Investigation of a possibility of spatial modelling of tree diversity using environmental and data mining algorithms
Biological diversity is the basis for a wide array of goods and services provided by forests. The variety of forest trees and shrubs plays a vital role in the daily life of forest communities. The purpose of this study is to investigate the possibility of modelling the diversity of tree species by c...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Czech Academy of Agricultural Sciences
2016-12-01
|
| Series: | Journal of Forest Science |
| Subjects: | |
| Online Access: | https://jfs.agriculturejournals.cz/artkey/jfs-201612-0005_investigation-of-a-possibility-of-spatial-modelling-of-tree-diversity-using-environmental-and-data-mining-algor.php |
| _version_ | 1797898750417436672 |
|---|---|
| author | A. Abdollahnejad D. Panagiotidis P. Surový |
| author_facet | A. Abdollahnejad D. Panagiotidis P. Surový |
| author_sort | A. Abdollahnejad |
| collection | DOAJ |
| description | Biological diversity is the basis for a wide array of goods and services provided by forests. The variety of forest trees and shrubs plays a vital role in the daily life of forest communities. The purpose of this study is to investigate the possibility of modelling the diversity of tree species by characteristics of topography, soil and climate, using data mining algorithms k-NN, RF and SVM in Dr. Bahramnia forestry plan in the north of Iran. Based on the basal area factor for each species in a total of 518 sample plots, diversity indices such as species richness, evenness and heterogeneity were calculated for each plot. Topographic maps of primary and secondary properties were prepared using the digital elevation model. Categories of the soil and climate maps database of Dr. Bahramnia forestry plan were extracted. Modelling rates of tree and shrub species diversity using data mining algorithms, with 80% of the sampling plots were taken. Assessment of the model accuracy, using 20% of samples and evaluation criteria, was conducted. Results showed that topographic features, especially elevation, had the highest impact on the species diversity index. The modelling results also showed that Camargo evenness index had lowest root mean square error (RMSE) (0.14) and RMSE% (24.35), compared to other indicators of diversity. In addition, the results of the comparison between the algorithms showed that the random forest algorithms were more accurate in modelling the diversity. |
| first_indexed | 2024-04-10T08:18:01Z |
| format | Article |
| id | doaj.art-0238147d300443fe807ea5868361bb7a |
| institution | Directory Open Access Journal |
| issn | 1212-4834 1805-935X |
| language | English |
| last_indexed | 2024-04-10T08:18:01Z |
| publishDate | 2016-12-01 |
| publisher | Czech Academy of Agricultural Sciences |
| record_format | Article |
| series | Journal of Forest Science |
| spelling | doaj.art-0238147d300443fe807ea5868361bb7a2023-02-23T03:42:45ZengCzech Academy of Agricultural SciencesJournal of Forest Science1212-48341805-935X2016-12-01621256257010.17221/97/2016-JFSjfs-201612-0005Investigation of a possibility of spatial modelling of tree diversity using environmental and data mining algorithmsA. Abdollahnejad0D. Panagiotidis1P. Surový2Department of Forest Management, Faculty of Forestry and Wood Sciences, Czech University of Life Sciences Prague, Prague, Czech RepublicDepartment of Forest Management, Faculty of Forestry and Wood Sciences, Czech University of Life Sciences Prague, Prague, Czech RepublicDepartment of Forest Management, Faculty of Forestry and Wood Sciences, Czech University of Life Sciences Prague, Prague, Czech RepublicBiological diversity is the basis for a wide array of goods and services provided by forests. The variety of forest trees and shrubs plays a vital role in the daily life of forest communities. The purpose of this study is to investigate the possibility of modelling the diversity of tree species by characteristics of topography, soil and climate, using data mining algorithms k-NN, RF and SVM in Dr. Bahramnia forestry plan in the north of Iran. Based on the basal area factor for each species in a total of 518 sample plots, diversity indices such as species richness, evenness and heterogeneity were calculated for each plot. Topographic maps of primary and secondary properties were prepared using the digital elevation model. Categories of the soil and climate maps database of Dr. Bahramnia forestry plan were extracted. Modelling rates of tree and shrub species diversity using data mining algorithms, with 80% of the sampling plots were taken. Assessment of the model accuracy, using 20% of samples and evaluation criteria, was conducted. Results showed that topographic features, especially elevation, had the highest impact on the species diversity index. The modelling results also showed that Camargo evenness index had lowest root mean square error (RMSE) (0.14) and RMSE% (24.35), compared to other indicators of diversity. In addition, the results of the comparison between the algorithms showed that the random forest algorithms were more accurate in modelling the diversity.https://jfs.agriculturejournals.cz/artkey/jfs-201612-0005_investigation-of-a-possibility-of-spatial-modelling-of-tree-diversity-using-environmental-and-data-mining-algor.phptopographic characteristicssuborder soilclimatenon-parametric algorithmsrichnessevenness indicators |
| spellingShingle | A. Abdollahnejad D. Panagiotidis P. Surový Investigation of a possibility of spatial modelling of tree diversity using environmental and data mining algorithms Journal of Forest Science topographic characteristics suborder soil climate non-parametric algorithms richness evenness indicators |
| title | Investigation of a possibility of spatial modelling of tree diversity using environmental and data mining algorithms |
| title_full | Investigation of a possibility of spatial modelling of tree diversity using environmental and data mining algorithms |
| title_fullStr | Investigation of a possibility of spatial modelling of tree diversity using environmental and data mining algorithms |
| title_full_unstemmed | Investigation of a possibility of spatial modelling of tree diversity using environmental and data mining algorithms |
| title_short | Investigation of a possibility of spatial modelling of tree diversity using environmental and data mining algorithms |
| title_sort | investigation of a possibility of spatial modelling of tree diversity using environmental and data mining algorithms |
| topic | topographic characteristics suborder soil climate non-parametric algorithms richness evenness indicators |
| url | https://jfs.agriculturejournals.cz/artkey/jfs-201612-0005_investigation-of-a-possibility-of-spatial-modelling-of-tree-diversity-using-environmental-and-data-mining-algor.php |
| work_keys_str_mv | AT aabdollahnejad investigationofapossibilityofspatialmodellingoftreediversityusingenvironmentalanddataminingalgorithms AT dpanagiotidis investigationofapossibilityofspatialmodellingoftreediversityusingenvironmentalanddataminingalgorithms AT psurovy investigationofapossibilityofspatialmodellingoftreediversityusingenvironmentalanddataminingalgorithms |