Search

Fractal theory has been widely applied in a variety of disciplines to understand the theory behind chaotic phenomena based on internal self-similarity. In this study, three ideal geological models are used to analyze the unreliability of the capacity dimension in the fractal calculation of geological bodies with different scales. Additionally, by varying the side length r of the statistical units, the geological meanings of the fractal dimension D and the correlation coefficient R2 are discussed. The points of information (POIs) are densely filled by binarizing the geological bodies to black/white. Based on the optimized r of a geological body, an algorithm is derived that divides the grids of the statistical units to determine the probability of the POIs falling into different grids. The information dimension (DI) and R2 of a geological body are obtained by fitting the variable data. An example calculation of the information dimension field in the Jinhu sag is presented to demonstrate the methodology and to test its reliability. The results show that determining the appropriate side length of the statistical unit is key to evaluating the fractal calculation. Compared to the capacity dimension, DI is more reliable in the fractal calculation of multi-scale geological bodies; DI is thereby the preferred fractal dimension to use in the analyses of these types of geological bodies.

In these years, Official Statistics has acknowledged the value of Big Data and has started exploring the use of diverse sources in several domains. For some of the sources, the object can be, a part from privacy restrictions, easily related to a statistical unit. In those cases, if a unit identifier was available, the opportunity to link Big data to already existing statistical data at micro-level could allow to enlarge the content, the coverage, the accuracy and the timeliness of official statistics. This could be the case, for instance, of the Internet-scraped data. In this setting, new challenges arise for data integration experts in official statistics, due to the deep differences with respect to the familiar framework in which administrative data have been integrated for a long time in order to produce statistical outputs, i.e. business and population registers. In this study, exploiting a real case as a starting point, we describe novelties and challenges in integrating Internet-scraped data with traditional statistical datasets, from the entity extraction and recognition phases to the unit matching algorithms, as well. As case study, we propose the linkage of Internet-scraped information with data related to agritourisms, as reported in the Italian Farm Register, that is obtained by the integration of seven administrative sources. In order to overcome limits and rigidities of the so far well-established linkage procedures, we explore new techniques not yet introduced in the official statistics production system. Finally, we devote the due attention to the output quality evaluation, in order to entirely understand benefits and risks of the integration and to allow the analysts to take into account potential integration errors in subsequent analyses.

Research on temporal characteristics of human dynamics has attracted much attentions for its contribution to various areas such as communication, medical treatment, finance, etc. Existing studies show that the time intervals between two consecutive events present different non-Poisson characteristics, such as power-law, Pareto, bimodal distribution of power-law, exponential distribution, piecewise power-law, et al. With the occurrences of new services, new types of distributions may arise. In this paper, we study the distributions of the time intervals between two consecutive visits to QQ and WeChat service, the top two popular instant messaging services in China, and present a new finding that when the value of statistical unit T is set to 0.001s, the inter-event time distribution follows a piecewise distribution of exponential and power-law, indicating the heterogeneous character of IM services users' online behavior in different time scales. We infer that the heterogeneous character is related to the communication mechanism of IM and the habits of users. Then we develop a combination model of exponential model and interest model to characterize the heterogeneity. Furthermore, we find that the exponent of the inter-event time distribution of the same service is different in two cities, which is correlated with the popularity of the services. Our research is useful for the application of information diffusion, prediction of economic development of cities, and so on.

This paper proposes an approach for hurricane damage level prediction using semantic web resources and matrix completion algorithms. Based on the statistical unit node set framework, streaming data from five hurricanes and damage levels from 48 counties in the USA were collected from the SRBench dataset and other web resources, and then trans-coded into matrices. At a time t, the pattern of possible highest damage levels at 6 hours into the future was estimated using a multivariate regression procedure based on singular value decomposition. We also applied soft-impute algorithm and k-nearest neighbours concept to improve the statistical unit node set framework in this research domain. Results showed that the model could deal with inaccurate, inconsistent and incomplete streaming data that were highly sparse, to learn future damage patterns and perform forecasting in near real-time. It was able to estimate the damage levels in several scenarios even if two-thirds of the relevant weather information was unavailable. The contributions of this work will be able to promote the applicability of the semantic web in the context of climate change.

