Modelling non-stationary ‘Big Data’
نویسندگان
چکیده
‘Fat big data’ characterise data sets that contain many more variables than observations. We discuss the use of both principal components analysis and equilibrium correction models to identify cointegrating relations handle stochastic trends in non-stationary fat data. However, most time series are wide-sense non-stationary—induced by joint occurrence distributional shifts—so we also latter saturation estimation. Seeking substantive relationships when there vast numbers potentially spurious connections cannot be achieved merely choosing best-fitting equation or trying hundreds empirical fits selecting a preferred one, perhaps contradicted others go unreported. Conversely, useful if they help ensure generation process is nested postulated model, increase power specification mis-specification tests without raising chances adventitious significance. model monthly UK unemployment rate, using macroeconomic Google Trends data, searching across 3000 explanatory variables, yet parsimonious, statistically valid, theoretically interpretable specification.
منابع مشابه
Data Computational Modelling of Multivariable Non-Stationary Noisy Linear Systems by MOESP_AOKI_VAR Algorithm
The main objective of this work is to develop a recursive algorithm for identification in the state-space of linear stochastic discrete multivariable non-stationary system; a computational process called MOESP_AOKI_VAR is proposed and implemented to achieve this. The proposed algorithm is based on the subspace methods: Multivariable Output-Error State Space (MOESP), used for computational model...
متن کاملWave data assimilation using non-stationary kriging
We present a fast sequential assimilation technique for conditioning wave simulations using buoy data. It is based on non-stationary kriging that uses a separable space-time variogram. This variogram is estimated from the history of simulations, as a collection of maps of variances of increments between the logarithms of simulated and observed wave heights. The method finally provides a set of ...
متن کاملData alignment with non-stationary shaping filters
Cross-correlation provides a method of calculating a static shift between two datasets. By cross-correlating patches of data, I can calculate a “warp function” that dynamically aligns the two datasets. By exploiting the link between cross-correlation and shaping filters, I calculate warp functions in a more general way, leveraging the full machinery of geophysical estimation. I compare warp fun...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: International Journal of Forecasting
سال: 2021
ISSN: ['1872-8200', '0169-2070']
DOI: https://doi.org/10.1016/j.ijforecast.2020.08.002