Random Forest Classification of Bridge Damage from Visual Inspection Results and Structure Age

Project info

• Field of studyCivil Engineering and Transportation
• LevelDoctoral dissertation (S3)
• Method typeClassification
• Estimated build time± 7 working days
• Toolsscikit-learn RandomForestClassifier, confusion_matrix, roc_auc_score
• PriceUS$31 - US$200

Research goal

Classify bridge damage category using the features structure age and crack width with Random Forest Classification and assess its accuracy.

Variables

Example result

• 0.824Accuracy
• 0.778Precision
• 0.907Recall
• 0.838F1-score
• 0.905ROC AUC
• 0.54Importance Structu…

The random forest model classifies Bridge damage category (low/high) from Structure age and Crack width with test accuracy 0.824, F1 0.838 and AUC 0.905.

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

About Random Forest Classification

Random forest classification combines the votes of many trees trained on bootstrap samples and random feature subsets. It suits tabular classification needing stable accuracy with little tuning.

Steps

1. 1Split into stratified training and test sets
2. 2Train RandomForestClassifier with hundreds of trees
3. 3Tune max_depth, max_features, and class_weight if classes are imbalanced
4. 4Evaluate accuracy, F1, and AUC on the test set
5. 5Show the confusion matrix
6. 6Compute feature importance

Assumptions to check

• Training data represent the target population
• No target leakage in the features
• Feature importance can be biased toward high-cardinality features

Reading and reporting the results

Report test metrics, the confusion matrix, and the most influential features. Include the OOB score or cross-validation to make accuracy claims more convincing.

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 RandomForestClassifier
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score, f1_score, roc_auc_score, confusion_matrix, ConfusionMatrixDisplay

rng = np.random.default_rng(31)
n = 500
structure_age = rng.normal(0, 1, n)
crack_width = rng.normal(0, 1, n)
x3 = rng.normal(0, 1, n)
bridge_damage_category = ((structure_age ** 2 + crack_width + 0.3 * rng.normal(size=n)) > 0.8).astype(int)

X = np.column_stack([structure_age, crack_width, x3])
X_tr, X_te, y_tr, y_te = train_test_split(X, bridge_damage_category, test_size=0.25, stratify=bridge_damage_category, random_state=31)
rf = RandomForestClassifier(n_estimators=300, oob_score=True, random_state=31)
rf.fit(X_tr, y_tr)
pred = rf.predict(X_te)
prob = rf.predict_proba(X_te)[:, 1]
print("OOB accuracy:", round(rf.oob_score_, 3))
print("Test accuracy:", round(accuracy_score(y_te, pred), 3))
print("F1:", round(f1_score(y_te, pred), 3), "AUC:", round(roc_auc_score(y_te, prob), 3))
print("Importance:", rf.feature_importances_.round(3))

cm = confusion_matrix(y_te, pred)
ConfusionMatrixDisplay(cm).plot()
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