Due to copyright restrictions, the access to the full text of this article is only available via subscription. Unit selection based text-to-speech synthesis (TTS) has been the dominant TTS approach of the last decade. Despite its success, unit selection approach has its disadvantages. One of the most significant disadvantages is the sudden discontinuities in speech that distract the listeners (Speech Commun 51:1039-1064, 2009). The second disadvantage is that significant expertise and large amounts of data is needed for building a high-quality synthesis system which is costly and time-consuming. The statistical speech synthesis (SSS) approach is a promising alternative synthesis technique. Not only that the spurious errors that are observed in the unit selection system are mostly not observed in SSS but also building voice models is far less expensive and faster compared to the unit selection system. However, the resulting speech is typically not as natural-sounding as speech that is synthesized with a high-quality unit selection system. There are hybrid methods that attempt to take advantage of both SSS and unit selection systems. However, existing hybrid methods still require development of a high-quality unit selection system. Here, we propose a novel hybrid statistical/unit selection system for Turkish that aims at improving the quality of the baseline SSS system by improving the prosodic parameters such as intonation and stress. Commonly occurring suffixes in Turkish are stored in the unit selection database and used in the proposed system. As opposed to existing hybrid systems, the proposed system was developed without building a complete unit selection synthesis system. Therefore, the proposed method can be used without collecting large amounts of data or utilizing substantial expertise or time-consuming tuning that is typically required in building unit selection systems. Listeners preferred the hybrid system over the baseline system in the AB preference tests. TÜBİTAK

Summary Objectives: In diagnostic studies, proportions such as sensitivities are often to be calculated and to be compared between different diagnostic procedures. As statistical unit of analysis, multiple observational units may be assessed within each patient, i.e., multiple lesions in an organ. As a requirement, these are to be assessed by multiple blinded readers. In this paper we propose a method to cover correlations between units within patients, correlations between procedures and correlations between different raters assessing each observational unit. Methods: The proposed approach is a two-step method to analyze clustered data with multiple measurements to compare diagnostic procedures in a paired modality design and the correlation between the readers in a paired reader design. The performance of the approach was compared to a generalized estimation equations model (GEEs) by power simulations. Results: Power simulations suggest, that the two-step approach is not inferior to GEEs with regard to the single readers as well as with regard to the average reader. Conclusions: An intuitive approach was developed next to established methods to analyze “paired modality, paired reader” and “unpaired modality, paired reader” studies with binary endpoints when estimating proportions and differences in proportions for clustered data with multiple measurements. This two-step approach is an alternative method to cover routine designs of diagnostic studies where the difference of proportions is to be estimated directly along with confidence intervals.

After collecting data, several transformation steps are needed so that they can be used to provide stable and efficient estimates of a population’s parameters. Statisticians devote considerable effort to their survey designs, using auxiliary information to select a set of units that is representative of the population from which it was selected. In other words, if the sample considers many characteristics of the population, it will probably generate estimates very close to the true and unknown parameters Furthermore, the concept of the quality of the estimates is strongly linked with the possibility that we can measure and control the sampling errors of the design-estimator.

One of the aspects that have the greatest impact on the quality of the results of a survey is the type of frame used to select the units from the population. Nevertheless this aspect is very often underestimated in survey methods. The sampling frame is any material or device used to obtain observational access to the finite population of interest (Sarndal et al. 1992). The sampling frame has significant implications on the cost and the quality of any agricultural or non-agricultural survey.

Classic sampling theory textbooks do not devote a considerable amount of attention to the necessary activities between designing the sample and estimating the parameters of the variables of interest. This disconnect is mainly because the methods used for these activities are often poorly developed. However, these phases of the survey have a considerable impact on both the cost and quality of the results. This chapter presents the various activities that occur during data collection and processing and how they should be organized and conducted.

