Estimation of standard penetration test value on cohesive soil using artificial neural network without data normalization

Artificial neural networks (ANNs) are often used recently by researchers to solve complex and nonlinear problems. Standard penetration test (SPT) cone (CPT) field tests that obtain soil parameters. There have been many previous studies examined the value obtained through SPT with CPT test, but research carried out still uses equations linear. This will conduct an estimated of on cohesive using data in form end resistance blanket resistance, laboratory such as effective overburden pressure, liquid limit, plastic limit percentage sand, silt clay. study 242 testing areas several cities island Sumatra, Indonesia. The developed artificial network be created without normalization. final results this root mean square error (RMSE) values 3.441, absolute (MAE) 2.318 R2 0.9451 for training RMSE 2.785, MAE 2.085, 0.9792 data. RMSE, indicate ANN has is considered quite good efficient estimating value.