• Journal of the Korean Data and Information Science Society
• /
• v.23 no.5
• /
• pp.949-959
• /
• 2012

This article is to find the right size of decision trees that performs better for boosting algorithm. First we defined the tree size D as the depth of a decision tree. Then we compared the performance of boosting algorithm with different tree sizes in the experiment. Although it is an usual practice to set the tree size in boosting algorithm to be small, we figured out that the choice of D has a significant influence on the performance of boosting algorithm. Furthermore, we found out that the tree size D need to be sufficiently large for some dataset. The experiment result shows that there exists an optimal D for each dataset and choosing the right size D is important in improving the performance of boosting. We also tried to find the model for estimating the right size D suitable for boosting algorithm, using variables that can explain the nature of a given dataset. The suggested model reveals that the optimal tree size D for a given dataset can be estimated by the error rate of stump tree, the number of classes, the depth of a single tree, and the gini impurity.

• Journal of Korean Society of Forest Science
• /
• v.95 no.5
• /
• pp.595-600
• /
• 2006

The plot design influences the budgets and the precision of forest inventory results. The objective of this study is to determine the efficiency of estimating forest variables such as tree density, basal area, volume, and species richness based on various plot sizes using fixed-area plot sampling in the natural deciduous forest of Pyeong-Chang County, Gang-won Province, Korea. In this study, 108 reference plots were established with a fixed plot size and shape of 0.09 ha ($30m{\times}30m$). In order to determine the optimal plot size for the interest of variables, each sample plot was established using different shapes (square, circle, and rectangle) and was divided into different plot sizes from 100 to $900m^2$. The mean relative difference (MRD) for the sum of the basal area and volume, and tree density per hectare decreased as plot size increased. But the MRD for three variables were only below 13% at the plot size of $500m^2$. Species richness for each reference stand observed ranging from 2 to 15 species, demonstrated highly positive significant relationships with plot size. The minimum plot size for the estimation of tree density, the sum of the BA and volume was determined to be about $400m^2$, whereas the estimation of species richness required a minimum plot size of $500m^2$.

• Journal of the Korean Data and Information Science Society
• /
• v.27 no.1
• /
• pp.255-264
• /
• 2016

Classification is a predictive modeling for a categorical target variable. Various classification ensemble methods, which predict with better accuracy by combining multiple classifiers, became a powerful machine learning and data mining paradigm. Well-known methodologies of classification ensemble are boosting, bagging and random forest. In this article, we assume that decision trees are used as classifiers in the ensemble. Further, we hypothesized that tree size affects classification accuracy. To study how the tree size in uences accuracy, we performed experiments using twenty-eight data sets. Then we compare the performances of ensemble algorithms; bagging, double-bagging, boosting and random forest, with different tree sizes in the experiment.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.09a
• /
• pp.330-333
• /
• 2003

Initial cluster size for clustering of partitioning methods is very important to the clustering result. In K-means algorithm, the result of cluster analysis becomes different with optimal cluster size K. Usually, the initial cluster size is determined by prior and subjective information. Sometimes this may not be optimal. Now, more objective method is needed to solve this problem. In our research, we propose a hybrid genetic algorithm, a tree induction based evolution algorithm, for determination of optimal cluster size. Initial population of this algorithm is determined by the number of terminal nodes of tree induction. From the initial population based on decision tree, our optimal cluster size is generated. The fitness function of ours is defined an inverse of dissimilarity measure. And the bagging approach is used for saying computational time cost.

• Journal of Forest and Environmental Science
• /
• v.28 no.1
• /
• pp.56-62
• /
• 2012

There is little information about appropriate greenspace structures to satisfy aesthetic function in Korea. The purpose of this study was to analyze Korean's aesthetic preferences for greenspace structures concerned with urban tree plantings of an areal type to explore desirable greenspace landscapes. The study considered 5 structural variables of greenspace which were species composition, tree density, tree size, vertical and horizontal structure, and tree layout pattern. A photo-questionnaire was prepared through color simulations of different landscape types for each structural variable. Preference responses of an interval-scale rating from 214 respondents were statistically analyzed between landscape types and between respondent groups. Respondents preferred greenspace landscapes with diverse tree species to single species, higher tree density to lower density, larger trees to many smaller trees, multilayered and grouped plantings to single-layered and sparse plantings, and informal pattern to formal pattern. These preferences tended to be relatively higher for educated specialist and student groups than for other generalist group. Thus, multilayered and dense plantings in natural pattern including larger trees of diverse species, which are similar to ecological plantings, are recommended to increase aesthetic function of greenspace.

• Korean Journal of Remote Sensing
• /
• v.27 no.6
• /
• pp.703-715
• /
• 2011

Forests have been considered one of the most important ecosystems on the earth, affecting the lives and environment. The sustainable forest management requires accurate and timely information of forest and tree parameters. Appropriately interpreted remotely sensed imagery can provide quantitative data for deriving forest information temporally and spatially. Especially, analysis of individual tree detection and crown delineation is significant issue, because individual trees are basic units for forest management. Individual trees in aerial imagery have reflectance characteristics according to tree species, crown shape and hierarchical status. This study suggested a method that identified individual trees and delineated crown boundaries through adopting gradient method algorithm to amplified greenness data using red and green band of aerial imagery. The amplification of specific band value improved possibility of detecting individual trees, and gradient method algorithm was performed to apply to identify individual tree tops. Additionally, tree crown boundaries were explored using spectral intensity pattern created by geometric characteristic of tree crown shape. Finally, accuracy of result derived from this method was evaluated by comparing with the reference data about individual tree location, number and crown boundary acquired by visual interpretation. The accuracy ($\hat{K}$) of suggested method to identify individual trees was 0.89 and adequate window size for delineating crown boundaries was $19{\times}19$ window size (maximum crown size: 9.4m) with accuracy ($\hat{K}$) at 0.80.

