Análisis de sistemas dinámicos y complejos en la Liga Profesional de Balonmano de España = Complex and dynamical systems analysis in Spanish Professional Handball League

El propósito de esta tesis fue estudiar el rendimiento ofensivo de los equipos de balonmano de élite cuando se considera el balonmano como un sistema dinámico complejo no lineal. La perspectiva de análisis dinámica dependiente del tiempo fue adoptada para evaluar el rendimiento de los equipos durante el partido. La muestra general comprendió los 240 partidos jugados en la temporada 2011-2012 de la liga profesional masculina de balonmano de España (Liga ASOBAL). En el análisis posterior solo se consideraron los partidos ajustados (diferencia final de goles ≤ 5; n = 142). El estado del marcador, la localización del partido, el nivel de los oponentes y el periodo de juego fueron incorporados al análisis como variables situacionales. Tres estudios compusieron el núcleo de la tesis. En el primer estudio, analizamos la coordinación entre las series temporales que representan el proceso goleador a lo largo del partido de cada uno de los dos equipos que se enfrentan. Autocorrelaciones, correlaciones cruzadas, doble media móvil y transformada de Hilbert fueron usadas para el análisis. El proceso goleador de los equipos presentó una alta consistencia a lo largo de todos los partidos, así como fuertes modos de coordinación en fase en todos los contextos de juego. Las únicas diferencias se encontraron en relación al periodo de juego. La coordinación en los procesos goleadores de los equipos fue significativamente menor en el 1er y 2º periodo (0–10 min y 10–20 min), mostrando una clara coordinación creciente a medida que el partido avanzaba. Esto sugiere que son los 20 primeros minutos aquellos que rompen los partidos. En el segundo estudio, analizamos los efectos temporales (efecto inmediato, a corto y a medio plazo) de los tiempos muertos en el rendimiento goleador de los equipos. Modelos de regresión lineal múltiple fueron empleados para el análisis. Los resultados mostraron incrementos de 0.59, 1.40 y 1.85 goles para los periodos que comprenden la primera, tercera y quinta posesión de los equipos que pidieron el tiempo muerto. Inversamente, se encontraron efectos significativamente negativos para los equipos rivales, con decrementos de 0.50, 1.43 y 2.05 goles en los mismos periodos respectivamente. La influencia de las variables situacionales solo se registró en ciertos periodos de juego. Finalmente, en el tercer estudio, analizamos los efectos temporales de las exclusiones de los jugadores sobre el rendimiento goleador de los equipos, tanto para los equipos que sufren la exclusión (inferioridad numérica) como para los rivales (superioridad numérica). Se emplearon modelos de regresión lineal múltiple para el análisis. Los resultados mostraron efectos negativos significativos en el número de goles marcados por los equipos con un jugador menos, con decrementos de 0.25, 0.40, 0.61, 0.62 y 0.57 goles para los periodos que comprenden el primer, segundo, tercer, cuarto y quinto minutos previos y posteriores a la exclusión. Para los rivales, los resultados mostraron efectos positivos significativos, con incrementos de la misma magnitud en los mismos periodos. Esta tendencia no se vio afectada por el estado del marcador, localización del partido, nivel de los oponentes o periodo de juego. Los incrementos goleadores fueron menores de lo que se podría esperar de una superioridad numérica de 2 minutos. Diferentes teorías psicológicas como la paralización ante situaciones de presión donde se espera un gran rendimiento pueden ayudar a explicar este hecho. Los últimos capítulos de la tesis enumeran las conclusiones principales y presentan diferentes aplicaciones prácticas que surgen de los tres estudios. Por último, se presentan las limitaciones y futuras líneas de investigación. ABSTRACT The purpose of this thesis was to investigate the offensive performance of elite handball teams when considering handball as a complex non-linear dynamical system. The time-dependent dynamic approach was adopted to assess teams’ performance during the game. The overall sample comprised the 240 games played in the season 2011-2012 of men’s Spanish Professional Handball League (ASOBAL League). In the subsequent analyses, only close games (final goal-difference ≤ 5; n = 142) were considered. Match status, game location, quality of opposition, and game period situational variables were incorporated into the analysis. Three studies composed the core of the thesis. In the first study, we analyzed the game-scoring coordination between the time series representing the scoring processes of the two opposing teams throughout the game. Autocorrelation, cross-correlation, double moving average, and Hilbert transform were used for analysis. The scoring processes of the teams presented a high consistency across all the games as well as strong in-phase modes of coordination in all the game contexts. The only differences were found when controlling for the game period. The coordination in the scoring processes of the teams was significantly lower for the 1st and 2nd period (0–10 min and 10–20 min), showing a clear increasing coordination behavior as the game progressed. This suggests that the first 20 minutes are those that break the game-scoring. In the second study, we analyzed the temporal effects (immediate effect, short-term effect, and medium-term effect) of team timeouts on teams’ scoring performance. Multiple linear regression models were used for the analysis. The results showed increments of 0.59, 1.40 and 1.85 goals for the periods within the first, third and fifth timeout ball possessions for the teams that requested the timeout. Conversely, significant negative effects on goals scored were found for the opponent teams, with decrements of 0.59, 1.43 and 2.04 goals for the same periods, respectively. The influence of situational variables on the scoring performance was only registered in certain game periods. Finally, in the third study, we analyzed the players’ exclusions temporal effects on teams’ scoring performance, for the teams that suffer the exclusion (numerical inferiority) and for the opponents (numerical superiority). Multiple linear regression models were used for the analysis. The results showed significant negative effects on the number of goals scored for the teams with one less player, with decrements of 0.25, 0.40, 0.61, 0.62, and 0.57 goals for the periods within the previous and post one, two, three, four and five minutes of play. For the opponent teams, the results showed positive effects, with increments of the same magnitude in the same game periods. This trend was not affected by match status, game location, quality of opposition, or game period. The scoring increments were smaller than might be expected from a 2-minute numerical playing superiority. Psychological theories such as choking under pressure situations where good performance is expected could contribute to explain this finding. The final chapters of the thesis enumerate the main conclusions and underline the main practical applications that arise from the three studies. Lastly, limitations and future research directions are described.