Let us begin with the first step of the intelligence cycle: data collection. Many businesses gather crucial information – on expenditures and sales, say – but few enter it into a central database for systematic evaluation. The first task of the statistician is to mine this valuable information. Often, this requires skills of persuasion: employees may be hesitant to give up data for the purpose of systematic analysis, for this may reveal past failures.

Data mining, as already noted, is a component of the knowledge discovery process. It can be defined as a set of techniques that allows data analysis and exploration in order to discover significant rules or hidden models within large archives by means of an entirely or partially automated procedure (Berry and Linoff 1997).

Iran is one of the seismically active areas of the world due to its position in the Alpine-Himalayan mountain system. Tehran has several faults hence huge earthquakes will permeates human settlement there. Production of seismic vulnerability map could help disaster management organizations to develop and implement a plan to promote awareness of earthquake vulnerability and implementation of seismic vulnerability reduction measures in Tehran. The process of seismic vulnerability assessment is a supervised classification problem which undertaken by implementation of classification rules obtained from relationships between classes defined by a set of attributes and a unified decision of a group of experts. Therefore, seismic vulnerability assessment is a multidisciplinary problem which needs a multi criteria decision making. The influencing factors make the problem and the process of decision making a complicated disaster management problem. To overcome this problem, this paper proposes an integrated model based upon the granular computing and Dempster–Shafer to extract classification rules for classification of urban areas regarding seismic vulnerability. One of the significant properties of granular computing is induction of more compatible rules having no inconsistency. In this paper, Dempster–Shafer theory is used to integrate and model the conflict among different experts’ viewpoints to get an informed decision regarding the measure of seismic vulnerability in each statistical unit in the study area.

Dealing with large amounts of data or data flows, it can be convenient or necessary to process them in different ‘pieces’; if the data in question refer to different occasions or positions in time or space, a comparative analysis of data stratified in batches can be suitable. The present approach combines clustering and factorial techniques to study the association structure of binary attributes over homogeneous subsets of data; moreover, it seeks to update the result as new statistical units are processed in order to monitor and describe the evolutionary patterns of association.

Due to copyright restrictions, the access to the full text of this article is only available via subscription. Despite its success, unit selection based text-to-speech synthesis (TTS) has has some disadvantages such as sudden discontinuities in speech that distract the listeners. The HMM-based TTS (HTS) approach has been increasingly getting more attention from the TTS research community. One of the advantage is the lack of spurious errors that are observed in the unit selection scheme. Another advantage of the HTS system is the small memory footprint requirement which makes it attractive for embedded devices. Here, we propose a novel hybrid statistical unit selection TTS system for agglutinative languages that aims at improving the quality of the baseline HTS system while keeping the memory footprint small. The intelligibility and quality scores of the baseline system are comparable to the MOS scores of English reported in the Blizzard Challenge tests. Listeners preferred the hybrid system over the baseline system in the A/B preference tests. TÜBİTAK

One of the main characteristics of Semantic Web (SW) data is that it is notoriously incomplete: in the same domain a great deal might be known for some entities and almost nothing might be known for others. A popular example is the well known friend-of-a-friend data set where some members document exhaustive private and social information whereas, for privacy concerns and other reasons, almost nothing is known for other members. Although deductive reasoning can be used to complement factual knowledge based on the ontological background, still a tremendous number of potentially true statements remain to be uncovered. The paper is focused on the prediction of potential relationships and attributes by exploiting regularities in the data using statistical relational learning algorithms. We argue that multivariate prediction approaches are most suitable for dealing with the resulting high-dimensional sparse data matrix. Within the statistical framework, the approach scales up to large domains and is able to deal with highly sparse relationship data. A major goal of the presented work is to formulate an inductive learning approach that can be used by people with little machine learning background. We present experimental results using a friend-of-a-friend data set.

The aim is to show how statistical disclosure methods can be applied to data using the R-packages sdcMicro and sdcTable.

