Statistics for Spatial Data.

1. Introduction and Overview. 2. Detecting Influential Observations and Outliers. 3. Detecting and Assessing Collinearity. 4. Applications and Remedies. 5. Research Issues and Directions for Extensions. Bibliography. Author Index. Subject Index.

Regression Diagnostics: Identifying Influential Data and Sources of Collinearity

Climate change 2014 - Synthesis Report

We evaluate 179 classifiers arising from 17 families (discriminant analysis, Bayesian, neural networks, support vector machines, decision trees, rule-based classifiers, boosting, bagging, stacking, random forests and other ensembles, generalized linear models, nearest-neighbors, partial least squares and principal component regression, logistic and multinomial regression, multiple adaptive regression splines and other methods), implemented in Weka, R (with and without the caret package), C and Matlab, including all the relevant classifiers available today. We use 121 data sets, which represent the whole UCI data base (excluding the large-scale problems) and other own real problems, in order to achieve significant conclusions about the classifier behavior, not dependent on the data set collection. The classifiers most likely to be the bests are the random forest (RF) versions, the best of which (implemented in R and accessed via caret) achieves 94.1% of the maximum accuracy overcoming 90% in the 84.3% of the data sets. However, the difference is not statistically significant with the second best, the SVM with Gaussian kernel implemented in C using LibSVM, which achieves 92.3% of the maximum accuracy. A few models are clearly better than the remaining ones: random forest, SVM with Gaussian and polynomial kernels, extreme learning machine with Gaussian kernel, C5.0 and avNNet (a committee of multi-layer perceptrons implemented in R with the caret package). The random forest is clearly the best family of classifiers (3 out of 5 bests classifiers are RF), followed by SVM (4 classifiers in the top-10), neural networks and boosting ensembles (5 and 3 members in the top-20, respectively).

/pdf/do-we-need-hundreds-of-classifiers-to-solve-real-world-1igtahh9y9.pdf

Do we need hundreds of classifiers to solve real world classification problems

A wide range of methods for analysis of airborne- and satellite-derived imagery continues to be proposed and assessed. In this paper, we review remote sensing implementations of support vector machines (SVMs), a promising machine learning methodology. This review is timely due to the exponentially increasing number of works published in recent years. SVMs are particularly appealing in the remote sensing field due to their ability to generalize well even with limited training samples, a common limitation for remote sensing applications. However, they also suffer from parameter assignment issues that can significantly affect obtained results. A summary of empirical results is provided for various applications of over one hundred published works (as of April, 2010). It is our hope that this survey will provide guidelines for future applications of SVMs and possible areas of algorithm enhancement.

https://www.aboutgis.com/Publications/Mountrakis_SVM_review_in_remote_sensing_ISPRS2010.pdf

Support vector machines in remote sensing: A review

(1) Department of Geodesy and Geoinformation, Vienna University of Technology, Vienna, Austria (ana.djuricic@geo.tuwien.ac.at), (2) Geological Paleontological Department, Natural History Museum Vienna, Vienna, Austria (mathias.harzhauser@nhm-wien.ac.at), (3) D-IT GmbH, Austria (office@4d-it.com), (4) Interdisziplinäres Ökologisches Zentrum, TU Bergakademie Freiberg, Germany (balazs.szekely@ioez.tu-freiberg.de), (5) Department of Geophysics and Space Science, Eötvös Loránd University, Hungary (molnar@sas.elte.hu)

Parameter Estimation of Fossil Oysters from High Resolution 3D Point Cloud and Image Data

We present a set of methods to improve the automation of the parametric 3D modeling of historic roof structures using terrestrial laser scanning (TLS) point clouds. The final product of the TLS point clouds consist of 3D representation of all objects, which were visible during the scanning, including structural elements, wooden walking ways and rails, roof cover and the ground; thus, a new method was applied to detect and exclude the roof cover points. On the interior roof points, a region-growing segmentation-based beam side face searching approach was extended with an additional method that splits complex segments into linear sub-segments. The presented workflow was conducted on an entire historic roof structure. The main target is to increase the automation of the modeling in the context of completeness. The number of manually counted beams served as reference to define a completeness ratio for results of automatically modeling beams. The analysis shows that this approach could increase the quantitative completeness of the full automatically generated 3D model of the roof structure from 29% to 63%.

/pdf/historic-timber-roof-structure-reconstruction-through-3puwoull.pdf

Historic Timber Roof Structure Reconstruction through Automated Analysis of Point Clouds

A proof of concept: Airborne LIDAR-measured ellipsoidal heights of a lake surface correspond to a local geoid model

High-resolution Digital Terrain Models (DTMs) representing the bare Earth are a fundamental input for various applications in geomorphology. Airborne laser scanning (ALS) is established as a standard tool for deriving DTMs over large areas with unprecedented accuracy. Due to advances in sensor technology and in processing algorithms in the recent years the obtainable accuracy is still increasing. Accuracy is understood as the deviation from the elevation at one specified point to its true value. These advances may lead to a more efficient data acquisition, if reduced accuracy is targeted, but also allow data acquisition schemes with more detail becoming visible, i.e. small features of the relief. For the latter a high internal precision, i.e. repeatability, is necessary. The essential advances in the technologies are improvements in ranging through the introduction of full-waveform (FWF) laser scanning and rigorous models of strip adjustment.

/pdf/state-of-the-art-in-high-accuracy-high-detail-dtms-derived-4f4r7vkf5b.pdf

Norbert Pfeifer

Papers

Parameter Estimation of Fossil Oysters from High Resolution 3D Point Cloud and Image Data

Historic Timber Roof Structure Reconstruction through Automated Analysis of Point Clouds

A proof of concept: Airborne LIDAR-measured ellipsoidal heights of a lake surface correspond to a local geoid model

State of the art in high accuracy high detail DTMs derived from ALS

The potential of UAV laser scanning in forest inventories