scholarly journals A comparison of parameter selection measures for sensor learning from financial news events

2018 ◽  
Author(s):  
Alex S. Farias ◽  
Solange O. Rezende ◽  
Ricardo M. Marcacini
Keyword(s):  
Time Series ◽  
Digital Media ◽  
Recurrent Events ◽  
Parameter Selection ◽  
General Idea ◽  
External Information ◽  
Promising Alternative ◽  
Financial News ◽  
Selection Measures ◽  
News Events

The popularization of web platforms promoted a significant increase in the publication of financial news and reports in digital media. In this sense, a multidisciplinary research area called “learning to sense” (or sensor learning) has received attention recently. Unlike traditional machine learning methods, in sensor learning there is an interest in obtaining a time series that indicates the activity of a particular topic over time. A sensor is represented by a set of parameters learned from a historical news events dataset. The sensor generates time series as news events are processed and these time series are used in decision support systems. This paper presents an overview of sensor learning for financial news. We compared six parameter selection measures for sensor learning, with the differential of considering an unsupervised scenario. The general idea is to use the concept of k-recurrent events, i.e, news events that are similar and occur together in different periods of up-trends and down-trends of a financial time series. Thus, if a specific event (extracted from news) occurred at least k times in the past always associated with up-trends, then such news is labeled as positive news. Analogously, it can be labeled as negative. The experimental results from real data provided evidence that the approach investigated in this work is a promising alternative for sensor learning from financial news events, especially in contexts where there are no domain experts or external information to label a training set.

scholarly journals EXTERNAL INFORMATION AND MONETARY POLICY TRANSMISSION IN NEW EU MEMBER STATES: RESULTS FROM FAVAR MODELS

2016 ◽  
Vol 21 (2) ◽  
pp. 311-335 ◽  
Author(s):  
Zlatina Balabanova ◽  
Ralf Brüggemann
Keyword(s):  
Time Series ◽  
Monetary Policy ◽  
External Information ◽  
Response Patterns ◽  
Member States ◽  
Vector Autoregressive ◽  
Eu Member States ◽  
Policy Shocks ◽  
Economic Developments ◽  
New Eu Member States

We investigate the effects of monetary policy shocks in the new European Union (EU) member states the Czech Republic, Hungary, Poland, and Slovakia. In contrast to existing studies, we explicitly account for external developments in European Monetary Union (EMU) countries and in other acceding countries. We do so by using factor-augmented vector autoregressive models that employ information from nonstationary factor time series. One set of VAR models includes factors obtained from a large cross section of time series from EMU countries, whereas another set includes factors obtained from other acceding countries. We find that including EMU factors does change impulse response patterns in some but not all acceding countries. In contrast, including factors from other acceding countries leads to substantial changes in impulse responses and to economically more plausible results. Overall, our analysis highlights that taking external economic developments properly into account is crucial for the analysis of monetary policy in the new EU member states.

scholarly journals Estimation of Irregular Wave Runup on Intermediate and Reflective Beaches Using a Phase-Resolving Numerical Model

2020 ◽  
Vol 8 (12) ◽  
pp. 993
Author(s):  
Jonas Pinault ◽  
Denis Morichon ◽  
Volker Roeber
Keyword(s):  
Time Series ◽  
Best Practices ◽  
Numerical Models ◽  
Laboratory Data ◽  
Irregular Waves ◽  
Wave Transformation ◽  
Model Parameters ◽  
Wave Runup ◽  
Data Set ◽  
Promising Alternative

Accurate wave runup estimations are of great interest for coastal risk assessment and engineering design. Phase-resolving depth-integrated numerical models offer a promising alternative to commonly used empirical formulae at relatively low computational cost. Several operational models are currently freely available and have been extensively used in recent years for the computation of nearshore wave transformations and runup. However, recommendations for best practices on how to correctly utilize these models in computations of runup processes are still sparse. In this work, the Boussinesq-type model BOSZ is applied to calculate runup from irregular waves on intermediate and reflective beaches. The results are compared to an extensive laboratory data set of LiDAR measurements from wave transformation and shoreline elevation oscillations. The physical processes within the surf and swash zones such as the transfer from gravity to infragravity energy and dissipation are accurately accounted for. In addition, time series of the shoreline oscillations are well captured by the model. Comparisons of statistical values such as R2% show relative errors of less than 6%. The sensitivity of the results to various model parameters is investigated to allow for recommendations of best practices for modeling runup with phase-resolving depth-integrated models. While the breaking index is not found to be a key parameter for the examined cases, the grid size and the threshold depth, at which the runup is computed, are found to have significant influence on the results. The use of a time series, which includes both amplitude and phase information, is required for an accurate modeling of swash processes, as shown by computations with different sets of random waves, displaying a high variability and decreasing the agreement between the experiment and the model results substantially. The infragravity swash SIG is found to be sensitive to the initial phase distribution, likely because it is related to the short wave envelope.