Many papers refer to Tukey’s (1977) treatise on exploratory data analysis as the contribution that transformed statistical thinking. In actual fact, new ideas introduced by Tukey prompted many statisticians to give a more prominent role to data visualization and more generally to data. However, J.W. Tukey in 1962 had already begun his daring provocation when at the annual meeting of the Institute of Mathematical Statistics he gave his talk entitled “The Future of Data Analysis” (Tukey, 1962).

Many real world phenomena are better represented by non-precise data rather than by single-valued data. In fact, non-precise data represent two sources of variability: the natural phenomena variability and the variability or uncertainty induced by measurement errors or determined by specific experimental conditions. The latter variability source is named imprecision. When there are information about the imprecision distribution the fuzzy data coding is used to represent the imprecision. However, in many cases imprecise data are natively defined only by the minimum and maximum values. Technical specifications, stock-market daily prices, survey data are some examples of such kind of data. In these cases, interval data represent a good data coding to take into account the imprecision. This paper aims at describing multiple imprecise data by means of a suitable Principal Component Analysis that is based on specific interval data coding taking into account both sources of variation.

Classification, within a PLS regression framework, is classically meant in the sense of the SIMCA methodology, i.e. as the assignment of statistical units to a-priori defined classes. As a matter of fact, PLS components are built with the double objective of describing the set of explanatory variables while predicting the set of response variables. Taking into account this objective, a classification algorithm is developed that allows to build typologies of statistical units whose different local PLS models have an intrinsic explanatory power higher than the initial global PLS model. The typology induced by the algorithm may undergo a non parametric validation procedure based on bootstrap. Finally, the definition of a compromise model is investigated.

We study the problem of efficiently removing equal frequency n-gram substrings from an n-gram set, formally called Statistical Substring Reduction (SSR). SSR is a useful operation in corpus based multi-word unit research and new word identification task of oriental language processing. We present a new SSR algorithm that has linear time (O(n)) complexity, and prove its equivalence with the traditional O(n2) algorithm. In particular, using experimental results from several corpora with different sizes, we show that it is possible to achieve performance close to that theoretically predicated for this task. Even in a small corpus the new algorithm is several orders of magnitude faster than the O(n2) one. These results show that our algorithm is reliable and efficient, and is therefore an appropriate choice for large scale corpus processing.

Consider a class of statistical units, in which each unit may be an aggregate of individual statistical units. Each unit is decribed by an interval of values for each variable. Our aim is to develop a partition of this class of aggregated statistical units in which each part of the partition is described by a conjunction of characteristic properties. We use a stepwise top-down binary tree method and we introduce a probabilistic approach to assign units to the nodes of the tree. At each step we select the best variable and its best split to optimize simultaneously a discrimination criterion given by a prior partition and a homogeneity criterion. Finally, we present an example of real data.

With the emergence of the knowledge based society, great emphasis is put on the development of qualitative and quantitative policy tools for analysing the science and innovation system. In this chapter an overview is given of the available R&D input data at (supra) national and regional level. With the Frascati Manual the OECD provides a methodological framework for setting up national surveys to collect these data. This methodology is used to produce standardised measurements of human and financial resources devoted to R&D by OECD member countries. EUROSTAT adapted and extended this methodology to produce for the EU countries R&D input data at regional level. To measure the performance of national and regional R&D systems, input and output data have to be combined. The methodologies for collecting input and output data have, however, been developed largely independently from each other. The resulting limitations on their use in performance indicators are discussed, and suggestions are formulated for a more integrated approach to construct input and output data.

Clustering methods and principal axes techniques as well play a major role in the computerized exploration of textual corpora. However, most of the outputs of these unsupervised procedures are difficult to assess. We will focus on the two following issues: External validation, involving external data and allowing for classical statistical tests. Internal validation, based on resampling techniques such as bootstrap and other Monte Carlo methods. In the domain of textual data, these techniques can efficiently tackle the difficult problem of the plurality of statistical units (words, lemmas, segments, sentences, respondents).

