## ALS Forest Regression Analysis in LiDAR360

### Introduction

The basic idea of regression analysis is to first build regression models using independent variables (average elevation, maximum elevation, minimum elevation, elevation percentiles, etc.) and dependent variables (growing stock volume, biomass, etc.) from sample datasets, and then utilize the regression models and characteristic variables of the point cloud data to estimate growing stock volume, biomass, and other metrics of the sample plot or larger areas. Popescu et al.(2004) has shown that regression analysis can achieve high accuracy in its predictions. However, to build a reliable regression model, large quantities of survey data are required.

There are four regression algorithms in LiDAR360 to choose from: Linear Regression, Support-Vector Machine (SVM), Artificial Neural Network (ANN), and Random Forest. The exercises in this tutorial cover the estimation of biomass in LiDAR360 using **ALS Forest** module’s **Regression Analysis** functions coupled with survey data from sample plots. Please install the LiDAR360 software, download required sample data, and use it to complete the exercises following instructions.

### Software requirement

Please download the latest version of LiDAR360 from the GreenValley International official website, and install and activate following instructions in the User Guide.

### Sample data

Please download the required sample dataset RegressionAnalysisSampleData for exercises in this tutorial.

Your sample dataset should consist of the following data files:

- ALSData.LiData: ALS forest point cloud dataset
- SampleData.txt: Sample data