• Journal of Ecology and Environment
• /
• v.29 no.3
• /
• pp.185-189
• /
• 2006

Optimum tapping size and number for sap collection were investigated to minimize damage on tree growth and secure resources for sap by sap collection of Acer mono. The amounts of sap collected by tapping size of 12 mm, 10 mm and 8 mm was 189 L, 140 L and 193 L, respectively. Fusion rates by tapping size were 100% for 8 mm, 89% for 10 mm and 85% for 12 mm, respectively. Amount of sap by size and number of tapping were much at three, two and three tappings to small, middle and large diameter tree in case of 8 mm tapping, and it was large in quantity at two tappings of 12 mm tapping to large diameter tree, respectively. Trees tapped from one to three tappings of 8 mm size for small diameter tree with treatment of DB ointment (mixture of thiophane ointment and 2% bamboo charcoal powder) were completely filled up within 6 months. Diameter growth by number of tapping of 8 mm size was 0.60 mm for one tapping, 1.12 mm for two tappings and 0.47 mm for three tappings to small, middle and large diameter tree, respectively. In case of 12 mm tapping size, diameter growth was fast in the order of large (0.55 mm), middle (0.30 mm) and small (0.23 mm) diameter tree, respectively.

• The KIPS Transactions:PartA
• /
• v.15A no.6
• /
• pp.325-334
• /
• 2008

Flash memory is widely used as a storage medium for mobile devices such as cell phones, MP3 players, PDA's due to its tiny size, low power consumption and shock resistant characteristics. Additionally, some computer manufacturers try to replace hard-disk drives used in Laptops or personal computers with flash memory. More recently, there are some literatures on developing a flash memory-aware $B^+$-tree index for an efficient key-based search in the flash memory storage system. They focus on minimizing the number of "overwrites" resulting from inserting or deleting a sequence of key values to/from the $B^+$-tree. However, in addition to this factor, the size of a physical page allocated to a node can affect the maintenance cost of the $B^+$-tree. In this paper, with diverse experiments, we compare and analyze the costs of construction and search of the $B^+$-tree and the space requirement on flash memory as the node size increases. We also provide sorting-based or non-sorting-based algorithms to be used when inserting a key value into the node and suggest an header structure of the index node for searching a given key inside it efficiently.

• Journal of Ecology and Environment
• /
• v.44 no.4
• /
• pp.229-240
• /
• 2020

Background: Genus Quercus is a successful group that has occupied the largest area of forest around the world including South Korea. The acorns are an important food source for both wild animals and humans. Although the reproductive characteristics of this genus are highly variable, it had been rarely studied in South Korea. Therefore, in Seoraksan and Odaesan National Parks (i) we measured the acorn production of Quercus mongolica, an overwhelmingly dominant species in South Korea, for 3 years (2017-2019), (ii) evaluated the spatial-temporal variation of acorn production, and (iii) analyzed the effects of oak- and site-related variables on the acorn production. Results: The annual acorn production of Q. mongolica increased 36 times from 1.2 g m-2 in 2017 to 43.2 g m-2 in 2018, and decreased to 16.7 g m-2 in 2019, resulting in an annual coefficient of variation of 104%. The coefficient of spatial variation was high and reached a maximum of 142%, and the tree size was the greatest influencing factor. That is, with an increase in tree size, acorn production increased significantly (2018 F = 16.3, p < 0.001; 2019 F = 8.2, p < 0.01). Elevation and slope also significantly affected the production in 2019. However, since elevation and tree size showed a positive correlation (r = 0.517, p < 0.001), the increase in acorn production with increasing elevation was possibly due to the effect of tree size. The acorn production of Odaesan for 3 years was 2.2 times greater than that of Seoraksan. This was presumed that there are more distribution of thick oak trees and more favorable site conditions such as deep soil A-layer depth, high organic matter, and slower slopes. Conclusion: As reported for other species of the genus Quercus, the acorn production of Q. mongolica showed large spatial and annual variations. The temporal variability was presumed to be a weather-influenced masting, while the spatial variability was mainly caused by oak tree size.

• Journal of Plant Biotechnology
• /
• v.3 no.2
• /
• pp.45-57
• /
• 2001

Trees have the ability to undergo secondary growth and produce a woody body. This tree-specific growth is affected by the secondary vascular system and the developmental continuum of secondary phloem and xylem. Secondary growth is one of the most important biological processes on earth. Considering its economic and environmental significance, our knowledge of tree growth and development is surprisingly limited. Trees have received little attention as model species in plant science, as most Plant biology questions can be best addressed by using herbaceous model species, such as Arabidopsis. Furthermore, tree biology is difficult to study mainly due to the inherent problems of tree species, including large size, long generation time, large genome size, and recalcitrance to biotechnological manipulations. Despite all of this, one must rely on trees as models to study tree-specific questions, such as secondary growth, which cannot be studied effectively in non-woody model species. Recent advances in genomics technology provide a unique opportunity to overcome these inherent tree-related problems. Several groups, including our own, have been successful in studying the biology of wood formation with a variety of hardwood and softwood species. In this article, 1 first review the current understanding of tree growth and then discuss the recent attempts to fully explore and realize the potential of molecular biology as a tool for enhanced understanding of secondary growth.