Recurrence Quantification Techniques of vegetation time-series indices in semiarid grasslands

2020 ◽  
Author(s):  
Andres Almeida-Ñauñay ◽  
Rosa M. Benito ◽  
Miguel Quemada ◽  
Juan Carlos Losada ◽  
Ana Maria Tarquis
Keyword(s):  
Time Series ◽  
Recurrent Events ◽  
Vegetation Index ◽  
Near Infrared ◽  
Normalized Difference Vegetation Index ◽  
Dynamic Properties ◽  
Distinct Pattern ◽  
Climatic Variables ◽  
Recurrence Quantification ◽  
Semi Arid

<p>Grassland ecosystems are extremely complex and set up intricate structures, whose characteristics and dynamic properties are greatly influenced by climate and meteorological patterns. Climate change and global warming are factors that could impact negatively in the quality and productivity of these ecosystems.</p><p>Remote sensing techniques have been demonstrated as a powerful tool for monitoring extensive areas. In this study, two semi-arid grassland plots were selected in the centre of Spain. This region is characterized by low precipitation and moderate productivity per unit. Through scientific research, spectral vegetation indices (VIs) have been developed to characterize vegetation cover. The most common VI is the Normalized Difference Vegetation Index (NDVI). However, in vegetation scarcity conditions, bare soil reflectance is increased, and the feasibility of NDVI is reduced. This study aims to perform a method to compare soil and agro-climatic variables effect on vegetation time-series indices.</p><p>The construction of the time series was based on multispectral images of MODIS TERRA (MOD09A1.006) product acquired from 2002 till 2018. Three pixels with a temporal resolution of 8 days and a spatial resolution of 500 x 500 m were chosen in each area. To estimate and analyse VIs series, Red (620-670 nm) and Near Infrared (841-876 nm) channels were extracted and filtered by the quality of pixel. All spectral bands showed statistically significant differences confirming that both areas presented different soil properties. Moreover, average annual precipitation was different in each area of study.</p><p>NDVI calculation is only based on NIR and RED bands. To improve the estimation of vegetation in semi-arid areas, several indices have been developed to minimize the soil effect. Each one of them incorporates soil influence in a different way, i.e., Soil Adjusted Vegetation Index (SAVI) adds a constant soil adjustment factor (L), whereas, MSAVI, incorporate an L variable and dependant on soil characteristics.</p><p>Recurrence plots (RP) and recurrence quantification analysis (RQA) were computed to characterize the influence of agro-climatic variables in vegetation index dynamics. Characterization was based on various RQA measures, such as Determinism (DET), average diagonal length (LT) or entropy (ENT).</p><p>Our results showed different RPs depending on the area, VI utilized and precipitation. MSAVI patterns were further distinct, meanwhile, NDVI showed a noisy pattern. LT values in MSAVI were higher than in SAVI implying that MSAVI recurrent events are much longer than SAVI. Simultaneously, LT and DET values in ZSO, with a higher rain, were above ZEA values in MSAVI.</p><p>This indicates that incorporating more detailed information of soil and precipitation reinforce vegetation index estimation and allow to obtain a more distinct pattern of the time series. Therefore, in arid-semiarid grasslands, they should be considered.</p><p><strong>ACKNOWLEDGEMENTS</strong></p><p>The authors acknowledge support from Project No. PGC2018-093854-B-I00 of the Spanish <em>Ministerio de Ciencia Innovación y Universidades</em> of Spain and the funding from the Comunidad de Madrid (Spain) and Structural Funds 2014-2020 512 (ERDF and ESF), through project AGRISOST-CM S2018/BAA-4330, are highly appreciated.</p>

scholarly journals Event Study: Advanced Machine Learning and Statistical Technique for Analyzing Sustainability in Banking Stocks

Mathematics ◽  
2021 ◽  
Vol 9 (24) ◽  
pp. 3319
Author(s):  
Varun Dogra ◽  
Aman Singh ◽  
Sahil Verma ◽  
Abdullah Alharbi ◽  
Wael Alosaimi
Keyword(s):  
Machine Learning ◽  
Stock Prices ◽  
Research Work ◽  
Abnormal Returns ◽  
Second Phase ◽  
Financial News ◽  
News Events ◽  
News Event ◽  
The Impact

