Showing papers by "Kai Puolamäki published in 2010"

Precipitation and large herbivorous mammals II: application to fossil data

Abstract: Background: We developed a method to estimate precipitation using mammalian ecomorphology, specifically the relative height of the molars of herbivores (see companion paper, this issue) Question: If we apply the new method to paleoenvironments, do the results agree with previous results from fossil mammals and paleobotanical proxies? Data: Large herbivorous fossil mammals of Eurasia Data from NOW database covers 23–22 Ma and is Eurasia-wide Method: We apply the new precipitation estimation method (based on present-day mammalian ecomorphology) to fossil assemblages from different localities Conclusions: The early Miocene retained the overall humid conditions of the late Paleogene A shift to more arid conditions began during the middle Miocene The late Miocene as a whole was a time of large changes, and there was continent-wide restructuring of the distribution of environments Our new results agree with previous investigations and the mammal proxy data are in good agreement with palaeovegetation data Mammals and vegetation produce similar precipitation values and large-scale patterns

Precipitation and large herbivorous mammals I: estimates from present-day communities

Abstract: Question: How can mammalian community characteristics be used to estimate regional precipitation? Data: Global distribution data of large mammals and their ecomorphology; global climate data. Research methods: Non-linear regression-tree analysis and linear regression. Conclusions: The methods unravelled the complex relationships between the environment and the characteristics of mammalian communities. The regression trees described here provide a reasonably accurate estimate of precipitation values for today’s world. The strongest correlations are for annual precipitation versus diet (R 2 = 0.665), precipitation versus tooth crown height (R 2 = 0.658), and precipitation versus diet and tooth crown height combined (R 2 = 0.742)

Contextual information access with Augmented Reality

Abstract: We have developed a prototype platform for contextual information access in mobile settings. Objects, people, and the environment are considered as contextual channels or cues to more information. The system infers, based on gaze, speech and other implicit feedback signals, which of the contextual cues are relevant, retrieves more information relevant to the cues, and presents the information with Augmented Reality (AR) techniques on a handheld or head-mounted display. The augmented information becomes potential contextual cues as well, and its relevance is assessed to provide more information. In essence, the platform turns the real world into an information browser which focuses proactively on the information inferred to be the most relevant for the user. We present the first pilot application, a Virtual Laboratory Guide, and its early evaluation results.

Visually Controllable Data Mining Methods

Abstract: A large number of data mining methods are, as such, not applicable to fast, intuitive, and interactive use. Thus, there is a need for visually controllable data mining methods. Such methods should comply with three major requirements: their model structure can be represented visually, they can be controlled using visual, interaction, and they should be fast enough for visual interaction. We, define a framework for using data mining methods in, interactive visualization. These data mining methods are called, ``visually controllable'' and combine data mining with visualization, and user-interaction, bridging the gap between data mining and, visual analytics. Our main objective is to define the interactive, visualization scenario and the requirements for visually, controllable data mining. Basic data mining algorithms are reviewed, and it is demonstrated how they can be controlled visually. We also discuss how existing visual analytics tools fit to the proposed framework., From a data mining perspective, this work creates a reference framework for, designing and evaluating visually controllable algorithms and visual analytics systems.

Introduction to the special issue on visual analytics and knowledge discovery

Abstract: The papers in this Special Issue present the state of the art in Visual Analytics and Knowledge Discovery and Data Mining (KDD), as well as propose potential extensions and research questions to further advance and integrate these two fields.

The smallest set of constraints that explains the data: a randomization approach

Abstract: Randomization methods can be used to assess statistical significance of data mining results. A randomization method typically consists of a sampler which draws data sets from a null distribution, and a test statistic. If the value of the test statistic on the original data set is more extreme than the test statistic on randomized data sets we can reject the null hypothesis. It is often not immediately clear why the null hypothesis is rejected. For example, the cost of clustering can be significantly lower in the original data than in the randomized data, but usually we would also like to know why the cost is small. We introduce a methodology for finding the smallest possible set of constraints, or patterns, that explains the data. In principle any type of patterns can be used as long as there exists an appropriate randomization method. We show that the problem is, in its general form, NP-hard, but that in a special case an exact solution can be computed fast, and propose a greedy algorithm that solves the problem. The proposed approach is demonstrated on time series data as well as on frequent itemsets in 0–1 matrices, and validated theoretically and experimentally.

Canonical analysis of individual vocabulary profiling data

Abstract: Canonical analysis is an ordination technique which allows a direct comparison of two data matrices. It is often applied in ecological studies and here it is shown how canonical analysis, in particular redundancy analysis, can be used to analyze sensory evaluation data. The example data are obtained with individual vocabulary profiling study of concert hall acoustics. The redundancy analysis is compared with hierarchical multiple factor analysis. In addition, it is shown how redundancy analysis can be used to explain the subjective data with different objective data, in this case room acoustical parameters and physical measures of the studied concert halls.