In letteratura sono state proposte differenti tecniche per la valutazione della Customer Satisfaction (CS), spesso basate sul concetto di gap. In questo contributo, nel caso di più rilevazioni sulle stesse unità statistiche, viene proposta una strategia di analisi integrata che consente di scomporre l’analisi del gap tra qualità percepita ed attesa in ulteriori componenti da analizzare ai fini di un miglioramento del servizio.

Production data of oil fields are provided as decline curves (oil and water production vs time), that the user wants to gather in a limited number of clusters. Preprocessing of data is required to remove noise, and provides a complete data set, involving for each statistical unit (wells) extraction of attributes from smoothed or modelized curves. Hierarchical clustering is performed in two steps to avoid smaller or outlier cluster ; firstly the centroid clustering method is used to recognize and then discard clusters having a lower frequency, this is followed by application of the Ward-method. Finally, using the central part of these previous (Ward) clusters, discriminant analysis is performed, including all the discarded units. This sequence avoids the disturbing influence of outlying units, and also gives the probability for each unit to be classified in the clusters.

In the present paper we initially intend to show the fundamental ideas behind the methodological achievements of Non Symmetrical Data Analysis from a geometrical point of view. We then focus on some of the most recent extensions by stressing and giving insights on their application aspects. Finally, we outline what seem to be the most promising directions of further research in this field.

In a series of 16 oxyphilic follicular neoplasms of the thyroid (8 adenomas and 8 carcinomas), three different approaches for the analysis of morphometric data were evaluated. It was shown that the statistical design of morphometric studies is by nature nested due to subsampling of cells within each patient. Therefore, the most appropriate analysis would be to account for this hierarchical structure. However, related statistical methods are not at present well established, especially as far as classification rules are concerned. Therefore, the nested design is converted into the simple factorial one by considering only one kind of statistical unit – either patients or cells. The results of the study presented indicate that ignoring the patient as unit of analysis leads to a substantial error in statistical output, regardless of the particular procedure applied. Moreover, the size of the error can be neither diminished nor controlled. Choosing patients as primary units assures accurate results and also has an advantage of gaining some additional information by calculating several distributional estimates in each patient. However, this approach often requires a reduction of dimensions and, furthermore, is not encouraged in certain fields of quantitative cytology. Advantages and disadvantages of all approaches have been summarized and practical recommendations for their use have been worked out.

In multiple time series analysis, when there are a very large number of series, a classification into homogeneous clusters might be useful to reduce the problem’s complexity and eliminate possible redundancies (Zani, 1983). Furthermore, when we have different classifications, one for each statistical unit (e. g. spatial units), a consensus classification allows one to obtain a classification which summarizes the given classifications. The present paper focuses on the problem of identifying consensus classifications in a set of multiple time series (panel data), using a consensus method (Vichi, 1993, 1994). First, a distance among time series is defined and a hierarchical classification among time series, for each temporal lag and for each unit, is performed. Then, a consensus classification among different units for the same temporal lag is carried out. Finally, a hierarchical classification among the different consensus classifications, with the same temporal lag, is carried out.

In this paper we describe a new language for statistical data modelling, which offers a general framework for the representation of elementary and summary data. Three are the main characteristics of the language: 1) the types of modeling primitives of the language are particularly suited for representing objects from a statistical point of view; 2) the language includes a rich set of structuring mechanisms for both elementary and summary data, which are given a formal semantics by means of logic; 3) the language is equipped with specialized inference procedures, allowing for performing different kinds of checks on the representation.

Producing the Principal Components Analysis of a data table requires choices which need to be explained in order to acquire complete understanding of the results. This explicitness opens the road to other possible choices, leading to the oretical research and many practical applications. Changes of scale, changes of variables, weighting of statistical units, decentering of the representations, and elimination of dependence between individuals are dealt with. After reviewing the usual methods from t h i s perspective, it can be seen that it is possible to transform them in order to better adapt mathematical abstractions to concrete situations.

