Time Series

ForeCA: Forecastable Component Analysis

by Georg Goerg

Submitted for publication.

Blind source separation (BSS) techniques are often applied to multivariate time series with the goal to obtain better... more Blind source separation (BSS) techniques are often applied to multivariate time series with the goal to obtain better forecasts. But BSS and the need for better forecasts are often treated separately, in the sense that finding an optimally transformed (sub-)space has nothing to do with the aim to predict well. Here I introduce Forecastable Component Analysis (ForeCA), a new BSS technique for temporally dependent signals that uses forecastability as the explicit objective in finding an optimal transformation. It separates the signal into the forecastable, $\mathbf{F}$, and the orthogonal white noise space, $\mathbf{F}^{\bot}$. Simulations and applications to financial data show that ForeCA successfully finds signals that can be used to forecast. ForeCA therefore automatically discovers informative structure in multivariate signals. The R package ForeCA (http://cran.r-project.org/web/packages/ForeCA/index.html) will be publicly available on CRAN upon publication of the manuscript.

Aplicação do algoritmo "Perceptually Important Points" em séries temporais de datacenters

by Ronald Kaiser

Sorry, only in portuguese.

The main purpose of this work is helping operators of  datacenters in the task of visualizing the behaviour of... more The main purpose of this work is helping operators of  datacenters in the task of visualizing the behaviour of their devices and services through time, represented by large
time series. In order to accomplish that, a technique used in pattern recognition from the financial market context was choosed. The “Perceptually Important Points” algorithm
gives a method for dimensionality reduction and a mechanism to automatically extract the most important points from a human observer perspective, favouring compression and a good visualization of time series with high dimensionality. The implementation of the algorithm and its integration in an existing monitoring system was explored and encompasses the content of this work.

THE ROOTS OF SUCCESS: INDUSTRIAL GROWTH IN ITALY RECONSIDERED, 1911–1951

by Emanuele Felice

co-authored with Albert Carreras

This article reconsiders the growth of Italian industry from the First World War to the eve of the economic miracle,... more This article reconsiders the growth of Italian industry from the First World War to the eve of the economic miracle, with the aid of sector-specific new value-added series, at three different price-bases. The new estimates reduce growth during the First World War, making the Italian case com-parable to the other belligerent countries, while improving the performance of the 1920s. The 1929 crisis looks more profound than before, while the recovery after 1933 is now stronger. During the 1920s and the 1930s, a significant shift from traditional to more advanced activities took place: when confronted with the rest of Europe, the interwar period was a relative success, which laid the ground for the following economic boom.

Echo State Networks for Online Prediction of Movement Data --- Comparing Investigations

by Sören Strauss

Proceedings of the 18th international conference on Artificial Neural Networks, 2008

Postmaterialism as a lifetime learning process

by Raul Tormos

presented at the conference of the European Survey Research Association, Warsaw 2009.

Vector-dependent Functionally Pooled ARX Models for the Identification of Systems Under Multiple Operating Conditions

by Fotis Kopsaftopoulos

Proceedings of the 16th IFAC Symposium on System Identification, (SYSID), Brussels, Belgium, July 2012.

Neural Network Time Series Models for Financial Risk Management

by Francesco Virili

Nonstationarity and data preprocessing for neural network predictions of an economic time series

by Francesco Virili

Econometric analysis of global climate change

by David Stern

Investigation of Time Series Modelling to Explicate the Learning Related Activity Observed in fMRI Data

by Krishna Prasad Miyapuram

V. S. Chandrasekhar Pammi, K. P. Miyapuram, Raju S. Bapi, Chakravarthy Bhagavati,T. Brahrnaiah, Ch. Chandra Sekhar Rao
Intemational Conference on Systernies. Cybernetics and Informatics. February 12-15, 2004

Categories and Subject Descriptors
Cybernetics: Neural and Cognitive Modeling
Paper Identification Number: NCM-6
This paper has been published by the Pentagram Research Centre (R) Limited.

