Study on settlement prediction of soft ground considering multiple feature parameters based on ISSA-RF model

Area of Science:

• Geotechnical Engineering
• Machine Learning
• Computational Intelligence

Background:

• Accurate settlement prediction is crucial for the success of preloading engineering projects on soft foundations.
• Existing settlement prediction models often face challenges in accuracy and efficiency.

Purpose of the Study:

• To develop a highly accurate settlement prediction model for soft foundations.
• To enhance the performance of the random forest (RF) model using an improved optimization algorithm.

Main Methods:

• A settlement prediction database was created using data from preloading engineering projects.
• An improved sparrow search algorithm (ISSA) was developed, incorporating chaotic mapping, adaptive weights, and Levy flight.
• The ISSA was used to optimize the hyperparameters of the random forest (RF) model, creating the ISSA-RF model.

Main Results:

• The ISSA demonstrated superior accuracy and stability compared to other optimization algorithms on benchmark functions.
• The ISSA-RF model showed significantly improved prediction accuracy and applicability over the standard RF model in practical applications.

Conclusions:

• The ISSA-RF model offers a robust and accurate solution for soft foundation settlement prediction.
• This approach provides valuable guidance for planning and executing preloading engineering projects effectively.