TY - JOUR ID - 158235 TI - Soft Computing Based Prediction of Unconfined Compressive Strength of Fly Ash Stabilised Organic Clay JO - Journal of Soft Computing in Civil Engineering JA - SCCE LA - en SN - AU - Gnananandarao, Tammineni AU - Dutta, Rakesh Kumar AU - Khatri, Vishwas Nandakishor AU - Kumar, Mummidivarapu Satish AD - Assistant Professor, Department of Civil Engineering, Aditya College of Engineering and Technology, Surampalem, Andhra Pradesh, India AD - Professor, National Institute of Technology, Hamirpur, Himachal Pradesh, Pin No: 177005, India AD - Assistant Professor, Indian Institute of Technology (Indian School of Mines), Dhanbad, Jharkhand, Pin No: 826004, India Y1 - 2022 PY - 2022 VL - 6 IS - 4 SP - 43 EP - 58 KW - Unconfined compressive strength KW - Organic clay KW - Fly ash KW - Random forest regression KW - Artificial Neural Networks KW - SVM Poly kernel KW - SVMRBK kernel KW - M5P model tree DO - 10.22115/scce.2022.339698.1429 N2 - The current study uses machine learning techniques such as Random Forest Regression (RFR), Artificial Neural Networks (ANN), Support Vector Machines Ploy kernel (SVMP), Support Vector Machines Radial Basis Function Kernel (SVMRBK), and M5P model tree (M5P) to estimate unconfined compressive strength of organic clay stabilized with fly ash. The unconfined compressive strength of stabilized clay was computed by considering the different input variables namely i) the ratio of Cao to Sio2, ii) organic content (OC), iii) fly ash (FAper) content, iv) the unconfined compressive strength of organic clay without fly ash (UCS0) and v) the pH of soil-fly ash (pHmix). By comparing the performance measure parameters, each model performance is evaluated. The result of present study can conclude the random forest regression (RFR) model predicts the unconfined compressive strength of the organic clay stabilized with fly ash with least error followed by Support Vector Machines Radial Basis Function Kernel (SVMRBK), Support Vector Machines Ploy kernel (SVMP), Artificial Neural Networks (ANN) and M5P model tree (M5P). When compared to the semi-empirical model available in the literature, all of the model predictions given in this study perform well. Finally, the RFR and SVMRBK sensitivity analyses revealed that the CaO/SiO2 ratio was the most relevant parameter in the prediction of unconfined compressive strength. UR - https://www.jsoftcivil.com/article_158235.html L1 - https://www.jsoftcivil.com/article_158235_ac5ba3299fe4e2c30b6196e191edfe6b.pdf ER -