Constructor
new SessionData(proofPercentage, dataSet, _useDefaultStandardization, _callbackStandardizeopt)
Creates an instance of SessionData.
Parameters:
| Name | Type | Attributes | Description |
|---|---|---|---|
proofPercentage |
number | A value 0-1 exclusive used to determine
the number of cases reserved for generalization testing. These cases
are never seen by the model during training. A common value is 0.2 (20%). |
|
dataSet |
DataSet | The data used to train and test models. | |
_useDefaultStandardization |
boolean | If true, the input values
will be modified internally such that each feature has a mean of zero
and a variance of one. If standardization callbacks are supplied, this argument is ignored. |
|
_callbackStandardize |
function |
<optional> |
A function invoked with the
input data prior to the grid search. It provides an opportunity to
preprocess the data internally before it's transformed into tensors. Arguments: unstandardizedInputs: Array Returns: void |
- Source:
Methods
(static) CountLeafElements(inputData)
Returns the length of the most deeply nested array.
Parameters:
| Name | Type | Description |
|---|---|---|
inputData |
TFNestedArray |
- Source:
(static) FindMean(data) → {number}
Calculate the average of a set of numbers.
Parameters:
| Name | Type | Description |
|---|---|---|
data |
Array.<number> |
Returns:
- Type
- number
- Source:
(static) FindStandardDeviation(data, mean) → {number}
Calculate the standard deviation of a set of numbers.
Parameters:
| Name | Type | Description |
|---|---|---|
data |
Array.<number> | |
mean |
number |
Returns:
- Type
- number
- Source:
(static) StandardizeInputs(inputData)
Adjusts input data such that each feature has a mean of zero and a variance
of one.
Parameters:
| Name | Type | Description |
|---|---|---|
inputData |
TFNestedArray |
- Source: