Random Forest Regression to Estimate Mangrove Aboveground Biomass from Satellite Vegetation Index

Project info

• Field of studyForestry and Environment
• LevelMaster's thesis (S2)
• Method typeRegression
• Estimated build time± 5 working days
• Toolsscikit-learn RandomForestRegressor, feature_importances_, train_test_split
• PriceUS$22 - US$144

Research goal

Model aboveground biomass from ndvi and canopy height using Random Forest Regression and evaluate the model performance.

Variables

Example result

• 0.851R2 (test)
• 0.843OOB R2
• 1.624RMSE (test)
• 1.286MAE (test)
• 0.838Top importance
• 150Trees

Random forest regression predicts Aboveground biomass with test R2 0.851 and RMSE 1.624; the most important predictor is NDVI (importance 0.838).

Computed from illustrative synthetic data, not from a real study.

About Random Forest Regression

Random forest regression averages many decision trees trained on bootstrap samples, capturing nonlinear relations and interactions. It suits theses with tabular data whose patterns are not linear.

Steps

1. 1Prepare features and target, then split into training and test sets
2. 2Train RandomForestRegressor with enough trees
3. 3Tune max_depth and min_samples_leaf if it overfits
4. 4Evaluate R2 and RMSE on the test set (and the OOB score)
5. 5Compute feature importance
6. 6Plot actual versus predicted and the importance bars

Assumptions to check

• Training data represent the data to be predicted
• No target information leaks into the features
• The model cannot extrapolate beyond the training range

Reading and reporting the results

Report R2, RMSE, and feature importance ranking as an indication of dominant variables, not causation. Compare with a linear baseline to show the added value.

Example Python code

A short starting script with this project's variable names. The delivered version is adapted to your data and fully commented.

import numpy as np
import matplotlib.pyplot as plt
from sklearn.ensemble import RandomForestRegressor
from sklearn.model_selection import train_test_split
from sklearn.metrics import r2_score, mean_squared_error

rng = np.random.default_rng(7)
n = 400
ndvi = rng.uniform(0, 10, n)
canopy_height = rng.uniform(0, 5, n)
x3 = rng.normal(size=n)
aboveground_biomass = np.sin(ndvi) * 3 + canopy_height ** 2 + 0.2 * x3 + rng.normal(0, 0.5, n)

X = np.column_stack([ndvi, canopy_height, x3])
X_tr, X_te, y_tr, y_te = train_test_split(X, aboveground_biomass, test_size=0.2, random_state=7)
rf = RandomForestRegressor(n_estimators=300, min_samples_leaf=3, oob_score=True, random_state=7)
rf.fit(X_tr, y_tr)
pred = rf.predict(X_te)
print("OOB R2:", round(rf.oob_score_, 3))
print("Test R2:", round(r2_score(y_te, pred), 3))
print("Test RMSE:", round(mean_squared_error(y_te, pred) ** 0.5, 3))
print("Importance:", rf.feature_importances_.round(3))

fig, ax = plt.subplots(1, 2, figsize=(10, 4))
ax[0].scatter(y_te, pred)
ax[0].plot([y_te.min(), y_te.max()], [y_te.min(), y_te.max()], "r--")
ax[0].set_xlabel("Actual"); ax[0].set_ylabel("Predicted")
ax[1].bar(["x1", "x2", "x3"], rf.feature_importances_)
ax[1].set_title("Feature importance")
plt.tight_layout()
plt.show()

Packages and prices

Choose how much help you need. Each higher package includes everything in the lower ones. Prices exclude 11% VAT.

• Payment: 50% down payment before work starts, balance after the deliverables are handed over
• Tax: Prices exclude 11% VAT
• Data and academic integrity: Results are computed from your own research data; you write and stand behind your manuscript, we help with the analysis and your understanding of it
• Revisions: Up to 2 revisions are included in the price (Package 3: 3 rounds)

Frequently asked questions