August 25, 2025Deleted JournalOpen Access

Optimizing multivariate adaptive regression splines (MARS) with coordinate descent to accurately select the best model for house price prediction

View Full Paper

Ask AI

Bookmark

The proposed CD-MARS model shows a mean squared error significantly lower than traditional MARS, enhancing predictive accuracy.
The new model achieved an R-squared value of 0.931, outperforming the traditional MARS score of 0.8952 by 3.58%.
Using coordinate descent as a penalty in model selection improves generalizability and reduces over-fitting.
House prices dataset analysis demonstrates the effectiveness of CD-MARS in non-parametric regression applications.

Discussion

No takes yet. Share an insight, caveat, or question.

Proposed CD-MARS model improves R-squared to 93.1% in house price prediction, indicating effective model selection.

Adonia et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark