nuget
A matrix for make the sample input normalized.(进行所输入的样本数据的归一化的矩阵)
NormalizeMatrix
01 Syntax
Microsoft.VisualBasic.MachineLearning.ComponentModel.StoreProcedure.NormalizeMatrix
02 Methods
| Name | Overloads | Summary |
|---|---|---|
| DoNormalize | 1 | Normalize a single property value by its property name. |
| doNormalInternal | 1 | Normalize a single value with the given sample distribution and normalization method. |
| NormalizeInput | 2 | Normalize the sample inputs Sample.label to value range [0, 1] |
| CreateFromSamples | 2 | Create the normalization matrix from a SampleList object. |
03 Properties
04 Members
DoNormalize(
String, Double, Methods)Normalize a single property value by its property name.
Parameters
| Name | Type | Description |
|---|---|---|
name | String | The property name, it should be one of the elements in the NormalizeMatrix.names list. |
value | Double | The raw property value that will be normalized. |
method | Methods | The normalization method, the default value is |
Returns
The normalized value.
doNormalInternal(
SampleDistribution, Double, Methods)Normalize a single value with the given sample distribution and normalization method.
Parameters
| Name | Type | Description |
|---|---|---|
dist | SampleDistribution | The |
x | Double | The raw value that will be normalized. |
method | Methods | The normalization method, the default value is |
Returns
The normalized value.
NormalizeInput(Sample,
Methods)Normalize the sample inputs Sample.label to value range [0, 1]
Parameters
| Name | Type | Description |
|---|---|---|
sample | Sample | The target sample whose input vector will be normalized. |
method | Methods | The normalization method, the default value is |
Returns
An array of the normalized input values.
NormalizeInput(
IEnumerable(Of Double), Methods)Normalize the sample inputs Sample.label to value range [0, 1]
Parameters
| Name | Type | Description |
|---|---|---|
sample | IEnumerable(Of Double) | The target input vector which will be normalized, the element order should be identical with the NormalizeMatrix.matrix and the NormalizeMatrix.names list. |
method | Methods | The normalization method, the default value is |
Returns
An array of the normalized input values.
Create the normalization matrix from a SampleList object.
Parameters
| Name | Type | Description |
|---|---|---|
sampleList | SampleList | The training sample collection. |
names | IEnumerable(Of String) | The property names, not sample id names. |
estimateQuantile | Boolean | Whether the quantile value of the sample distribution should be estimated? The default value is |
Returns
A NormalizeMatrix object.
CreateFromSamples(
IEnumerable(Of Sample), IEnumerable(Of String), Boolean)神经网络会要求输入的属性值之间是可以直接进行比较的, 所以为了能够直接进行比较, 在这里将sample的每一个属性都按列归一化为[0,1]之间的结果
Parameters
| Name | Type | Description |
|---|---|---|
samples | IEnumerable(Of Sample) | A collection of the training Sample data. |
names | IEnumerable(Of String) | The property names, not sample id names |
Returns
A NormalizeMatrix object which contains the distribution of each property column of the given samples.
matrix
每一个属性都具有一个归一化区间
Returns
A list of the SampleDistribution objects, one for each property column.
names
属性名称列表,这个序列的长度是和NormalizeMatrix.matrix的长度一致的,并且元素的顺序一一对应的
Returns
An array of the property names.