Systematic illiquidity, characteristic illiquidity, and stock returns: Timeseries analysis

The objective of this research is to investigate the relationship between illiquidity and stock prices on the Tunisian stock exchange. While previous researches tended to focus on one form of illiquidity to examine this relationship, our study unifies three forms of illiquidity at the same time. Indeed, we simultaneously consider illiquidity as systematic risk, as a characteristic of the market, and as a characteristic of the stock. The aggregate illiquidity of the market is the average of individual stock illiquidity. The illiquidity risk is the sensitivity of the stock price to illiquidity shocks. Shocks of market illiquidity are estimated by the innovations in the expected market illiquidity. Results show that investors on the Tunisian stock exchange do not require higher returns when they expect a rise of market illiquidity, whereas investors on U.S markets are compensated for higher expected market illiquidity. In addition, shocks of market illiquidity provoke a fall in stock prices of small caps, while large caps are not sensitive to market illiquidity shocks. This differs slightly from results based on U.S. data where illiquidity shocks reduce all stock prices but most notably those of small caps. Robustness tests validate our findings. Our results are consistent with previous studies which reported that the “zero-return” ratio predicts significantly the return-illiquidity relationship on emerging markets.

Geodetic investigation of triggered slip below Fiordland

<p>Secretary Island, at the head of Doubtful Sound in Fiordland, has been seismically active in past 30 years, with earthquakes larger than M w 6.5: the 1989 Doubtful Sound, 1993 Secretary Island, and 2003 Fiordland earthquakes. These events were approximately coincident with the 17° bend in the strike of the young, obliquely-converging, and steeply dipping Puysegur Subduction Zone. This section of the plate interface also has a history of triggered slip: the 1989 earthquake is inferred to have triggered the 1993 earthquake and, further north at George Sound, triggered afterslip was reported following the 2009 Dusky Sound earthquake. We have used L-band (23.6 cm-wavelength) Synthetic Aperture Radar (SAR) data from the ALOS1 and ALOS2 satellites, and C-band (5.5 cm-wavelength) SAR data from Sentinel 1A/B satellites, to test the hypothesis that triggered slip also occurred in the vicinity of Secretary Island following the 2007 George Sound, 2009 Dusky Sound and 2016 Kaikōura earthquakes. SAR images were aligned, interfered, filtered, and unwrapped using GMTSAR processing tools. Long-wavelength ionosphere noise was removed by inverting for the best-fitting linear plane, and we assumed a linear function of height to remove short-wavelength atmospheric noise. Small Baseline Subset (SBAS) timeseries analysis indicated a localised deformation signal centred on Secretary Island following the Dusky Sound earthquake. A re-analysis was undertaken of the co- and post-seismic deformation caused by the Dusky Sound earthquake so that any surface deformation centred on Secretary Island could be isolated. Campaign and continuous Global Positioning System (GPS) data were simultaneously inverted with co- and post-seismic interferograms using a statistical Bayesian modelling approach to determine the optimal Dusky Sound earthquake source parameters. Limitations arising from orbital drift, the frequency of SAR acquisitions and the observation geometry hindered our ability to constrain the timing, magnitude and location of reactivated slip from a source similar to the 2003 Secretary Island earthquake. Our findings indicate that slip was not triggered following either the 2007 George Sound earthquake or 2016 Kaikōura earthquake. However, we cannot rule out triggered slip near Secretary Island following the 2009 Dusky Sound earthquake. Any such slip likely occurred on an area of c. 350 km² (c. 15 km updip of the Secretary Island epicentre) with an average slip of 1–3 m, producing motion away from the satellite of c. 25 mm at Secretary Island.</p>

Geodetic investigation of triggered slip below Fiordland

<p>Secretary Island, at the head of Doubtful Sound in Fiordland, has been seismically active in past 30 years, with earthquakes larger than M w 6.5: the 1989 Doubtful Sound, 1993 Secretary Island, and 2003 Fiordland earthquakes. These events were approximately coincident with the 17° bend in the strike of the young, obliquely-converging, and steeply dipping Puysegur Subduction Zone. This section of the plate interface also has a history of triggered slip: the 1989 earthquake is inferred to have triggered the 1993 earthquake and, further north at George Sound, triggered afterslip was reported following the 2009 Dusky Sound earthquake. We have used L-band (23.6 cm-wavelength) Synthetic Aperture Radar (SAR) data from the ALOS1 and ALOS2 satellites, and C-band (5.5 cm-wavelength) SAR data from Sentinel 1A/B satellites, to test the hypothesis that triggered slip also occurred in the vicinity of Secretary Island following the 2007 George Sound, 2009 Dusky Sound and 2016 Kaikōura earthquakes. SAR images were aligned, interfered, filtered, and unwrapped using GMTSAR processing tools. Long-wavelength ionosphere noise was removed by inverting for the best-fitting linear plane, and we assumed a linear function of height to remove short-wavelength atmospheric noise. Small Baseline Subset (SBAS) timeseries analysis indicated a localised deformation signal centred on Secretary Island following the Dusky Sound earthquake. A re-analysis was undertaken of the co- and post-seismic deformation caused by the Dusky Sound earthquake so that any surface deformation centred on Secretary Island could be isolated. Campaign and continuous Global Positioning System (GPS) data were simultaneously inverted with co- and post-seismic interferograms using a statistical Bayesian modelling approach to determine the optimal Dusky Sound earthquake source parameters. Limitations arising from orbital drift, the frequency of SAR acquisitions and the observation geometry hindered our ability to constrain the timing, magnitude and location of reactivated slip from a source similar to the 2003 Secretary Island earthquake. Our findings indicate that slip was not triggered following either the 2007 George Sound earthquake or 2016 Kaikōura earthquake. However, we cannot rule out triggered slip near Secretary Island following the 2009 Dusky Sound earthquake. Any such slip likely occurred on an area of c. 350 km² (c. 15 km updip of the Secretary Island epicentre) with an average slip of 1–3 m, producing motion away from the satellite of c. 25 mm at Secretary Island.</p>

