": 1,
      "name": "rbf",
      "object": "categorical_value"
    },
    {
      "enum_index": 2,
      "name": "poly",
      "object": "categorical_value"
    },
    {
      "enum_index": 3,
      "name": "sigmoid",
      "object": "categorical_value"
    }
  ],
  "conditions": {},
  "default_value": null,
  "name": "kernel",
  "object": "parameter",
  "precision": null,
  "prior": null,
  "transformation": null,
  "type": "categorical"
}

Last updated