Brown trout thermal niche and climate change: expected changes in the distribution of cold-water fish in central Spain

This paper addresses the determination of the realized thermal niche and the effects of climate change on the range distribution of two brown trout populations inhabiting two streams in the Duero River basin (Iberian Peninsula) at the edge of the natural distribution area of this species. For reaching these goals, new methodological developments were applied to improve reliability of forecasts. Water temperature data were collected using 11 thermographs located along the altitudinal gradient, and they were used to model the relationship between stream temperature and air temperature along the river continuum. Trout abundance was studied using electrofishing at 37 sites to determine the current distribution. The RCP4.5 and RCP8.5 change scenarios adopted by the International Panel of Climate Change for its Fifth Assessment Report were used for simulations and local downscaling in this study. We found more reliable results using the daily mean stream temperature than maximum daily temperature and their respective seven days moving-average to determine the distribution thresholds. Thereby, the observed limits of the summer distribution of brown trout were linked to thresholds between 18.1ºC and 18.7ºC. These temperatures characterise a realised thermal niche narrower than the physiological thermal range. In the most unfavourable climate change scenario, the thermal habitat loss of brown trout increased to 38% (Cega stream) and 11% (Pirón stream) in the upstream direction at the end of the century; however, at the Cega stream, the range reduction could reach 56% due to the effect of a ?warm-window? opening in the piedmont reach.

What is a linear process?

We argue that given even an infinitely long data sequence, it is impossible (with any test statistic) to distinguish perfectly between linear and nonlinear processes (including slightly noisy chaotic processes). Our approach is to consider the set of moving-average (linear) processes and study its closure under a suitable metric. We give the precise characterization of this closure, which is unexpectedly large, containing nonergodic processes, which are Poisson sums of independent and identically distributed copies of a stationary process. Proofs of these results will appear elsewhere.

Impact of artifact correction methods on R-R interbeat signals to quantifying heart rate variability (HRV) according to linear and nonlinear methods.

