HomePage esProc Function Reference Functions lasso() ＜Previous | Next ＞

lasso()

Read（1176） Label: model, prediction, vector,

Description:

Build models and performs predictions using the Lasso regression method.

Syntax:

lasso(X, Y, learning_rate, iterations)

The function fits together matrix X and vector Y using the Lasso regression method and returns model information that includes coefficient matrix and intercept. The model information can act as parameter F in lasso(X’, F) to perform a fitting computation

lasso(X’, F) 

The function fits together two matices that have same number of columns – that is, perform predicitions on another matrix X’ using model F, and returns a vector

Note:

An external library function (See External Library Guide) that builds models and performs predictions using the Lasso regression method.

Parameter:

X

A matrix

Y

A vector having the same number of rows as matrix X

learning_rate

Learning rate that is a decimal between 0 and 1; default value is 0.01

iterations

Number of iterations; default is 1000

X’

A matrix that has same number of columns as matrix X

F

The return result of lasso (X, Y, learning_rate, iterations)

Return value:

Matrix/Vector

Example:

 

A

 

1

[[19,1],[25,1],[31,1],[38,1],[44,1]]

 

2

[19,32.3,49,73.3,97.8]

 

3

=lasso(A1,A2,0.001,10000)

Fit A1 and A2 together using Lasso regression method and return coefficient matrix A3(1) and intercept A3(2).

4

=lasso(A1,A3)

Perform prediction on A1 using model A3; the result can be compared with actual values in A2.