Functional Magnetic Resonance imaging (fMRI) technique has so far demonstrated its usefulness in revealing the neural... more Functional Magnetic Resonance imaging (fMRI) technique has so far demonstrated its usefulness in revealing the neural correlates for higher-order cognitive functions such as memory, attention, emotion. language. etc. We report results from our efforts toward understanding the neural !lases for learning visuo¬motor sequences. Using the fMRI technique, we tested learning related effects corresponding to a sequential finger-pressing task
2x6 task. Behavioural analysis revealed improvements in performance measures of subjects such as speed and accuracy, pointing out successful learning of the sequence of finger movements accomplished by subjects. Traditional imaging analysis using the Statistical Parametric Mapping (SPM) approach revealed activation in several brain areas. However. SPM analysis averages out temporal changes and presents gross differences between the early and late stages of learning. As the main focus of the study is to understand the dynamic changes in brain activation induced by the learning process, the time course of activity of voxels in this area is modelled by statistical methods of time series analysis. Results from ARIMA model of activation values from the different brain regions revealed that the model parameters are distinctly different for the control/follow condition and the test/sequence condition. In the control condition subjects made finger movements in response to random visual cues and hence there was no learning. In contrast, in the sequence condition subjects progressively learned a fixed sequence of movements by responding to visual cues. A detailed analysis is required to tease out learning specific activity in the time series corresponding to the sequence epochs, for example to bring out differences in the early and late periods of the learning process. Although these results are still preliminary, they demonstrate that explicit modelling of fMRI time series is important in revealing and characterizing learning induced changes in the brain activity. Further, these methods seem more useful for investigating transient and long-term changes across the learning period than the traditional analysis methods.

Similarity-based forecasting with simultaneous previews: A river plot interface for time series forecasting

by Paolo Buono

Paolo Buono, Catherine Plaisant, Adalberto Simeone, Aleks Aris, Galit Shmueli, and Wolfgang Jank. 2007. Similarity-Based Forecasting with Simultaneous Previews: A River Plot Interface for Time Series Forecasting. In Proceedings of the 11th International Conference Information Visualization (IV '07). IEEE Computer Society, Washington, DC, USA, 191-196. DOI=10.1109/IV.2007.101 http://dx.doi.org/10.1109/IV.2007.101

Time-series forecasting has a large number of applications. Users with a partial time series for auctions, new stock... more Time-series forecasting has a large number of applications. Users with a partial time series for auctions, new stock offerings, or industrial processes desire estimates of the future behavior. We present a data driven forecasting method and interface called Similarity-Based Forecasting (SBF). A pattern matching search in an historical time series dataset produces a subset of curves similar to the partial time series. The forecast is displayed graphically as a river plot showing statistical information about the SBF subset. A forecasting preview interface allows users to interactively explore alternative pattern matching parameters and see multiple forecasts simultaneously. User testing with 8 users demonstrated advantages and led to improvements.

Interactive pattern search in time series

by Paolo Buono

The need for pattern discovery in long time series data led researchers to develop algorithms for similarity search.... more The need for pattern discovery in long time series data led researchers to develop algorithms for similarity search. Most of the literature about time series focuses on algorithms that index time series and bring the data into the main storage, thus providing fast information retrieval on large time series. This paper reviews the state of the art in visualizing time series, and focuses on techniques that enable users to interactively query time series. Then it presents TimeSearcher 2, a tool that enables users to explore multidimensional data using coordinated tables and graphs with overview+detail, filter the time series data to reduce the scope of the search, select an existing pattern to find similar occurrences, and interactively adjust similarity parameters to narrow the result set. This tool is an extension of previous work, TimeSearcher 1, which uses graphical timeboxes to interactively query time series data.

Interactive shape specification for pattern search in time series

by Paolo Buono

Paolo Buono and Adalberto Lafcadio Simeone. 2008. Interactive shape specification for pattern search in time series. In Proceedings of the working conference on Advanced visual interfaces (AVI '08). ACM, New York, NY, USA, 480-481. DOI=10.1145/1385569.1385666 http://doi.acm.org/10.1145/1385569.1385666

Time series analysis is a process whose goal is to understand phenomena. The analysis often involves the search for a... more Time series analysis is a process whose goal is to understand phenomena. The analysis often involves the search for a specific pattern. Finding patterns is one of the fundamental steps for time series observation or forecasting. The way in which users are able to specify a pattern to use for querying the time series database is still a challenge. We hereby propose an enhancement of the SearchBox, a widget used in TimeSearcher, a well known tool developed at the University of Maryland that allows users to find patterns similar to the one of interest.