In the analysis of heart rate variability (HRV) are used temporal series that contains the distances between successive heartbeats in order to assess autonomic regulation of the cardiovascular system. These series are obtained from the electrocardiogram (ECG) signal analysis, which can be affected by different types of artifacts leading to incorrect interpretations in the analysis of the HRV signals. Classic approach to deal with these artifacts implies the use of correction methods, some of them based on interpolation, substitution or statistical techniques. However, there are few studies that shows the accuracy and performance of these correction methods on real HRV signals. This study aims to determine the performance of some linear and non-linear correction methods on HRV signals with induced artefacts by quantification of its linear and nonlinear HRV parameters. As part of the methodology, ECG signals of rats measured using the technique of telemetry were used to generate real heart rate variability signals without any error. In these series were simulated missing points (beats) in different quantities in order to emulate a real experimental situation as accurately as possible. In order to compare recovering efficiency, deletion (DEL), linear interpolation (LI), cubic spline interpolation (CI), moving average window (MAW) and nonlinear predictive interpolation (NPI) were used as correction methods for the series with induced artifacts. The accuracy of each correction method was known through the results obtained after the measurement of the mean value of the series (AVNN), standard deviation (SDNN), root mean square error of the differences between successive heartbeats (RMSSD), Lomb\'s periodogram (LSP), Detrended Fluctuation Analysis (DFA), multiscale entropy (MSE) and symbolic dynamics (SD) on each HRV signal with and without artifacts. The results show that, at low levels of missing points the performance of all correction techniques are very similar with very close values for each HRV parameter. However, at higher levels of losses only the NPI method allows to obtain HRV parameters with low error values and low quantity of significant differences in comparison to the values calculated for the same signals without the presence of missing points.

Sur les tests lisses d'ajustement dans le context des series chronologiques

La plupart des modèles en statistique classique repose sur une hypothèse sur la distribution des données ou sur une distribution sous-jacente aux données. La validité de cette hypothèse permet de faire de l’inférence, de construire des intervalles de confiance ou encore de tester la fiabilité du modèle. La problématique des tests d’ajustement vise à s’assurer de la conformité ou de la cohérence de l’hypothèse avec les données disponibles. Dans la présente thèse, nous proposons des tests d’ajustement à la loi normale dans le cadre des séries chronologiques univariées et vectorielles. Nous nous sommes limités à une classe de séries chronologiques linéaires, à savoir les modèles autorégressifs à moyenne mobile (ARMA ou VARMA dans le cas vectoriel). Dans un premier temps, au cas univarié, nous proposons une généralisation du travail de Ducharme et Lafaye de Micheaux (2004) dans le cas où la moyenne est inconnue et estimée. Nous avons estimé les paramètres par une méthode rarement utilisée dans la littérature et pourtant asymptotiquement efficace. En effet, nous avons rigoureusement montré que l’estimateur proposé par Brockwell et Davis (1991, section 10.8) converge presque sûrement vers la vraie valeur inconnue du paramètre. De plus, nous fournissons une preuve rigoureuse de l’inversibilité de la matrice des variances et des covariances de la statistique de test à partir de certaines propriétés d’algèbre linéaire. Le résultat s’applique aussi au cas où la moyenne est supposée connue et égale à zéro. Enfin, nous proposons une méthode de sélection de la dimension de la famille d’alternatives de type AIC, et nous étudions les propriétés asymptotiques de cette méthode. L’outil proposé ici est basé sur une famille spécifique de polynômes orthogonaux, à savoir les polynômes de Legendre. Dans un second temps, dans le cas vectoriel, nous proposons un test d’ajustement pour les modèles autorégressifs à moyenne mobile avec une paramétrisation structurée. La paramétrisation structurée permet de réduire le nombre élevé de paramètres dans ces modèles ou encore de tenir compte de certaines contraintes particulières. Ce projet inclut le cas standard d’absence de paramétrisation. Le test que nous proposons s’applique à une famille quelconque de fonctions orthogonales. Nous illustrons cela dans le cas particulier des polynômes de Legendre et d’Hermite. Dans le cas particulier des polynômes d’Hermite, nous montrons que le test obtenu est invariant aux transformations affines et qu’il est en fait une généralisation de nombreux tests existants dans la littérature. Ce projet peut être vu comme une généralisation du premier dans trois directions, notamment le passage de l’univarié au multivarié ; le choix d’une famille quelconque de fonctions orthogonales ; et enfin la possibilité de spécifier des relations ou des contraintes dans la formulation VARMA. Nous avons procédé dans chacun des projets à une étude de simulation afin d’évaluer le niveau et la puissance des tests proposés ainsi que de les comparer aux tests existants. De plus des applications aux données réelles sont fournies. Nous avons appliqué les tests à la prévision de la température moyenne annuelle du globe terrestre (univarié), ainsi qu’aux données relatives au marché du travail canadien (bivarié). Ces travaux ont été exposés à plusieurs congrès (voir par exemple Tagne, Duchesne et Lafaye de Micheaux (2013a, 2013b, 2014) pour plus de détails). Un article basé sur le premier projet est également soumis dans une revue avec comité de lecture (Voir Duchesne, Lafaye de Micheaux et Tagne (2016)).

