Examinando por Autor "Sipols, Ana E."

Mostrando 1 - 5 de 5

Clustering and Forecasting Urban Bus Passenger Demand with a Combination of Time Series Models
(MDPI, 2022) Mariñas Collado, Irene; Sipols, Ana E.; Santos Martín, M. Teresa; Frutos Bernal, Elisa
The present paper focuses on the analysis of large data sets from public transport networks, more specifically, on how to predict urban bus passenger demand. A series of steps are proposed to ease the understanding of passenger demand. First, given the large number of stops in the bus network, these are divided into clusters and then different models are fitted for a representative of each of the clusters. The aim is to compare and combine the predictions associated with traditional methods, such as exponential smoothing or ARIMA, with machine learning methods, such as support vector machines or artificial neural networks. Moreover, support vector machine predictions are improved by incorporating explanatory variables with temporal structure and moving averages. Finally, through cointegration techniques, the results obtained for the representative of each group are extrapolated to the rest of the series within the same cluster. A case study in the city of Salamanca (Spain) is presented to illustrate the problem.
Distributed lags using elastic‑net regularization for market response models: focus on predictive and explanatory capacity
(palgrave macmillan, 2022) Martínez, Andres; Salafranca, Alfonso; Sipols, Ana E.; Simón de Blas, Clara; Van Hengel, Daniel
For many decades, considerable research has been conducted on Market Response models. Mostly without any attempts to validate the results in strictly predictive tasks and often ignoring if the methods comply with the underlying assumptions and conditions, like the method’s ability to outline the broadly accepted effects of advertising actions. This work presents an enhanced method for market response models consistent with the underlying assumptions of such. Our method is based on Distributed Lag Models with the novelty of introducing regularization in its estimation, a cross-validation framework, and hold-out testing, next to present an empirical manner of extracting its effects. This approach allows the construction of models in an exploratory and simple manner, unlocking the possibility of extracting the underlying effects and being suitable for large samples and many variables. Last, we conduct a practical example using real-world data, accompanied by an unprecedented set of empirical explainability assessments next to a high level of predictive capability in similar circumstances to how it would be used for decision-making in a corporate setup.
Non‑linear Cointegration Test, Based on Record Counting Statistic
(Springer, 2023-11-27) Atil, Lynda; Fellag, Hocine; Sipols, Ana E.; Santos Martín, M.T; Simón de Blas, Clara
Traditional tests fail to detect the presence of nonlinearities in series that are cointegrated, so in this paper a new procedure for cointegration tests is proposed by modifying the two-step Engle and Granger (EG) test (Engle and Granger in Econometrica 55:251–276, 1987), incorporating the RUR and the FB-RUR test of Aparicio et al. (J Time Ser Anal 27:545–576, 2006). The statistics of these non-parametric tests, which are constructed as functions of order statistics, endow the test with desirable properties such as invariance to non-linear transformations of the series and robustness to the presence of significant parameter shifts. As no prior estimation of the cointegrating parameter is required, the new tests lead to parameter-free asymptotic null distributions. Monte Carlo simulations are used to analyze the test properties and evaluate the power at different sample sizes. The robustness of the procedure is tested by performing a comparison of different tests of cointegration in real exchange rate relationships. These tests are able to find evidence of cointegration while standard cointegration tests fail to detect it.
Prediction of crop biophysical variables with panel data techniques and radar remote sensing imagery
(Elsevier, 2021) Blas, Clara Simón de; Valcarce-Diñeiro, Rubén; Sipols, Ana E.; Sánchez Martín, Nilda; Arias-Pérez, Benjamín; Santos-Martín, M. Teresa
Since the late 1970s, remote sensing techniques have been proven to be suitable for characterizing and monitoring plants and crops. In particular, synthetic aperture radar (SAR) missions contribute considerably to this prediction effort. However, the main issue when using SAR image series together with field observations is the scarcity of data due to the difficulty of acquiring field measurements. This research aimed to contribute to solving this problem with an alternative statistical model that can overcome the lack of a long, robust series of field-based ground truth observations. The main novelty of this research is the evaluation of the potential of a panel data approach to radar remote sensing imagery for predicting crop biophysical variables. For this purpose, RADARSAT-2 imagery was acquired over the study area in central Spain. Simultaneously, a field campaign was deployed to estimate crop parameters in the same area and to validate the results of the modelling. The analysis of the influence of the crop type on the incidence angle and the polarimetric parameters showed a strong influence of the co-polar channels (HH, VV), the entropy (H) and the coherence between the co-polar channels (γHHVV), with the differences being higher at 25°. The panel data analysis method demonstrated that good predictions, with R2 greater than 0.78, were achieved for all biophysical variables analysed in this study. Overall, this novel statistical approach with remote sensing data showed great applicability for the prediction of crop variables, even with a short series of observations.
Time Series of Quad-Pol C-Band Synthetic Aperture Radar for the Forecasting of Crop Biophysical Variables of Barley Fields Using Statistical Techniques
(MDPI, 2022-01-27) Sipols, Ana E.; Valcarce Diñeiro, Ruben; Santos Martín, María Teresa; Sánchez, Nilda; Simón de Blas, Clara
This paper aims to both fit and predict crop biophysical variables with a SAR image series by performing a factorial experiment and estimating time series models using a combination of forecasts. Two plots of barley grown under rainfed conditions in Spain were monitored during the growing cycle of 2015 (February to June). The dataset included nine field estimations of agronomic parameters, 20 RADARSAT-2 images, and daily weather records. Ten polarimetric observables were retrieved and integrated to derive the six agronomic and monitoring variables, including the height, biomass, fraction of vegetation cover, leaf area index, water content, and soil moisture. The statistical methods applied, namely double smoothing, ARIMAX, and robust regression, allowed the adjustment and modelling of these field variables. The model equations showed a positive contribution of meteorological variables and a strong temporal component in the crop’s development, as occurs in natural conditions. After combining different models, the results showed the best efficiency in terms of forecasting and the influence of several weather variables. The existence of a cointegration relationship between the data series of the same crop in different fields allows for adjusting and predicting the results in other fields with similar crops without re-modelling.