Structural health monitoring by Lyapunov exponents of non‐linear time series

by Sara Casciati

Casciati F. and Casciati S. (2006). “Structural health monitoring by Lyapunov exponents of nonlinear time series”. Structural Control & Health Monitoring, 13(1), 132-146. ISSN: 1545-2255.

DATE AND PLACE OF PUBLICATION: January-February 2006; John Wiley & Sons, Ltd., Chichester PO19 8SQ, W Sussex, England.

ABSTRACT. In this study, structural health monitoring is pursued by collecting multi-channel measurements and by... more ABSTRACT. In this study, structural health monitoring is pursued by collecting multi-channel measurements and by computing, directly from them, the Lyapunov exponents. The latter quantities are invariants of the dynamic system, so that their different values, associated with different time histories obtained from the same structure, denote damage. First, the problem is framed in the general theory. The structural health monitoring strategy is then formulated, with special care being devoted to its capability of localizing damage. The procedure is finally validated by using the time histories which were collected during the experimental tests on the model of a monumental arch.

KEY WORDS: dynamic system; Kolmogorov entropy; Lyapunov dimension; Lyapunov exponents; observed variables space; structural health monitoring

Testing for white noise against locally stationary alternatives

by Georg Goerg

Winner of the 2012 JSM student paper competition in Statistical Learning and Data Mining; Under revision for Statistical Analysis and Data Mining (SAM)

Many real-world systems have dynamics that evolve over time, yet stationary models still remain a popular choice in... more Many real-world systems have dynamics that evolve over time, yet stationary models still remain a popular choice in empirical time series studies. In this work I show that one reason for seemingly correct stationary ts is a very low power of classic white noise tests against locally varying dynamics. In particular, if autocorrelations change over time but on average equal zero, standard white noise tests cannot detect this deviation from the null hypothesis due to their fundamental design. Here I introduce a moving-window version of the Ljung-Box statistic with an asymptotic chi-square distribution under the null and much larger power facing processes with time-varying autocorrelations. Simulations and a case study of tree-ring data demonstrate the importance of the new test for applied time series studies.

Hegel e la filosofia contemporanea del tempo [Hegel and the contemporary philosophy of time]

by Federico PERELDA

Draft version of "Hegel e la filosofia contemporanea del tempo", published in  "Verifiche" XXXIX, 1-4, 2010, pp. 135-85.

In this essay I compare Hegel’s theory of time and becoming with the contemporary debate, aiming on the one hand (A)... more In this essay I compare Hegel’s theory of time and becoming with the contemporary debate, aiming on the one hand (A) at presenting Hegel’s thought in contemporary terms, and on the other, (B) at offering new inputs to the present metaphysical debate from a Hegelian point of view. From a close reading of selected Hegelian texts I argue (1) that Hegel advocates a form of presentism and shares McTaggart’s thesis that the B-series (chronological time) presupposes the A-series (dynamical time); (2) that his position is pe-culiar because, although he admits that change is inconsistent, he puts in jeopardy the law of non contradiction (at least in its universality), instead of denying the reality of time and change, like McTaggart did. These considerations will then lead to Hegel’s speculative logic. According to the so called coherentistic reading of Hegel’s thought, he never seriously questioned the principle of non contradiction: he would be just a very sophisticated Aristo-telian, after all. I oppose this view, arguing (3) that Hegel was a proponent of an articulated form of dialetheism.

Relaxed Selection Techniques for Querying Time-Series Graphs

by Christian Holz

Holz, C. and Feiner, S. 2009. Relaxed Selection Techniques for Querying Time-Series Graphs. In Proceedings of UIST '09, 213–222.

presentation: http://www.christianholz.net/2009-uist09-holz-relaxed_selection_techniques-presentation.pdf
more information: http://www.christianholz.net/relaxed_selection_techniques.html

Time-series graphs are often used to visualize phenomena that change over time. Common tasks include comparing values... more Time-series graphs are often used to visualize phenomena that change over time. Common tasks include comparing values at different points in time and searching for specified patterns, either exact or approximate. However, tools that support time-series graphs typically separate query specification from the actual search process, allowing users to adapt the level of similarity only after specifying the pattern. We introduce relaxed selection techniques, in which users implicitly define a level of similarity that can vary across the search pattern, while creating a search query with a single-gesture interaction. Users sketch over part of the graph, establishing the level of similarity through either spatial deviations from the graph, or the speed at which they sketch (temporal deviations). In a user study, participants were significantly faster when using our temporally relaxed selection technique than when using traditional techniques. In addition, they achieved significantly higher precision and recall with our spatially relaxed selection technique compared to traditional techniques.

