Search

Opisuje se predloženi alternatnivni opći regresijski model neuronske mreže, koji omogućuje analizu sumarnih podataka dobivenih istraživanjem likvefakcije uobičajenim metodama. U tu su svrhu u model uneseni podaci s ukupno trideset bušotina izvedenih terenskim istraživanjima lociranim u Babolu, u iranskoj pokrajini Mazandaran. Rezultati se slažu s rezultatima dobivenim metodom koju su uveli Seet i drugi. Predloženi je model pouzdan za ocjenjivanje likvefakcijskog potencijala., The authors propose an alternative general regression model based on neural networks, which enables analysis of summary data obtained by liquefaction analysis according to usual methods. For that purpose, the data from some thirty boreholes made during field investigations in Babol, in the Iranian province of Mazandaran, were entered in the model. The results correspond well with the results obtained by the method introduced by Seet and others. The proposed model has been found reliable for prediction of liquefaction potential.