Narrow Results

Elemental concentrations of light and medium weight elements ranging from about 0.2% to less than 1ppm were determined by thick target PIXE in a variety of South African woods from different localities. Correspondence analysis, using the measured contents of K, Ca, Mn, Fe, Cu, Zn, and Rb, resulted in groupings which were type-dependent and locality-dependent. The possibility of substituting one variety for another in the same group, for paper production, is discussed.

This paper examines the evolution of inequality, poverty and welfare in the South Caucasian Asian states, using an aggregate index based on consumer durables available to the households. Instead of using principal components analysis to aggregate asset indicators into an overall asset index, we propose an ordinal approach to using data on assets, when estimating the wealth of a household (or individual). Using three different approaches (the concept of order of acquisition of durable goods, item response theory and correspondence analysis), we first derive the order of importance of the different assets. Then, we compute indices introduced recently to measure inequality, poverty and achievement, when only ordinal variables are available. Our empirical analysis, based on data collected by the Caucasus Barometer and covering the three states in South Caucasus shows that there exists such an order, that it does not really depend on the statistical approach adopted and that it was very similar in 2009 and 2013. Our results show that among the three countries analyzed, Armenia had the lowest degree of inequality in asset ownership and Georgia the highest. Whatever the approach or index used, ordinal inequality, increased in Armenia and Azerbaijan between 2009 and 2013, but slightly decreased during that period in Georgia. It also appears that poverty decreased in all three countries during this period. There was also important growth, when measuring the latter via the welfare index recently introduced by Apouey et al. (2019), assuming their parameter $\alpha \to 1$.

In order to extract and to visualize qualitative information from a high-dimensional time series, we apply ideas from symbolic dynamics. Counting certain ordinal patterns in the given series, we obtain a series of matrices whose entries are symbol frequencies. This matrix series is explored by simple methods from nominal statistics and information theory. The method is applied to detect and visualize qualitative changes of EEG data related to epileptic activity.

In this note we describe a simple method for visualizing time-dependent similarities and dissimilarities between the components of a high-dimensional time series. On the base of symbolic dynamics, the time series is turned into a series of matrices whose rows quantify pattern types in the components of the original series. For different scales we introduce distances between the components via the obtained pattern type distributions and approximate them in a one-dimensional manner. The method is illustrated for 19-channel EEG data.

Correspondence Analysis (CA), a multivariate statistical technique, allows a visual representation of the association between categorical variables through a contingency table consisting of frequencies representing the existence of relationships. Despite being a widely used statistical technique, the classical CA is not able to demonstrate the uncertainty in real-life problems. To address this issue, a new Interval-valued Hesitant Fuzzy CA approach is proposed to represent the uncertainty caused by human doubt. Due to the nature of operations defined on Hesitant Fuzzy Sets, it is hard to integrate the fuzzy calculations directly into the classical CA. Thus, a new hesitant expected value method is proposed to reveal the independence between two categorical variables. As the output of the proposed approach, an interval-valued hesitant fuzzy correspondence map consisting of rectangles of different sizes representing the amount of the hesitancy is constructed. The applicability of the proposed approach is demonstrated by a simple but effective illustrative example.

It is not always easy, not only for beginners but also for experts, to find out newly research subjects or grasp analytically research trends (or directions). This work is one of the intelligent tasks. Traditionally, some investigations or research projects focused on the challengeable task of human-oriented creativeness: The efforts are very stable and the results make the basic framework or fundamental ideas clear one by one for strictly-constrained application of human-activity. This task is closely dependent on the global framework of knowledge management with respect to the support from human creative activity. In this paper, we discuss a computer-support method to arrange already-read papers analytically and then discuss the research topics related to currently focused research fields or research trends derived from the existing papers and researcher's interests. Our idea is to realise this paper inquiry method in three procedural steps, such as arranging paper contents, grasping relationships among papers and investigating research subjects followed by already-referred papers, and also make users interact with individually-generated results by controlling spirally three steps.

We begin with pervasive ultrametricity due to high dimensionality and/or spatial sparsity. How extent or degree of ultrametricity can be quantified leads us to the discussion of varied practical cases when ultrametricity can be partially or locally present in data. We show how the ultrametricity can be assessed in time series signals. An aspect of importance here is that to draw benefit from this perspective the data may need to be recoded. Such data recoding can also be powerful in proximity searching, as we will show, where the data is embedded globally and not locally in an ultrametric space.

Real-world economic and business applications are in general demanding information processing tasks, whereby high volumes of incomplete, noisy data and high information costs are often the most remarkable features. Hence the need for advanced information analysis methods, such as those developed in the pattern recognition literature. This chapter shows real-world applications of a number of methods, most of them still quite new to the economic and financial analysts. The focus is not on the full coverage of all relevant information analysis approaches (a challenging task indeed), rather on the understanding of how the underlying data structures and data representations influence the interpretation activity. To this purpose, the presentation is organized by application domains, where we address broadly different areas including as examples real-time trading, financial analysis and long-term economic analysis, as well as the related data and interpretation goals. Correspondingly, several different methods used in these application domains are described and shown at work, including: statistics, grammars, neural networks, and qualitative modeling.