Sur les tests lisses d'ajustement dans le context des series chronologiques

La plupart des modèles en statistique classique repose sur une hypothèse sur la distribution des données ou sur une distribution sous-jacente aux données. La validité de cette hypothèse permet de faire de l’inférence, de construire des intervalles de confiance ou encore de tester la fiabilité du modèle. La problématique des tests d’ajustement vise à s’assurer de la conformité ou de la cohérence de l’hypothèse avec les données disponibles. Dans la présente thèse, nous proposons des tests d’ajustement à la loi normale dans le cadre des séries chronologiques univariées et vectorielles. Nous nous sommes limités à une classe de séries chronologiques linéaires, à savoir les modèles autorégressifs à moyenne mobile (ARMA ou VARMA dans le cas vectoriel). Dans un premier temps, au cas univarié, nous proposons une généralisation du travail de Ducharme et Lafaye de Micheaux (2004) dans le cas où la moyenne est inconnue et estimée. Nous avons estimé les paramètres par une méthode rarement utilisée dans la littérature et pourtant asymptotiquement efficace. En effet, nous avons rigoureusement montré que l’estimateur proposé par Brockwell et Davis (1991, section 10.8) converge presque sûrement vers la vraie valeur inconnue du paramètre. De plus, nous fournissons une preuve rigoureuse de l’inversibilité de la matrice des variances et des covariances de la statistique de test à partir de certaines propriétés d’algèbre linéaire. Le résultat s’applique aussi au cas où la moyenne est supposée connue et égale à zéro. Enfin, nous proposons une méthode de sélection de la dimension de la famille d’alternatives de type AIC, et nous étudions les propriétés asymptotiques de cette méthode. L’outil proposé ici est basé sur une famille spécifique de polynômes orthogonaux, à savoir les polynômes de Legendre. Dans un second temps, dans le cas vectoriel, nous proposons un test d’ajustement pour les modèles autorégressifs à moyenne mobile avec une paramétrisation structurée. La paramétrisation structurée permet de réduire le nombre élevé de paramètres dans ces modèles ou encore de tenir compte de certaines contraintes particulières. Ce projet inclut le cas standard d’absence de paramétrisation. Le test que nous proposons s’applique à une famille quelconque de fonctions orthogonales. Nous illustrons cela dans le cas particulier des polynômes de Legendre et d’Hermite. Dans le cas particulier des polynômes d’Hermite, nous montrons que le test obtenu est invariant aux transformations affines et qu’il est en fait une généralisation de nombreux tests existants dans la littérature. Ce projet peut être vu comme une généralisation du premier dans trois directions, notamment le passage de l’univarié au multivarié ; le choix d’une famille quelconque de fonctions orthogonales ; et enfin la possibilité de spécifier des relations ou des contraintes dans la formulation VARMA. Nous avons procédé dans chacun des projets à une étude de simulation afin d’évaluer le niveau et la puissance des tests proposés ainsi que de les comparer aux tests existants. De plus des applications aux données réelles sont fournies. Nous avons appliqué les tests à la prévision de la température moyenne annuelle du globe terrestre (univarié), ainsi qu’aux données relatives au marché du travail canadien (bivarié). Ces travaux ont été exposés à plusieurs congrès (voir par exemple Tagne, Duchesne et Lafaye de Micheaux (2013a, 2013b, 2014) pour plus de détails). Un article basé sur le premier projet est également soumis dans une revue avec comité de lecture (Voir Duchesne, Lafaye de Micheaux et Tagne (2016)).

