Prediction of Soil Pore Water Pressure Using Polynomial Kernel Support Vector Machine

Knowledge in soil pore water pressure is important in slope stability analysis. The
fluctuation of soil pore water pressure is mainly affect by the rainfall intensity. The
pore water pressure can cause effect on the soil strength. A few problems has been
identified in obtaining the pore water pressure reading which are time consuming and
intensive labor work that are from field instrumentation. However, the development of
soft-computing in nowadays has become focus as alternative technology to monitor
the changes of pore water pressure in soil, and scope in this study is to use support
vector machine. Then, a study on developing a model to predict the soil pore water
using polynomial kernel support vector machine and to evaluate the model
performance has been conducted. A considerable good correlation between observed
and predicted pore water pressure has been found by the end of the studies with
performance evaluation of R2 and RMSE of 0.93 and 0.60 respectively A slope in
Universiti Teknologi PETRONAS, Perak has been chosen as study scope to predict
the soil pore water pressure.