Time Series Techniques Research Articles

Forecasting stock market returns using deep learning and time series techniques: a comparative and empirical study using technical indicators

Forecasting stock market returns using deep learning and time series techniques: a comparative and empirical study using technical indicators

Forecasting of Electricity Consumption by Seasonal Autoregressive Integrated Moving Average Model in Assam, India

Sustainable electricity consumption, which is about balancing economic growth, social development and environmental protection are the core principles of the Sustainable Development Goal (SDG). Accurate forecasting of electricity consumption is very important for attaining SDG - 7 which aims to ensure access to affordable, reliable, sustainable and modern energy for all. Through this paper, attempt is made to forecast monthly electricity consumption in Assam. For this purpose, one of the most widely used time series techniques, viz., Seasonal Autoregressive Integrated Moving Average (SARIMA) is applied by considering the time period April, 2013 to February, 2023 to study the seasonal influence on the electricity consumption in Assam. By applying Augmented Dickey Fuller test, it is observed that the data series becomes stationary at first order difference and the result of Canova Hansen test reveals that no seasonal differencing is required for our considering time period. SARIMA (1,1,1) (1,0,1)12 has been selected for forecasting purpose by following the results of Akaike Information Criterion. By analyzing the model statistics, residual ACF and PACF plots of the selected model, it is found that SARIMA (1,1,1) (1,0,1)12 can be effectively recommended for forecasting of monthly electricity consumption in Assam with Mean Absolute Percentage Error as 3.12%.

Functional Time Series Models for K-Step Prediction of Sugar Production: A Case Study of Eswatini, Kenya, and South Africa

Sugar, produced from cane or beet, is a vital energy source and a globally traded commodity. Most business happens in a futures exchange market where speculators, hedgers, and institutional consumers and producers, make decisions based on their understanding of the future supply and demand situation. Sugar is sometimes bought and sold before the cane or beet is planted. Hence, improving sugar production forecast accuracy is vital to maximize gains and enable effective planning. The study considered the annual sugar production total as a function of the monthly output rather than a scalar quantity and analyzed historical data using three functional time series techniques. In particular, the methods used for k-steps and dynamic updating prediction of sugar production include Local Polynomial Regression (LPR), Ridge Regression (RR), and Penalized Least Squares (PLS). The performance of the three models was compared to that of the automatic ARIMA method using the Mean Squared Error (MSE) and the Mean Absolute Percentage Error (MAPE) measures to establish the suitability of each technique in sugar production forecasting. The LPR, RR, and PLS outperformed the ARIMA method in all three cases. However, the prediction accuracy was lower for highly volatile datasets from countries that overly depend on rainfall for cane development.

Forecasting of coronavirus active cases by utilizing logistic growth model and fuzzy time series techniques

Coronavirus has long been considered a global epidemic. It caused the deaths of nearly 7.01 million individuals and caused an economic downturn. The number of verified coronavirus cases is increasing daily, putting the whole human race at danger and putting strain on medical experts to eradicate the disease as rapidly as possible. As a consequence, it is vital to predict the upcoming coronavirus positive patients in order to plan actions in the future. Furthermore, it has been discovered all across the globe that asymptomatic coronavirus patients play a significant part in the disease’s transmission. This prompted us to incorporate similar examples in order to accurately forecast trends. A typical strategy for analysing the rate of pandemic infection is to use time-series forecasting technique. This would assist us in developing better decision support systems. To anticipate COVID-19 active cases for a few countries, we recommended a hybrid model utilizing a fuzzy time series (FTS) model mixed with a non-linear growth model. The coronavirus positive case outbreak has been evaluated for Italy, Brazil, India, Germany, Pakistan, and Myanmar through June 5, 2020 in phase-1, and January 15, 2022 in phase-2, and forecasts active cases for the next 26 and 14 days respectively. The proposed framework fitting effect outperforms individual logistic growth and the fuzzy time series techniques, with R-scores of 0.9992 in phase-1 and 0.9784 in phase-2. The proposed model provided in this article may be utilised to comprehend a country’s epidemic pattern and assist the government in developing better effective interventions.