Machine learning has grown in popularity in recent years as a method for evaluating financial text data, with promising results in stock price projection from financial news. Various research has looked at the relationship between news events and stock prices, but there is little evidence on how different sentiments (negative, neutral, and positive) of such events impact the performance of stocks or indices in comparison to benchmark indices. The goal of this paper is to analyze how a specific banking news event (such as a fraud or a bank merger) and other co-related news events (such as government policies or national elections), as well as the framing of both the news event and news-event sentiment, impair the formation of the respective bank’s stock and the banking index, i.e., Bank Nifty, in Indian stock markets over time. The task is achieved through three phases. In the first phase, we extract the banking and other co-related news events from the pool of financial news. The news events are further categorized into negative, positive, and neutral sentiments in the second phase. This study covers the third phase of our research work, where we analyze the impact of news events concerning sentiments or linguistics in the price movement of the respective bank’s stock, identified or recognized from these news events, against benchmark index Bank Nifty and the banking index against benchmark index Nifty50 for the short to long term. For the short term, we analyzed the movement of banking stock or index to benchmark index in terms of CARs (cumulative abnormal returns) surrounding the publication day (termed as D) of the news event in the event windows of (−1,D), (D,1), (−1,1), (D,5), (−5,−1), and (−5,5). For the long term, we analyzed the movement of banking stock or index to benchmark index in the event windows of (D,30), (−30,−1), (−30,30), (D,60), (−60,−1), and (−60,60). We explore the deep learning model, bidirectional encoder representations from transformers, and statistical method CAPM for this research.

How consumers' personalities, lifestyles and demographics predict SVOD genre and SVOD platform consumption

2021 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Anthony Palomba
Keyword(s):  
Digital Media ◽  
Structural Equation ◽  
Video On Demand ◽  
Equation Model ◽  
Consumer Information ◽  
External Information ◽  
Media Consumption ◽  
Content Type ◽  
Consumer Demographics ◽  
Consumer Data

PurposeStreaming video on demand (SVOD) services are comprised of digital media content creation and content distribution that provide a vast array of genre content playable on an assortment of different technology platforms. Additionally, these digital services are equipped to collect data and information on consumers. However, these services do not capture extensive consumer demographics, lifestyles or personalities information.Design/methodology/approachTo resolve this discrepancy, collecting external information such as complete demographics, personalities and lifestyles of consumers can be useful in advancing SVOD consumer behavior knowledge. This study examined how consumer demographics, lifestyles and personalities may predict SVOD genre consumption and SVOD platform consumption. A survey was executed and disseminated to collect consumer information across these dimensions. Multiple linear regressions and a structural equation model were formed to explicate variance.FindingsConsumer demographics, lifestyles and personalities’ information do predict SVOD genre consumption and SVOD platform consumption.Originality/valueMedia selection and trait theory have not been applied to understanding unexplained variance behind consumer media consumption, and are often used to predict media consumption variance among consumers. These findings illustrate that, while digital consumer touchpoints are necessary to collect and analyze, marketers should not lose sight of easily-obtainable consumer data, much of which dictates consumption choices.

Financial Time Series Prediction Based on BP Neural Network

2014 ◽  
Vol 631-632 ◽  
pp. 31-34 ◽  
Author(s):  
Jia Jia
Keyword(s):  
Neural Network ◽  
Time Series ◽  
Bp Neural Network ◽  
Financial Time Series ◽  
Risk Function ◽  
Risk Minimization ◽  
It Use ◽  
Promising Alternative ◽  
Minimization Principle ◽  
Financial Time

BP neural network is promising methods for the prediction of financial time series because it use a risk function consisting of the empirical error and a regularized term which is derived from the structural risk minimization principle. This study applies BP neural network to predicting the stock price index. In addition, this study examines the feasibility of applying BP neural network in financial forecasting. The experimental results show that BP neural network provides a promising alternative to stock market prediction.

scholarly journals A Novel Time Series Prediction Approach Based on a Hybridization of Least Squares Support Vector Regression and Swarm Intelligence

2014 ◽  
Vol 2014 ◽  
pp. 1-8 ◽  
Author(s):  
Nhat-Duc Hoang ◽  
Anh-Duc Pham ◽  
Minh-Tu Cao
Keyword(s):  
Time Series ◽  
Least Squares ◽  
Support Vector Regression ◽  
Time Series Data ◽  
Time Series Prediction ◽  
Hybrid Approach ◽  
Series Data ◽  
Support Vector ◽  
Promising Alternative ◽  
Prediction Approach

This research aims at establishing a novel hybrid artificial intelligence (AI) approach, named as firefly-tuned least squares support vector regression for time series prediction(FLSVRTSP). The proposed model utilizes the least squares support vector regression (LS-SVR) as a supervised learning technique to generalize the mapping function between input and output of time series data. In order to optimize the LS-SVR’s tuning parameters, theFLSVRTSPincorporates the firefly algorithm (FA) as the search engine. Consequently, the newly construction model can learn from historical data and carry out prediction autonomously without any prior knowledge in parameter setting. Experimental results and comparison have demonstrated that theFLSVRTSPhas achieved a significant improvement in forecasting accuracy when predicting both artificial and real-world time series data. Hence, the proposed hybrid approach is a promising alternative for assisting decision-makers to better cope with time series prediction.

Sign in / Sign up

Export Citation Format

Share Document