This chapter presents the basic definitions the collection of statistical data. The elementary unit is the unit—person, object, or event—whose characteristics are to be described in the course of the statistical inquiry. A statistical population universe is the aggregate of the statistical units considered. A list, map, or other description of all these units is called the frame of the inquiry. The scope of a statistical inquiry determines which characteristics of each statistical unit are to be recorded. The statistical inquiry may be intended to collect detailed information on a single topic. The design of a statistical inquiry determines the time when it is made, its periods, its method, its coverage, and its sources. The observer may note the number or proportion of units marked by the presence or absence of some qualitative characteristic, or may measure the actual magnitude of some measurable characteristic for each unit observed.

The PNS module is discussed as the building block for the synthesis of parallel, self-organizing, hierarchical, neural networks (PSHNNs). The PNS module contains three submodules (units), the first two of which are created as simple neural network constructs and the last of which is a statistical unit. The first two units are fractile in nature, meaning that each such unit may itself consist of a number of parallel PNS modules in a fractile fashion. Through a mechanism of statistical acceptance or rejection of input vectors for classification, the sample space is divided into a number of regions. The input vectors belonging to each region are classified by a dedicated set of PNS modules. This strategy results in considerably higher accuracy of classification and better generalization as compared to previous neural network models. If the delta rule network is used to generate the first two units, each region approximates a linearly separable region. In this sense, the total system becomes similar to a piecewise linear model. The various regions are determined nonlinearly by the first and third units of the PNS modules.

Jerome Cornfield's adventures in statistics have already been effectively and expertly encapsulated by Cornfield himself in his Presidential Address to the American Statistical Association (Cornfield, 1975). The address is revealing of a person with a broad range of interests and a developed or developing philosophy. One gets a glimpse of Cornfield as an undergraduate and sees also some of his earliest work as a statistician at the Bureau of Labor Statistics. Cornfield testifies that a certain minimization problem that he encountered in those early days was beyond his powers to develop-but he does not claim credit for his important contribution in identifying the problem and helping in its solution, whereby the subject of linear programming (input-output studies) was born. Only one specific instance of laboratory research is detailed in that Presidential Address, one relating to the toxicity of mixtures of amino acids (Gullino et al., 1956). But it is evident that Cornfield has identified himself as a member of the investigative team with which he was working, largely biochemists, making their concerns his concerns. Elsewhere in the address, Cornfield discusses the importance of the interaction between theory and practice that such collaboration brings about, emphasizing also the joy and satisfaction that it brought him. In the instant case involving amino acids, the first step of the investigation was to distinguish between response-wise and dose-wise additivity for mixtures of amino acids. Perhaps it was only because of the statistical concept of interaction, i.e., departure from simple additivity, that the response-wise additivity model was even in the running, but the data flatly contradicted so simplistic a model. As further investigations brought out, the dose-wise additive model was also contradicted but, nevertheless, it was a more logical null model by which to judge the activity of mixtures. Toxicologists and other workers with chemical combinations were aware of that point while many statisticians were still wedded to their own concepts of interaction. This state of identification of the statistician with the investigative team and of involvement with the problem at hand was greatly removed from what had earlier been the case. Cornfield had been recruited in 1947 by Dr. Harold F. Dorn to head up a statistical unit in Dr. Dorn's organization at the National Institutes of Health, which organization was subsequently blanketed into the National Cancer Institute. As additional professional staff for that statistics unit, Dorn recruited, between 1946 and 1948, Jacob Lieberman, Samuel Greenhouse, Marvin Schneiderman, and myself, all of whom, like Cornfield, had their prior statistical experience limited to areas in economics. Cornfield's wider experience at the Bureau of Labor Statistics probably came closest in statistical sophistication to what this team would be called on to do at the National Cancer Institute. Anyway, those early years probably marked Cornfield's, and his staff's, most active involvement in laboratory research, though that activity was little reflected in publications on which he was an author or co-author an occasional acknowledgement was as much as might be expected. But this was