A robust prediction model for evaluation of plastic limit based on sieve # 200 passing material using gene expression programming

Abstract

This study aims to propose a novel and high-accuracy prediction model of plastic limit (PL) based on soil particles passing through sieve # 200 (0.075 mm) using gene expression programming (GEP). PL is used for the classification of fine-grained soils which are particles passing from sieve # 200. However, it is conventionally evaluated using sieve # 40 passing material. According to literature, PL should be determined using sieve # 200 passing material. Although, PL200 is considered the accurate representation of plasticity of soil, its’ determination in laboratory is time consuming and difficult task. Additionally, it is influenced by clay and silt content along with sand particles. Thus, artificial intelligence-based techniques are considered viable solution to propose the prediction model which can incorporate multiple influencing parameters. In this regard, the laboratory experimental data was utilized to develop prediction model for PL200 using gene expression programming considering sand, clay, silt and PL using sieve 40 material (PL40) as input parameters. The prediction model was validated through multiple statistical checks such as correlation coefficient (R2), root mean square error (RMSE), mean absolute error (MAE) and relatively squared error (RSE). The sensitivity and parametric studies were also performed to further justify the accuracy and reliability of the proposed model. The results show that the model meets all of the criteria and can be used in the field.

Predictive Study on the Occurrence of Wheat Blossom Midges Based on Gene Expression Programming with Support Vector Machines

“…GEP has proven its effectiveness in various studies related to plant health prediction, showcasing its ability to handle complex agricultural data and develop models that are generalizable across different crops and regions. GEP's rapid model evolution and updating capabilities make it a promising approach for developing real-time prediction systems and early warning mechanisms [18][19][20][21].…”