Error-Correction as a Concept and as a Method: Time Series Analysis of Policy-Opinion Responsiveness

by Will Jennings

Will Jennings. (n.d.). ‘Error-Correction as a Concept and as a Method: Time Series Analysis of Policy-Opinion Responsiveness’. In Martin Lodge and Michael Bruter (Eds.) From the Engine Room: Methods and Approaches in the Social Sciences.

How can researchers analyse the inter-relationship between government policies and public opinion? What are the... more How can researchers analyse the inter-relationship between government policies and public opinion? What are the implications of choosing particular measures over others? And to what extent do the statistical peculiarities of model specification impact upon empirical findings? This chapter introduces theoretical approaches to analysis of policy-opinion responsiveness and proceeds to unravel the puzzle of how to measure, model and test interactions between public policy and public opinion over time. It considers the example of research on the opinion-responsiveness of the British asylum system over the period between 1994 and 2007 (Jennings 2009): a policy subsystem subject to intense and emotive media coverage, politicization and legislative activism throughout the past decade. The chapter discusses a number of features of the engine room: research context and case selection, the statistical properties of government policies and outputs and public opinion, normative implications of empirical findings, and the problem of spuriousness in statistical inference of opinion-responsiveness.

Macroeconomics, Crime Rates and Time: an analysis of recorded property crime, unemployment and income inequality in Britain 1961-2006

by Will Jennings

Will Jennings, Stephen Farrall and Shaun Bevan, (In Press) ‘The Economy, Crime and Time: an analysis of recorded property crime in Britain, 1961-2006’, forthcoming in the International Journal of Law, Crime and Justice.

We seek to determine whether one of the unanticipated side-effects of social and economic changes associated with the... more We seek to determine whether one of the unanticipated side-effects of social and economic changes associated with the adoption of neoliberal and monetarist economics during the 1970s/1980s was rising crime rates. Undertaking time series analysis of social and economic determinants of property crime (using official statistics on recorded crime for England and Wales from 1961-2006) we develop a model of the effect of changes in socio-economic variables (unemployment, inequality, welfare spending and incarceration) on property crime rates. We find that while three of these had significant effects on change in the property crime rate, income inequality did not. We conclude with a discussion of the extent to which neoliberal economic and welfare (and later criminal justice) policies can be held to have influenced the property crime rate since the early 1980s and what this tells us about the social and economic determinants of crime at the macro-level.

Effects of the Core Functions of Government on the Diversity of Executive Agendas

by Will Jennings

Will Jennings, Shaun Bevan, Arco Timmermans, Laura Chaques, Gerard Breeman, Sylvain Brouard, Christoffer Green-Pedersen, Peter John, Peter B. Mortensen and Anna Palau. (2011). ‘Effects of the Core Functions of Government on the Diversity of Executive Agendas’, Comparative Political Studies 44(8): 1001-1030.

The distribution of attention across issues is of fundamental importance to the political agenda and outputs of... more The distribution of attention across issues is of fundamental importance to the political agenda and outputs of government. This article presents an issue-based theory of the diversity of governing agendas where the core functions of government—defense, international affairs, the economy, government operations, and the rule of law—are prioritized ahead of all other issues. It undertakes comparative analysis of issue diversity of the executive agenda of several European countries and the United States over the postwar period. The results offer strong evidence of the limiting effect of core issues—the economy, government operations, defense, and international affairs—on agenda diversity. This suggests not only that some issues receive more attention than others but also that some issues are attended to only at times when the agenda is more diverse. When core functions of government are high on the agenda, executives pursue a less diverse agenda—focusing the majority of their attention on fewer issues. Some issues are more equal than others in executive agenda setting.