Evaluating slope failures in Ali-shan, Taiwan by artificial neural networks

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)


This paper was aimed at creating an empirical model in predicting highway slope failures associated with heavy rainfalls by artificial neural networks. The slope failure database for the model development in the Ali-shan area was established by the data collected from the Directorate General of Highways and field investigations. This database included 955 slope failure records from 1992 to 2001. The main factors extracted from the slope database belonged to three categories: topography, geology, and hydrology. The database was adopted to train and test the intended neural network. The learning and momentum rates used in the neural network are 0.7 and 0.3, respectively. The error rate of the trained neural network was 19.4%, which meant the slope failure model had an accuracy of greater than 80%. On the other hands, multivariate statistical analysis was also performed in analyzing the same slope failure database, and the accuracy is 75.29%. The accuracy comparison of the two models shows that the artificial neural networks performed better than the conventional multivariate statistical analysis. Copyright ASCE 2008.

Original languageEnglish
Title of host publicationProceedings of Sessions of GeoCongress 2008 - GeoCongress 2008
Subtitle of host publicationCharacterization, Monitoring, and Modeling of GeoSystems, GSP 179
Number of pages8
Publication statusPublished - 2008
EventGeoCongress 2008: Characterization, Monitoring, and Modeling of GeoSystems - New Orleans, LA, United States
Duration: 2008 Mar 92008 Mar 12


OtherGeoCongress 2008: Characterization, Monitoring, and Modeling of GeoSystems
Country/TerritoryUnited States
CityNew Orleans, LA

All Science Journal Classification (ASJC) codes

  • Civil and Structural Engineering
  • Architecture
  • Building and Construction
  • Geotechnical Engineering and Engineering Geology


Dive into the research topics of 'Evaluating slope failures in Ali-shan, Taiwan by artificial neural networks'. Together they form a unique fingerprint.

Cite this