Skip to main content
Financial Innovation

Table 2 Details of the machine learning cases tested in this study

From: Predicting the returns of the US real estate investment trust market: evidence from the group method of data handling neural network

Case

Training Samples

Testing samples

Kernel function

Input variables

Linear-30/30

30

30

linear function

All variables

Linear-100/30

100

30

linear function

All variables

Linear-200/30

200

30

linear function

All variables

Linear-300/30

300

30

linear function

All variables

Linear-600/30

600

30

linear function

All variables

Linear-300/60

300

60

linear function

All variables

Linear-cov-30/30

30

30

linear covariance function

All variables

Linear-cov-100/30

100

30

linear covariance function

All variables

Linear-cov-200/30

200

30

linear covariance function

All variables

Linear-cov-300/30

300

30

linear covariance function

All variables

Linear-cov-300/60

300

60

linear covariance function

All variables

Quadratic-30/30

30

30

quadratic function

All variables

Quadratic-100/30

100

30

quadratic function

All variables

Quadratic-200/30

200

30

quadratic function

All variables

Quadratic-300/30

300

30

quadratic function

All variables

Quadratic-600/30

600

30

quadratic function

All variables

Quadratic-300/60

300

60

quadratic function

All variables

  1. Note: The linear function, linear covariance function, quadratic function indicates the kernel functions described in Eqs. 2, 3, and 4, respectively
Back to article page