DIGITALNA ARHIVA ŠUMARSKOG LISTA
prilagođeno pretraživanje po punom tekstu




ŠUMARSKI LIST 3-4/2017 str. 25     <-- 25 -->        PDF

Development of an ensemble classifier with data from description sheets To classify forest stands in site qualities
RAZVOJ KOMPOZITNOG KLASIFIKATORA S PODACIMA IZ OPISNIH LISTOVA ZA KLASIFIKACIJU BONITETA ŠUMSKIH SASTOJINA
Kyriaki KITIKIDOU, Elias MILIOS, Panagiota PALAVOUZI
ABSTRACT
Aim of study: In this work, we tested the technique of combining the predictions of classifiers for the development of a single, ensemble classifier, in order to classify forest stands in site qualities. Area of study: We used data of the forest stands of Dadia – Lefkimi – Soufli forest (north-eastern Greece). Materials and methods: The variables that we used as input were the altitude, slope, age and canopy density. For the ensemble classifier development we applied the boosting algorithm. Main results: The canopy density was the most important predictor; topography which replaced altitudes and slopes was the second important predictor, while the developed ensemble classifier gave a percentage of correct classification up to 98.59% (for the worst site quality). Research essentials: If we consider that the initial site classification comprised over 70% of the Dadia-Lefkimi –Soufli forest area in the worst site quality, then the usage of boosting method for creating a collective classifier for site qualities in the studied forest can be characterized as fully successful. The application of this method using these input parameters do not need background information regarding the tree age and (or) other difficult to access information. Moreover, in a quite high degree, this site classification is not influenced by disturbances. The boosting method for creating a collective classifier for site qualities obviously will give far more accurate classifications of site productivity, if a more sophisticated scheme of data collection is used.
KEY WORDS: ensemble classifiers; forest stands; site qualities.
Introduction
UVOD
The technique of combining the predictions of many classifiers, for the creation of a single, collective classifier (ensemble classifier), has preoccupied researchers (Breiman 1996, Clemen 1989, Perrone 1993, Wolpert 1992, Opitz and Shavlik 1996). A collective classifier is effective when it is more accurate than any classifier that participates in the