Spatial predictive modeling of prehistoric sites in the
Bohemian-Moravian Highlands based on graph similarity analysis

J 2018

Spatial predictive modeling of prehistoric sites in the Bohemian-Moravian Highlands based on graph similarity analysis

MERTEL, Adam, Peter ONDREJKA and Klára ŠABATOVÁ

Basic information

Original name

Spatial predictive modeling of prehistoric sites in the Bohemian-Moravian Highlands based on graph similarity analysis

Authors

MERTEL, Adam (703 Slovakia, guarantor, belonging to the institution), Peter ONDREJKA (703 Slovakia, belonging to the institution) and Klára ŠABATOVÁ (203 Czech Republic, belonging to the institution)

Edition

Open Geosciences, Berlin, De Gruyter Open, 2018, 2391-5447

Other information

Language

English

Type of outcome

Článek v odborném periodiku

Field of Study

60102 Archaeology

Country of publisher

Germany

Confidentiality degree

není předmětem státního či obchodního tajemství

References:

URL

Impact factor

Impact factor: 0.788

RIV identification code

RIV/00216224:14210/18:00103055

Organization unit

Faculty of Arts

DOI

http://dx.doi.org/10.1515/geo-2018-0020

UT WoS

000439357300002

Keywords in English

graph analysis; machine learning; spatialmodeling; archaeological predictive modeling; historical landscape

Abstract

V originále

This paper presented a new method for identifying promising areas for archaeological research. The method is based on graph analysis that iteratively compares and manipulates Hamming distances between graphs of input geographical parameters and graphs of human activity data in various historical periods. The weights learned from the comparison was used to build a prediction model to estimate the potential presence of an archaeological site of a certain time period in a given cadaster. This was applied in the Bohemian Moravian Highlands region based on the most complete archaeological dataset of the area. Resulting maps were analyzed from the archaeological and historical point of view to test against the existing knowledge of prehistoric population movement in the region. Overall, the method proved to overcome problems such as fragmentary inputs and is a good candidate for application in smaller and geographically diverse research areas. The aim of this work was to contribute to the methodology of the prediction of historical human activity, to facilitate greater comprehension of past local settlement dynamics, and to possibly ease the protection of cultural heritage.

Links

MUNI/A/0734/2017, interní kód MU

Name: Archeologické terénní prospekce, exkavace, dokumentace a muzejní prezentace VII

Investor: Masaryk University, Category A

MUNI/A/1419/2016, interní kód MU

Name: Integrovaný výzkum environmentálních změn v krajinné sféře Země II

Investor: Masaryk University, Category A

Citovat

MERTEL, Adam, Peter ONDREJKA and Klára ŠABATOVÁ. Spatial predictive modeling of prehistoric sites in the Bohemian-Moravian Highlands based on graph similarity analysis. Open Geosciences. Berlin: De Gruyter Open, 2018, vol. 10, No 1, p. 261-274. ISSN 2391-5447. Available from: https://dx.doi.org/10.1515/geo-2018-0020.

@article{1421365,
   author = {Mertel, Adam and Ondrejka, Peter and Šabatová, Klára},
   article_location = {Berlin},
   article_number = {1},
   doi = {http://dx.doi.org/10.1515/geo-2018-0020},
   keywords = {graph analysis; machine learning; spatialmodeling; archaeological predictive modeling; historical landscape},
   language = {eng},
   issn = {2391-5447},
   journal = {Open Geosciences},
   title = {Spatial predictive modeling of prehistoric sites in the Bohemian-Moravian Highlands based on graph similarity analysis},
   url = {https://doi.org/10.1515/geo-2018-0020},
   volume = {10},
   year = {2018}
}

TY  - JOUR
ID  - 1421365
AU  - Mertel, Adam - Ondrejka, Peter - Šabatová, Klára
PY  - 2018
TI  - Spatial predictive modeling of prehistoric sites in the Bohemian-Moravian Highlands based on graph similarity analysis
JF  - Open Geosciences
VL  - 10
IS  - 1
SP  - 261-274
EP  - 261-274
PB  - De Gruyter Open
SN  - 23915447
KW  - graph analysis
KW  - machine learning
KW  - spatialmodeling
KW  - archaeological predictive modeling
KW  - historical landscape
UR  - https://doi.org/10.1515/geo-2018-0020
N2  - This paper presented a new method for identifying promising areas for archaeological research. The method is based on graph analysis that iteratively compares and manipulates Hamming distances between graphs of input geographical parameters and graphs of human activity data in various historical periods. The weights learned from the comparison was used to build a prediction model to estimate the potential presence of an archaeological site of a certain time period in a given cadaster. This was applied in the Bohemian Moravian Highlands region based on the most complete archaeological dataset of the area. Resulting maps were analyzed from the archaeological and historical point of view to test against the existing knowledge of prehistoric population movement in the region. Overall, the method proved to overcome problems such as fragmentary inputs and is a good candidate for application in smaller and geographically diverse research areas. The aim of this work was to contribute to the methodology of the prediction of historical human activity, to facilitate greater comprehension of past local settlement dynamics, and to possibly ease the protection of cultural heritage.
ER  -

MERTEL, Adam, Peter ONDREJKA and Klára ŠABATOVÁ. Spatial predictive modeling of prehistoric sites in the Bohemian-Moravian Highlands based on graph similarity analysis. \textit{Open Geosciences}. Berlin: De Gruyter Open, 2018, vol.~10, No~1, p.~261-274. ISSN~2391-5447. Available from: https://dx.doi.org/10.1515/geo-2018-0020.

Detailed Information on Publication Record