Feature Scaling (DP_5)

This functionality normalizes the range of values for different features in the dataset

This process is very important because many machine learning algorithms use the Euclidean distance between two points in their computations, and if the ranges of the features are vastly different, then the algorithm will be sensitive to the feature with the larger range, and may produce unexpected results.

Sample Request

This request scales the feature sets defined in the range of columns for training and testing.

{
    "project_id": 1,
    "parent_id": 4,
    "block_id": 5,
    "function_code": "DP_5",
    "args": {
        "dataset": true,
        "xtrain": true,
        "xtest": true,
        "x": true,
        "ytrain": true,
        "ytest": true,
        "y": true,
        "scaler": "maxabs",
        "boundariestoscale": ":, 2:",
        "save_name": "testweights",
        "load_name": "testweights"
    }
}

Missing Data

Encode categorical values

POST https://autogon.ai/api/v1/engine/start

Encodes categorical data on specific columns with specified boundaries

Request Body

Name Type Description

Name	Type	Description
project_id*	int	current project ID
parent_id*	int	parent block ID
block_id*	int	current block ID
function_code*	String	block's function code
encode*	boolean	specify if variable is encoded
args	object	block arguments
remainder*		applied method to none specified columns; `drop` drops the unspecified columns for encoding, `passthrough` ignores unspecified columns
index*	int	column index to apply encoding technique
scaler	String	Type of scaler to use: `'minmax'` : transforms data to a 0-1 range. `'standard'` : transforms data to have zero mean and unit variance. `'robust'` : scaling technique that is less sensitive to outliers in the data. `'maxabs'` : scaling technique that scales the data by dividing each feature by its maximum absolute value, preserving the sign of the values while normalizing them.
save_name		name to save processing models with.
load_name	String	name to load processing models with. Used to switch to loading mode

project_id*

int

current project ID

parent_id*

int

parent block ID

block_id*

int

current block ID

function_code*

String

block's function code

encode*

boolean

specify if variable is encoded

args

object

block arguments

remainder*

applied method to none specified columns; drop drops the unspecified columns for encoding, passthrough ignores unspecified columns

index*

int

column index to apply encoding technique

scaler

String

Type of scaler to use: 'minmax' : transforms data to a 0-1 range.

'standard' : transforms data to have zero mean and unit variance.

'robust' : scaling technique that is less sensitive to outliers in the data.

'maxabs' : scaling technique that scales the data by dividing each feature by its maximum absolute value, preserving the sign of the values while normalizing them.

save_name

name to save processing models with.

load_name

String

name to load processing models with. Used to switch to loading mode

{
    "status": "true",
    "message": {
        "id": 3,
        "project": 1,
        "block_id": 3,
        "parent_id": 2,
        "dataset_url": "",
        "x_value_url": "",
        "y_value_url": ""
    }
}

// Some code

// Some code

PreviousData Split (DP_4)NextDrop Columns (DP_6)

Last updated 1 year ago