A Comparative Analysis between FFT, EMD, and EEMD for Epilepsy Detection

Context: Epilepsy is a neurological disease that affects more than 50 million people worldwide, causing recurrent seizures, with a significant impact on patients' quality of life due to abnormally synchronized neuronal activity. Method: This article discusses three methods used for signal analysis in patients diagnosed with epilepsy. Conventional signal decomposition methods, such as the fast Fourier transform, widely used in signal analysis based on time series techniques, have some issues when analyzing nonlinear and non-stationary signals, in addition to difficulties in detecting low-order frequencies. Results: To overcome these limitations, alternatives such as empirical mode decomposition and one of its variants, called ensemble empirical mode decomposition, have been developed. These techniques allow observing different oscillation modes through intrinsic mode functions and instantaneous frequencies. Conclusions: In this study, the results obtained through the aforementioned techniques were compared, revealing the impact of nonlinear methods on the reconstruction of brain activity.

Modelling and forecasting Masvingo Province maternal mortality using time series models

In light of the Zimbabwean government's efforts, particularly through the Ministry of Health and Child Care (MoHCC), to reduce maternal mortality rates and achieve the aim of Sustainable Development Goal (SDG) number three, target one (SDG 3.1), which aims to reduce the global maternal mortality rate to fewer than 70 deaths per 100,000 live births, maternal mortality rates in Zimbabwe continue to rise. Time series techniques were used to model and predict quarterly maternal mortality statistics for Masvingo Province from January 2014 to December 2021. The time plot analysis revealed significant fluctuations in mortality, with the highest rates of maternal deaths recorded in 2018. The application of the Box-Jenkins methodology identified the ARIMA(2, 1, 1) model as the most suitable for modeling and forecasting quarterly maternal deaths among the fitted models. The suitability of the model was validated by the Akaike Information Criterion (AIC), and its forecast accuracy was confirmed by the Mean Absolute Error (MAE) and the root mean square error (RMSE). Consequently, it was used to project future maternal deaths. The projected values indicate a slight increase in quarterly mortality rates during the period, but there appears to be a relatively stable trend with moderate fluctuations, suggesting that Zimbabwe has not met the targets of SDG 3.1. These results underscore the need to re-assess current intervention programs aimed at reducing maternal mortality. The findings of the study could guide the refinement of existing strategies and the implementation of innovative solutions to urgently address unacceptably high mortality rates. Using statistical models such as the one used in this study, the Ministry of Health and Child Care (MoHCC) can make informed decisions in the health sector and implement effective interventions to combat maternal mortality.

Climate Change Trend Using Descriptive Time Series Technique in Machine Learning: A Case of Jimma Zone, Southwestern Ethiopia

Understanding climate variability and monitoring time-series trends of temperature and rainfall is crucial for the sustainable development of our planet. This study utilized historical data from the Global Historical Climatology Network-Monthly (GHCN-M) provided by the National Centers for Environmental Information (NCEI) to analyze the temperature and rainfall data from 2015 to 2022. The analysis was conducted using Python 3.1.1 on Anaconda Jupyter Notebook and the package matplotlib 3.2.1 was used for data visualization. The results revealed a pattern of maximum rainfall between March to May for the years 2020, 2021, and 2022, while for the years 2017, 2018, and 2019, the maximum rainfall was recorded in October, December, and November. Additionally, the annual maximum rainfalls were recorded in the years 2020 and 2022, and the annual maximum temperatures for all study years were recorded in January, February, and March months. On the other hand, the annual minimum temperatures for all study years occurred in June, July, August, and September months. Similarly, annual average temperatures were recorded in January, February, and March months. This study emphasizes the importance of monitoring climate change and its impacts on our planet. By understanding climate variability and time-series trends, we can better prepare for the future and work towards a sustainable world.

Acumulación regulatoria y crecimiento económico

Nowadays, there is no consensus on the effects of regulation on economic growth. In this document, we analyze the relationship between regulation and economic growth for Colombia, using data on the accumulation of regulation and real GDP during the 1991 - 2019 period for 9 large productive sectors of the Colombian economy. Different models are estimated using time series and panel data techniques to present evidence at the economic sector level. The results show that, for some sectors, such as agriculture, forestry, fishing, mining, manufacturing, electricity, gas and water, commerce, restaurants and hotels, financial and business services and community, social and personal services, there exists an inverted U relationship between regulatory accumulation and GDP. These results offer important lessons for those responsible for the design and formulation of public policies on matters of regulatory improvement.