Modeling Seasonal and Weather Impacts on Cycling Count

Thesis (Master's)--University of Washington, 2016-06

Statistical modeling and projections of lung cancer mortality in 4 industrialized countries

The purpose of this work was to model lung cancer mortality as a function of past exposure to tobacco and to forecast age-sex-specific lung cancer mortality rates. A 3-factor age-period-cohort (APC) model, in which the period variable is replaced by the product of average tar content and adult tobacco consumption per capita, was estimated for the US, UK, Canada and Australia by the maximum likelihood method. Age- and sex-specific tobacco consumption was estimated from historical data on smoking prevalence and total tobacco consumption. Lung cancer mortality was derived from vital registration records. Future tobacco consumption, tar content and the cohort parameter were projected by autoregressive moving average (ARIMA) estimation. The optimal exposure variable was found to be the product of average tar content and adult cigarette consumption per capita, lagged for 2530 years for both males and females in all 4 countries. The coefficient of the product of average tar content and tobacco consumption per capita differs by age and sex. In all models, there was a statistically significant difference in the coefficient of the period variable by sex. In all countries, male age-standardized lung cancer mortality rates peaked in the 1980s and declined thereafter. Female mortality rates are projected to peak in the first decade of this century. The multiplicative models of age, tobacco exposure and cohort fit the observed data between 1950 and 1999 reasonably well, and time-series models yield plausible past trends of relevant variables. Despite a significant reduction in tobacco consumption and average tar content of cigarettes sold over the past few decades, the effect on lung cancer mortality is affected by the time lag between exposure and established disease. As a result, the burden of lung cancer among females is only just reaching, or soon will reach, its peak but has been declining for I to 2 decades in men. Future sex differences in lung cancer mortality are likely to be greater in North America than Australia and the UK due to differences in exposure patterns between the sexes. (c) 2005 Wiley-Liss, Inc.

Forecasting exchange rates with linear and nonlinear models

In this paper, the exchange rate forecasting performance of neural network models are evaluated against the random walk, autoregressive moving average and generalised autoregressive conditional heteroskedasticity models. There are no guidelines available that can be used to choose the parameters of neural network models and therefore, the parameters are chosen according to what the researcher considers to be the best. Such an approach, however,implies that the risk of making bad decisions is extremely high, which could explain why in many studies, neural network models do not consistently perform better than their time series counterparts. In this paper, through extensive experimentation, the level of subjectivity in building neural network models is considerably reduced and therefore giving them a better chance of Forecasting exchange rates with linear and nonlinear models 415 performing well. The results show that in general, neural network models perform better than the traditionally used time series models in forecasting exchange rates.

The design of a real-time nowcasting system for localised weather

National meteorological offices are largely concerned with synoptic-scale forecasting where weather predictions are produced for a whole country for 24 hours ahead. In practice, many local organisations (such as emergency services, construction industries, forestry, farming, and sports) require only local short-term, bespoke, weather predictions and warnings. This thesis shows that the less-demanding requirements do not require exceptional computing power and can be met by a modern, desk-top system which monitors site-specific ground conditions (such as temperature, pressure, wind speed and direction, etc) augmented with above ground information from satellite images to produce `nowcasts'. The emphasis in this thesis has been towards the design of such a real-time system for nowcasting. Local site-specific conditions are monitored using a custom-built, stand alone, Motorola 6809 based sub-system. Above ground information is received from the METEOSAT 4 geo-stationary satellite using a sub-system based on a commercially available equipment. The information is ephemeral and must be captured in real-time. The real-time nowcasting system for localised weather handles the data as a transparent task using the limited capabilities of the PC system. Ground data produces a time series of measurements at a specific location which represents the past-to-present atmospheric conditions of the particular site from which much information can be extracted. The novel approach adopted in this thesis is one of constructing stochastic models based on the AutoRegressive Integrated Moving Average (ARIMA) technique. The satellite images contain features (such as cloud formations) which evolve dynamically and may be subject to movement, growth, distortion, bifurcation, superposition, or elimination between images. The process of extracting a weather feature, following its motion and predicting its future evolution involves algorithms for normalisation, partitioning, filtering, image enhancement, and correlation of multi-dimensional signals in different domains. To limit the processing requirements, the analysis in this thesis concentrates on an `area of interest'. By this rationale, only a small fraction of the total image needs to be processed, leading to a major saving in time. The thesis also proposes an extention to an existing manual cloud classification technique for its implementation in automatically classifying a cloud feature over the `area of interest' for nowcasting using the multi-dimensional signals.

