Zoning the Landslides of Givichay River Basin by Using Multi Layer Perceptron Model
Landslides represent morphodynamic processes which occurs on the sloped lands, and may damage residential , industrial, gardens and croplands. In this investigation multi-layer perceptron model of back propagation (BP) was used in landslide zoning in Givichay river basin. To evaluate the created ne...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | fas |
| Published: |
University of Sistan and Baluchestan
2014-09-01
|
| Series: | جغرافیا و توسعه |
| Subjects: | |
| Online Access: | https://gdij.usb.ac.ir/article_1716_a6b38b44909aa4cf52b28cb5e3c63498.pdf |
| _version_ | 1827926920482783232 |
|---|---|
| author | Masoomeh Rajabi Mehdi Feyzolahpour |
| author_facet | Masoomeh Rajabi Mehdi Feyzolahpour |
| author_sort | Masoomeh Rajabi |
| collection | DOAJ |
| description | Landslides represent morphodynamic processes which occurs on the sloped lands, and may damage residential , industrial, gardens and croplands. In this investigation multi-layer perceptron model of back propagation (BP) was used in landslide zoning in Givichay river basin. To evaluate the created neural network, the dataset related to 41 landslide events were presented. Then for processing landslide data in MATLAB software, 8 Layers consisting slope, slope direction, DEM, lithology, distance from the fault, hydrographic network, land use and landslide distribution using field studies, topographic , geological maps and satellite images was prepared in ArcGIS software. These layers were scaled in the range between 1 and 0 based on the largest value for each normalized layer. Then the normalized data was fed to three-layer Perceptron (feed forward) with the back error propagation algorithm. These data was first trained and then tested in the network. The final structure of the network has 8 neurons in the input layer, 20 neurons in the hidden layer and 1 neuron in the output layer. 80 percent of data was considered for training and 20 percent for testing. Finally, with regard to the output weights, landslide zoning maps in five classes: very high, high, medium, low and very low risk were drawn. The obtained results showed that the formed geological structure by Certaceous lime and Andesit Porphyry and also access to high humid resources has made the eastern heights of Boogharodagh and Aladagh in the area of Talesh mountains to have a high potential in the occurrences of landslides |
| first_indexed | 2024-03-13T05:44:37Z |
| format | Article |
| id | doaj.art-3a0e97fcfaed452ea56e87642ffdf0de |
| institution | Directory Open Access Journal |
| issn | 1735-0735 2676-7791 |
| language | fas |
| last_indexed | 2024-03-13T05:44:37Z |
| publishDate | 2014-09-01 |
| publisher | University of Sistan and Baluchestan |
| record_format | Article |
| series | جغرافیا و توسعه |
| spelling | doaj.art-3a0e97fcfaed452ea56e87642ffdf0de2023-06-13T20:17:44ZfasUniversity of Sistan and Baluchestanجغرافیا و توسعه1735-07352676-77912014-09-01123616118010.22111/gdij.2014.17161716Zoning the Landslides of Givichay River Basin by Using Multi Layer Perceptron ModelMasoomeh RajabiMehdi FeyzolahpourLandslides represent morphodynamic processes which occurs on the sloped lands, and may damage residential , industrial, gardens and croplands. In this investigation multi-layer perceptron model of back propagation (BP) was used in landslide zoning in Givichay river basin. To evaluate the created neural network, the dataset related to 41 landslide events were presented. Then for processing landslide data in MATLAB software, 8 Layers consisting slope, slope direction, DEM, lithology, distance from the fault, hydrographic network, land use and landslide distribution using field studies, topographic , geological maps and satellite images was prepared in ArcGIS software. These layers were scaled in the range between 1 and 0 based on the largest value for each normalized layer. Then the normalized data was fed to three-layer Perceptron (feed forward) with the back error propagation algorithm. These data was first trained and then tested in the network. The final structure of the network has 8 neurons in the input layer, 20 neurons in the hidden layer and 1 neuron in the output layer. 80 percent of data was considered for training and 20 percent for testing. Finally, with regard to the output weights, landslide zoning maps in five classes: very high, high, medium, low and very low risk were drawn. The obtained results showed that the formed geological structure by Certaceous lime and Andesit Porphyry and also access to high humid resources has made the eastern heights of Boogharodagh and Aladagh in the area of Talesh mountains to have a high potential in the occurrences of landslideshttps://gdij.usb.ac.ir/article_1716_a6b38b44909aa4cf52b28cb5e3c63498.pdflandslideartificial neural networkmulti-layer perceptronzoninggivichay river basin |
| spellingShingle | Masoomeh Rajabi Mehdi Feyzolahpour Zoning the Landslides of Givichay River Basin by Using Multi Layer Perceptron Model جغرافیا و توسعه landslide artificial neural network multi-layer perceptron zoning givichay river basin |
| title | Zoning the Landslides of Givichay River Basin by Using Multi Layer
Perceptron Model |
| title_full | Zoning the Landslides of Givichay River Basin by Using Multi Layer
Perceptron Model |
| title_fullStr | Zoning the Landslides of Givichay River Basin by Using Multi Layer
Perceptron Model |
| title_full_unstemmed | Zoning the Landslides of Givichay River Basin by Using Multi Layer
Perceptron Model |
| title_short | Zoning the Landslides of Givichay River Basin by Using Multi Layer
Perceptron Model |
| title_sort | zoning the landslides of givichay river basin by using multi layer perceptron model |
| topic | landslide artificial neural network multi-layer perceptron zoning givichay river basin |
| url | https://gdij.usb.ac.ir/article_1716_a6b38b44909aa4cf52b28cb5e3c63498.pdf |
| work_keys_str_mv | AT masoomehrajabi zoningthelandslidesofgivichayriverbasinbyusingmultilayerperceptronmodel AT mehdifeyzolahpour zoningthelandslidesofgivichayriverbasinbyusingmultilayerperceptronmodel |