Multivariate data binning and examples generation to build a Diabetic Retinopathy classifier based on temporal clinical and analytical risk factors

In this paper, we explore the possibility of exploiting retrospective clinical data from Electronic Health Records (EHR) for classification tasks in chronic patients. The different intervals, short length and high class imbalance make it unfeasible to use traditional time series techniques. The first contribution of the paper is a preprocessing method to construct a multivariate time series dataset using EHR data, which infers missing data and regularizes the data frequency. The second contribution addresses class imbalance by using domain knowledge and existing short EHR series. We synthetically extrapolate patients’ data by using similar long time series and a fuzzy-based approach. The paper addresses the problem of detection of Diabetic Retinopathy (DR). Expert domain knowledge from ophthalmologists has been used in the proposed techniques to guide the processing of time series. The novelty in that case study consists in not using eye-fundus image analysis. Instead, the proposed methods are based solely on EHR data. Several multivariate multiclass time series classifiers are used to detect the four levels of DR severity from the pre-processed data sequences. Experiments prove the quality of the sequence preprocessing techniques proposed for EHR data. Results indicate that the TapNet classifier is the best one for DR grading. Despite being tested for DR detection, the proposed data preparation methods are applicable to other diseases with similar characteristics.

Modeling and forecasting the volatility of some industry development indicators in Ethiopia using multivariate GARCH models

Industry development indicators refer to a set of measures used to assess the performance and growth of industries. The main aim of this study was to assess the relationship between industry development indicators in Ethiopia using a multivariate GARCH model based on World Bank data from 1982 to 2021. A time series technique using annual data for the period 1982–2021 is utilized, and multivariate generalized autoregressive conditional heteroscedasticity was performed for volatility modeling. The results of the diagonal BEKK (1, 1)-GARCH model showed that there is strong evidence for a GARCH effect and the presence of a weaker ARCH effect, Equations show a statistically significant co-variation in shocks, which depends more on their lags than on past errors. Consequently, development indicator shocks are influenced by past information. The cross-volatility effects are higher than the own-volatility effects in Industry GDP, manufacturing GDP, and manufacturing exports. However, past volatility shocks in industry growth have less effect on cross-volatility than its own volatility shock. The implication of the study is that both domestic policymakers and development partners should support and motivate the growth of manufacturing sectors and manufacturing exports, since this is a necessary condition for promoting industry growth.

Sales Forecasting and Inventory Optimization at Big Mart using Time Series Technique: A Survey

Sales Forecasting and Inventory Optimization at Big Mart using Time Series Technique: A Survey

Series2vec: similarity-based self-supervised representation learning for time series classification

We argue that time series analysis is fundamentally different in nature to either vision or natural language processing with respect to the forms of meaningful self-supervised learning tasks that can be defined. Motivated by this insight, we introduce a novel approach called Series2Vec for self-supervised representation learning. Unlike the state-of-the-art methods in time series which rely on hand-crafted data augmentation, Series2Vec is trained by predicting the similarity between two series in both temporal and spectral domains through a self-supervised task. By leveraging the similarity prediction task, which has inherent meaning for a wide range of time series analysis tasks, Series2Vec eliminates the need for hand-crafted data augmentation. To further enforce the network to learn similar representations for similar time series, we propose a novel approach that applies order-invariant attention to each representation within the batch during training. Our evaluation of Series2Vec on nine large real-world datasets, along with the UCR/UEA archive, shows enhanced performance compared to current state-of-the-art self-supervised techniques for time series. Additionally, our extensive experiments show that Series2Vec performs comparably with fully supervised training and offers high efficiency in datasets with limited-labeled data. Finally, we show that the fusion of Series2Vec with other representation learning models leads to enhanced performance for time series classification. Code and models are open-source at https://github.com/Navidfoumani/Series2Vec

Modeling the Accessibility to Electricity in Nigeria using Time Series Technique

Modeling the Accessibility to Electricity in Nigeria using Time Series Technique

Advancing the Limits of InSAR to Detect Crustal Displacement from Low-Magnitude Earthquakes through Deep Learning

