| Title |
Comparison of Naive Bayesian Classifier and Self Organizing Maps Model for Selecting the Retaining Wall Methods |
| Keywords |
Retaining Wall Methods ; Naive Bayesian Classifier ; Self Organizing Maps ; Pattern Recognition |
| Abstract |
In construction planning stage, it’s very important and difficult to choose a proper earth retaining wall method. In general, retaining wall method is selected through the use of limited information by a highly experienced expert. However frequent changes order during construction happen and affect the aspects of construction cost and period. In this study we suggest and compare two classification method ? Naive Bayesian Classifier(NBC) and Self Organizing Maps(SOM) for selecting the retaining wall methods. Comparing the based on NBC with SOM from 209 retaining wall cases, the experiment results show how NBC and SOM approaches can be used for selecting the retaining wall methods. Especially it was found that SOM approaches is more stable than NBC and can provide additional features based on the complete visibility graph. |