Branching Processes with Immigration and Integer-valued Time Series

In this paper, we indicate how integer-valued autoregressive time series Ginar(d) of ordre d, d ≥ 1, are simple functionals of multitype branching processes with immigration. This allows the derivation of a simple criteria for the existence of a stationary distribution of the time series, thus proving and extending some results by Al-Osh and Alzaid [1], Du and Li [9] and Gauthier and Latour [11]. One can then transfer results on estimation in subcritical multitype branching processes to stationary Ginar(d) and get consistency and asymptotic normality for the corresponding estimators. The technique covers autoregressive moving average time series as well.

Functional Transfer Theorems for Maxima of Stationary Processes

2000 Mathematics Subject Classification: 60G70, 60F12, 60G10.

On Estimation and Testing for Pareto Tails

2010 Mathematics Subject Classification: 62F10, 62F12.

Estimation of computer waste quantities using forecasting techniques

Technology changes rapidly over years providing continuously more options for computer alternatives and making life easier for economic, intra-relation or any other transactions. However, the introduction of new technology “pushes” old Information and Communication Technology (ICT) products to non-use. E-waste is defined as the quantities of ICT products which are not in use and is bivariate function of the sold quantities, and the probability that specific computers quantity will be regarded as obsolete. In this paper, an e-waste generation model is presented, which is applied to the following regions: Western and Eastern Europe, Asia/Pacific, Japan/Australia/New Zealand, North and South America. Furthermore, cumulative computer sales were retrieved for selected countries of the regions so as to compute obsolete computer quantities. In order to provide robust results for the forecasted quantities, a selection of forecasting models, namely (i) Bass, (ii) Gompertz, (iii) Logistic, (iv) Trend model, (v) Level model, (vi) AutoRegressive Moving Average (ARMA), and (vii) Exponential Smoothing were applied, depicting for each country that model which would provide better results in terms of minimum error indices (Mean Absolute Error and Mean Square Error) for the in-sample estimation. As new technology does not diffuse in all the regions of the world with the same speed due to different socio-economic factors, the lifespan distribution, which provides the probability of a certain quantity of computers to be considered as obsolete, is not adequately modeled in the literature. The time horizon for the forecasted quantities is 2014-2030, while the results show a very sharp increase in the USA and United Kingdom, due to the fact of decreasing computer lifespan and increasing sales.

Off-line and on-line affective recognition of a computer user through a biosignal processing approach

Physiological signals, which are controlled by the autonomic nervous system (ANS), could be used to detect the affective state of computer users and therefore find applications in medicine and engineering. The Pupil Diameter (PD) seems to provide a strong indication of the affective state, as found by previous research, but it has not been investigated fully yet. ^ In this study, new approaches based on monitoring and processing the PD signal for off-line and on-line affective assessment ("relaxation" vs. "stress") are proposed. Wavelet denoising and Kalman filtering methods are first used to remove abrupt changes in the raw Pupil Diameter (PD) signal. Then three features (PDmean, PDmax and PDWalsh) are extracted from the preprocessed PD signal for the affective state classification. In order to select more relevant and reliable physiological data for further analysis, two types of data selection methods are applied, which are based on the paired t-test and subject self-evaluation, respectively. In addition, five different kinds of the classifiers are implemented on the selected data, which achieve average accuracies up to 86.43% and 87.20%, respectively. Finally, the receiver operating characteristic (ROC) curve is utilized to investigate the discriminating potential of each individual feature by evaluation of the area under the ROC curve, which reaches values above 0.90. ^ For the on-line affective assessment, a hard threshold is implemented first in order to remove the eye blinks from the PD signal and then a moving average window is utilized to obtain the representative value PDr for every one-second time interval of PD. There are three main steps for the on-line affective assessment algorithm, which are preparation, feature-based decision voting and affective determination. The final results show that the accuracies are 72.30% and 73.55% for the data subsets, which were respectively chosen using two types of data selection methods (paired t-test and subject self-evaluation). ^ In order to further analyze the efficiency of affective recognition through the PD signal, the Galvanic Skin Response (GSR) was also monitored and processed. The highest affective assessment classification rate obtained from GSR processing is only 63.57% (based on the off-line processing algorithm). The overall results confirm that the PD signal should be considered as one of the most powerful physiological signals to involve in future automated real-time affective recognition systems, especially for detecting the "relaxation" vs. "stress" states.^