Detecting surface deformation associated with low-magnitude (Mw≤5) seismicity using interferometric synthetic aperture radar (InSAR) is challenging due to the subtlety of the signal and the often challenging imaging environments. However, low-magnitude earthquakes are potential precursors to larger seismic events, and thus characterizing the crustal displacement associated with them is crucial for regional seismic hazard assessment. We combine InSAR time-series techniques with a Deep Learning (DL) autoencoder denoiser to detect the magnitude and extent of crustal deformation from the Mw=3.4 Gallina, New Mexico earthquake that occurred on 30 July 2020. Although InSAR alone cannot detect event-related deformation from such a low-magnitude seismic event, application of the DL method reveals maximum displacements as small as (±2.5 mm) in the vicinity of both the fault and earthquake epicenter without prior knowledge of the fault system. This finding improves small-scale displacement discernment with InSAR by an order of magnitude relative to previous studies. We additionally estimate best-fitting fault parameters associated with the observed deformation. The application of the DL technique unlocks the potential for low-magnitude earthquake studies, providing new insights into local fault geometries and potential risks from higher-magnitude earthquakes. This technique also permits low-magnitude event monitoring in areas where seismic networks are sparse, allowing for the possibility of global fault deformation monitoring.

Forecasting material quantity using machine learning and times series techniques

Abstract The current research is dedicated to harnessing cutting-edge technologies within the paradigm of Industry 5.0. The objective is to capitalize on advancements in Machine and Deep Learning techniques. This research endeavors to construct robust predictive models, utilizing historical data, for precise real-time predictions in estimating material quantities within a cement workshop. Machine Learning regressors evaluated based on several metrics, SVR (R-squared 0.9739, MAE 0.0403), Random Forest (R-squared 0.9990, MAE 0.0026), MLP (R-squared 0.9890, MAE 0.0255), Gradient Boosting (R-squared 0.9989, MAE 0.0042). The time series models LSTM and GRU yielded R-squared 0.9978, MAE 0.0100, and R-squared 0.9980, MAE 0.0099, respectively. The ultimate outcomes include improved and efficient production, optimization of production processes, streamlined operations, reduced downtime, mitigation of potential disruptions, and the facilitation of the factory’s evolution towards intelligent manufacturing processes embedded within the framework of Industry 5.0. These achievements underscore the potential impact of leveraging advanced machine learning techniques for enhancing the operational dynamics and overall efficiency of manufacturing facilities

THE ROLE OF FINANCIAL DEVELOPMENT ON THE SHARE OF RENEWABLE ENERGY IN JAPAN: A COMPREHENSIVE TIME-SERIES ANALYSIS

The purpose of this study is to analyze the effect of financial development (FD) on the share of renewable energy (RE) usage in Japan. The existing theoretical literature and empirical analyses covering different country cases reveal that FD might have positive, negative, or insignificant effect on RE use. Since there is no empirical investigation of the issue for Japan, the study aims to contribute to the literature. To that end, we used several time-series techniques to detect the association between RE usage and FD in Japan over the 1970–2020 period. The results obtained from the ARDL, Hatemi-J, Maki, Tsong et al. and NARDL cointegration tests showed that there is a significant cointegrated relationship between the share of RE use, FD, GDP per capita and trade openness. As for the long-run coefficients obtained from the ARDL, FMOLS, CCR and NARDL estimators revealed that increases in FD and trade openness raise the share of RE usage while increases in GDP per capita reduce it. Briefly, for the Japan case, we may suggest that improving the financial market structure of the country will bear fruit in terms of the share of cleaner and sustainable energy usage.

Impact of oil price, economic globalization, and inflation on economic output: Evidence from Latin American oil-producing countries using the quantile-on-quantile approach

This research examines the impact of oil prices, economic globalization, and inflation on the economic output in oil-producing countries in Latin America. We employ advanced time series techniques to achieve precision in findings and reliability of policy inferences. We use cointegration techniques with and without structural breaks because oil prices are highly volatile, and the economies analyzed are unstable in the face of political and social changes. Additionally, we performed a sensitivity analysis using standard quantile and the newly developed quantile-on-quantile regressions to measure the impacts of the covariates on real per capita output. Finally, we use causality Fourier techniques to identify the direction of causal relationships between the series. We found that the oil price, economic globalization, inflation, and output have a long-term equilibrium relationship in the presence of structural breaks. Likewise, quantile models show that the impacts of inflation, economic globalization, and oil prices on economic output are extensively non-linear across the quantile distribution. We found a negative connection between inflation and economic output in the surveyed territories. We propose that those responsible for energy policy in the countries studied should rethink crude oil policies to maximize the benefits of oil exploitation under stagnant globalization and growing inflation.

