@article {10.3844/ajeassp.2025.55.66,
article_type = {journal},
title = {Ecological Function Evaluation of Soil and Water Conservation Measures for Power Transmission and Transformation Project Construction Based on Remote Sensing and it’s Improvement Path},
author = {Han, Yang and Wang, Songsong and Yang, Yanhui and Rong, Xing and Zhaokun, Zhang and Haoran, Wang},
volume = {18},
number = {2},
year = {2025},
month = {Mar},
pages = {55-66},
doi = {10.3844/ajeassp.2025.55.66},
url = {https://thescipub.com/abstract/ajeassp.2025.55.66},
abstract = {Assessing ecological functions is critical for determining the efficacy of soil and water conservation measures, especially during the development of power transmission and transformation projects. Efficient conservation practices are essential for reducing ecological effects and guaranteeing sustainable development. This research aims to create a resilient model, Ecological Function Prediction for Conservation Effectiveness (EFP-CE) that will classify conservation effectiveness as either effective or ineffective. The goal is to improve prediction accuracy and offer actionable insights for better conservation tactics. The EFP-CE model combines many analytical methods: Missing values are imputed using Support Vector Regression (SVR) and outliers are detected and removed using Euclidean distance. Categorical variables are converted using label encoding, while numerical attributes are subjected to Min-Max normalization. An ensemble feature selection technique integrates filter and wrapper methods to find important predictors, while cluster-based oversampling fixes data imbalance. The dataset is separated into training and testing sets. A Bagged Gradient Boosting model is trained and assessed to forecast conservation efficiency. The proposed model was evaluated using a ten ecological function assessment attributes dataset. The Bagged Gradient Boosting model obtained 93% accuracy, 91% precision, 89% recall, an F1-score of 90%, and a Matthews Correlation Coefficient (MCC) of 82%, suggesting strong predictive effectiveness in evaluating conservation measures. The EFP-CE model demonstrates how machine learning methods can be integrated to improve the assessment of conservation measures. By enhancing prediction accuracy, this research presents helpful knowledge for policymakers and stakeholders participating in environmental safety during infrastructure projects, eventually adding to more sustainable construction procedures.},
journal = {American Journal of Engineering and Applied Sciences},
publisher = {Science Publications}
}