WooDensECEfier: A Machine Learning-Driven Approach for Accurate Wood Density Prediction Using Random Forest Regression Analysis

The wood industry is essential to the world economy as it provides materials for construction, furniture, and more. However, the main challenge in this industry is to estimate wood density precisely, which directly influences its quality and market value. The researchers designed different machine learning models, such as Decision Tree Regression, Support Vector Regression, Ridge Regression, Random Forest Regression, Gradient Boosting Regression, and Partial Least Squared Regression. To determine the most precise model, cross-validation and holdout validation of regression analysis with the metrics Mean Squared Error, Root Mean Squared Error, Mean Absolute Error, R-squared, and Explained Variance Score were used. The Random Forest Regression model was considered the most efficient in predicting the wood density of the CIRAD dataset, gaining $\mathbf{99.88 \%}$ and $\mathbf{99.84 \%}$ for R2 scores on cross and holdout validation of optimized parameters, respectively. This research advances the understanding of how wood density can help to be a predictive measure for wood characteristics and strength, holding significant importance in the wood industry, especially among furniture manufacturers.