Exploring the Sensitivity of Recurrent Neural Network Models for Forecasting Land Cover Change

Recurrent Neural Networks (RNNs), including Long Short-Term Memory (LSTM) architectures, have obtained successful outcomes in timeseries analysis tasks. While RNNs demonstrated favourable performance for Land Cover (LC) change analyses, few studies have explored or quantified the geospatial data characteristics required to utilize this method. Likewise, many studies utilize overall measures of accuracy rather than metrics accounting for the slow or sparse changes of LC that are typically observed. Therefore, the main objective of this study is to evaluate the performance of LSTM models for forecasting LC changes by conducting a sensitivity analysis involving hypothetical and real-world datasets. The intent of this assessment is to explore the implications of varying temporal resolutions and LC classes. Additionally, changing these input data characteristics impacts the number of timesteps and LC change rates provided to the respective models. Kappa variants are selected to explore the capacity of LSTM models for forecasting transitions or persistence of LC. Results demonstrate the adverse effects of coarser temporal resolutions and high LC class cardinality on method performance, despite method optimization techniques applied. This study suggests various characteristics of geospatial datasets that should be present before considering LSTM methods for LC change forecasting.

An innovative, timeseries-analysis-based method to extract rainfall-streamflow events from continuous timeseries

&lt;p&gt;A widely accepted objective methodology to select individual rainfall-streamflow events is missing and this makes it difficult to synthesize findings from independent research initiatives. In fact, the selection of individual events is a fundamental step in many hydrological studies, but the importance and impact of the choices made at this stage are largely unrecognised.&lt;/p&gt;&lt;p&gt;The event selection methods found in the literature start by looking at either the rainfall timeseries or the streamflow timeseries. Moreover, most of the methodologies involve hydrograph separation, which is a highly uncertain step and can be performed using many different algorithms. Further increasing the subjectivity of the procedure, a wide range of ad hoc conditions are usually applied (e.g. peak-over-threshold, minimum duration of rainfall event, minimum duration of dry spell, minimum rainfall intensity&amp;#8230;).&lt;/p&gt;&lt;p&gt;For these reasons, we present a new methodology to extract rainfall-streamflow events which minimizes the conceptual hypotheses and user&amp;#8217;s choices, and bases the identification of the events mainly on the joint fluctuations of the two signals. The proposed methodology builds upon a timeseries analysis technique to estimate catchment response time, the Detrending Moving-average Cross-correlation Analysis-based method.&lt;/p&gt;&lt;p&gt;The proposed method has the advantage of looking simultaneously at the evolution in time of rainfall and streamflow timeseries, providing a more systemic detection of events. Moreover, the presented method can easily be adapted to extract events at different time resolutions (provided the resolution is fine enough to capture the delay between the rainfall and streamflow responses).&lt;/p&gt;&lt;p&gt;Properties of the events extracted with the proposed method are compared with the ones of the events extracted with the most traditional approach (based on hydrograph separation) to show strengths and weaknesses of the two techniques and suggest in which situations the proposed method can be most useful.&lt;/p&gt;

Innovation activity in agriculture and the issues of its assessment

The article describes the level of innovative activities of Russian agricultural businesses based on currently accepted indicators and justifies the limitations for their application in agriculture, taking into account its peculiar features. As an example the article considers a large agro-holding, which conducts active innovative activities aimed at the development of precision agriculture. The article shows the potential of correlation-regression analysis for the detection of the influence of particular technological innovations on the obtained result, as well as the potential of timeseries analysis for the assessment of major agriculturally-significant climatic risks.

Statistical Explorations and Univariate Timeseries Analysis on COVID-19 Datasets to Understand the Trend of Disease Spreading and Death

“Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2)”, the novel coronavirus, is responsible for the ongoing worldwide pandemic. “World Health Organization (WHO)” assigned an “International Classification of Diseases (ICD)” code—“COVID-19”-as the name of the new disease. Coronaviruses are generally transferred by people and many diverse species of animals, including birds and mammals such as cattle, camels, cats, and bats. Infrequently, the coronavirus can be transferred from animals to humans, and then propagate among people, such as with “Middle East Respiratory Syndrome (MERS-CoV)”, “Severe Acute Respiratory Syndrome (SARS-CoV)”, and now with this new virus, namely “SARS-CoV-2”, or human coronavirus. Its rapid spreading has sent billions of people into lockdown as health services struggle to cope up. The COVID-19 outbreak comes along with an exponential growth of new infections, as well as a growing death count. A major goal to limit the further exponential spreading is to slow down the transmission rate, which is denoted by a “spread factor (f)”, and we proposed an algorithm in this study for analyzing the same. This paper addresses the potential of data science to assess the risk factors correlated with COVID-19, after analyzing existing datasets available in “ourworldindata.org (Oxford University database)”, and newly simulated datasets, following the analysis of different univariate “Long Short Term Memory (LSTM)” models for forecasting new cases and resulting deaths. The result shows that vanilla, stacked, and bidirectional LSTM models outperformed multilayer LSTM models. Besides, we discuss the findings related to the statistical analysis on simulated datasets. For correlation analysis, we included features, such as external temperature, rainfall, sunshine, population, infected cases, death, country, population, area, and population density of the past three months—January, February, and March in 2020. For univariate timeseries forecasting using LSTM, we used datasets from 1 January 2020, to 22 April 2020.