|
|
| |
| The Performance of Ridge Regression, LASSO and Elastic Net in Modeling Market Value Data*
Download PDF
|
Nur Izzati Nabeela Adenan, Mazni Mohamad, Adzhar Rambli
|
Abstract:
Modelling market value is important in determining contributing factors and ensuring the well-being of a business or a company. Many studies employed ordinary least squares (OLS) regression in modelling market value. OLS regression is no longer appropriate when data is high dimensionality. An alternative to OLS regression, known as penalized regression, is utilized in this study since high-dimensionality market value data is to be analyzed. Ridge regression, LASSO, and elastic-net are the three techniques of penalized regression employed in this research. Their performance is compared by means of root mean square error (RMSE) and coefficient of determination, R2. Elastic net outperforms other techniques as it has the smallest RMSE and largest R2.
|
|
|
| | |