Food Price Inflation in the United States as a Complex Dynamic Economic System

Abstract The issue of volatile food prices is a consistent problem for American consumers, as rising prices make it challenging to afford nutritious food that meets dietary standards. Various complex factors influence this price volatility, including economic conditions, weather patterns, global trade, energy prices, and more. Notably, the impact of food price increases is not equal for everyone. Low-income individuals and those in rural areas are disproportionately affected. A comprehensive understanding of the driving factors is essential to tackle this issue effectively. We employ advanced time-series techniques such as Vector Error Correction Models (VECM) and modern causal inference methods such as probabilistic graphical models implemented via machine learning and artificial intelligence approaches on monthly data from 2000 to 2021 to investigate the U.S. food price inflation issue. These methods help unravel the intricate dynamics among key variables driving food price inflation. The study aims to achieve several objectives. It intends to (1) clarify how factors influencing food price inflation in the U.S. change over time using VECM models, (2) establish causal relationships among interconnected variables to develop probabilistic graphical models using innovative search algorithms, and (3) create and validate forecasts related to U.S. food price inflation. The end goal is to provide actionable insights for policy design. Results show that food price inflation is heavily tied to commodity pricing and pricing for medical services. Additionally, historical decompositions for COVID-19 show ties between food price inflation and energy inflation.

A Comparative Study between Time Series and Machine Learning Technique to Predict Dengue Fever in Dhaka City

The dengue virus is the most dangerous one that mosquitoes may spread to people. Despite attempts by the government, dengue outbreaks are becoming increasingly common in Bangladesh. Interventions in public health rely heavily on KAP (knowledge, attitude, and practice) studies. The primary goal of this research is to forecast the occurrence of dengue disease in the city of Dhaka using methods from machine learning and time series analysis and then to compare the models in order to find the one with the lowest MAPE. From January 2016 through July 2021, monthly data were retrieved for this study from WHO and the Directorate General of Health and Services (DGHS). According to the findings of this research, neural networks outperform time series analysis when it comes to making predictions. The best-fitted neural network (NN) model was found in model 04 with 05 hidden layers which produced the minimum error model with the value of error 0.003032557, and the values of RMSE and MAPE are 7.588889e − 06 and 1.15273, respectively, for the prediction of the dengue fever in Dhaka city. In contrast, the original dengue data in the time series analysis is not stationary. Take the difference and run the unit root test by the augmented Dickey–Fuller (ADF) test to make it stationary. The dengue data series is stationary at the first-order difference, as evidenced by the ACF and PACF, which show no noticeable spike in the first-order difference. The ARIMA (6, 1, 1) model with the lowest AIC = −251.8, RMSE = 0.0310797, and MAPE = 15.2892 is the best choice model for predicting the dengue death rate. Therefore, from these two models, the NN model gives better prediction performance with the lowest value of MAPE. So, the neural network gives better prediction performance than time series analysis. The NN model forecasted 12-month death rates of dengue fever that suggest the death rate in dengue fever falling month by month. This study is more innovative than any other research because this research approach is different from any other research approach. The model selection criteria are based on the most effective performance metrics MAPE, indicating the lowest error and better prediction performance. Therefore, from this research, the author suggests machine learning gives better prediction performance than time series analysis for any other prediction performance.

What is the long-run relationship between military expenditures, foreign trade and ecological footprint? Evidence from method of Maki cointegration test

AbstractThe US leads the global economy regarding foreign trade volume and military expenditures. Increases in military expenditures and foreign trade can substantially affect environmental issues. This study's preliminary research objective is to explore the interaction of foreign trade and military expenditures with ecological concerns in the long-term of the US economy using current time series techniques. Ecological footprint, military expenditures, exports, imports, urbanization, and agricultural area variables are used in the analysis with annual data for the US economy for the period 1970-2018. A relationship in long-term between the variables is revealed by the Maki (2012) cointegration test results, considering structural break. As indicated by the findings obtained from the coefficient estimation, an increase in military expenditures, exports, and agricultural area decreases the ecological footprint, while an increase in imports increases it. In addition, the results of the time-varying causality test show that there are periodic causality relationships between the variables. To sum up, there is a negative relationship between military expenditures, exports, and agricultural area, and ecological footprint, and a positive relationship between imports and ecological footprint.