Evaluating Temporal Bias in Time Series Event Detection Methods

Escobar, Luciana; Salles, Rebecca; Lima, Janio; Gea, Cristiane; Baroni, Lais; Ziviani, Artur; Pires, Paulo F.; Delicato, Flávia Coimbra; Coutinho, Rafaelli; Assis, Laura Silva de; Ogasawara, Eduardo

doi:10.5753/jidm.2021.1968

Cited by 3 publications

(1 citation statement)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To show the usefulness of this strategy at advanced forecasting scales, a short time prediction model has been used to forecast the amount of new COVID-19 symptoms in a specific state in the U.S. and for a set of counties within the state [9]. Newly used models include Ace-Mod Australian Census-based Epidemic Model (Ace-Mod), SIDR, Fuzzy Clustering, SEIR, and DASS-21 for monitoring the growth and predicting COVID-19 [10]- [12]. In particular, [12] have used the Richards growth model, a sub-epidemic wave model, and a generalized logistic growth model to produce predictions for China.…”

Section: Review Of the Literaturementioning

confidence: 99%

Coronavirus Disease 2019 (COVID-19): Prediction Strategy Using Sequential Deep Learning Model

Surja

Iqbal

Faruk

2023

EJMED

View full text Add to dashboard Cite

Since the globe has faced extreme difficulties with COVID-19, Artificial Intelligence appeared to help to cope with this epidemic in an innumerable number of ways. Motivated by this, in this article, a robust prediction model called COVID-SDL has been proposed using Sequential Deep Learning (SDL) for predicting the total positive cases per day. In order to evaluate the performance of COVID-SDL, data samples used in the model have been collected from Italy’s COVID-19 situation reports. Besides this, the dataset has gone through the processes of cleaning, filtering, formatting and visualization. COVID-SDL utilizes the correlation information among the features that have strengthened the prediction capability. Also, the exploratory survey showed that 5 most salient features (Home Confinement, Deaths, Recovered, Current Positive Cases and Tests Performed) results better which is obtained from the mentioned dataset primarily composed of 17 features. In addition, to assist the prediction ability of COVID-SDL, ReLu (Rectified Linear Unit) activation function has been used which enhanced the robustness of the model. With a view to making the predictions highly accurate, Adam optimizer has been adopted which works by reducing the cost function and making further updates of the weights. Moreover, COVID-SDL has successfully obtained accuracy parameters such as MAE- 0.00037316, MSE- 0.00000018, RMSE- 0.00043476 and R2 Score- 0.99999 with providing the best fit curve of predicted data which covers 99.999% of the actual data. Furthermore, to prove the robustness of the COVID-SDL, a comparative test among the adaptive and non-adaptive optimizers has also been performed.

show abstract

Section: Review Of the Literaturementioning

confidence: 99%

Coronavirus Disease 2019 (COVID-19): Prediction Strategy Using Sequential Deep Learning Model

Surja

Iqbal

Faruk

2023

EJMED

View full text Add to dashboard Cite

show abstract

Multi-Scale Event Detection in Financial Time Series

de Salles,

Gea,

Mello

et al. 2024

Comput Econ

View full text Add to dashboard Cite

Determining an effective short term COVID-19 prediction model in ASEAN countries

Sharif¹,

Hasan

Rahman

2022

Sci Rep

View full text Add to dashboard Cite

The challenge of accurately short-term forecasting demand is due to model selection and the nature of data trends. In this study, the prediction model was determined based on data patterns (trend data without seasonality) and the accuracy of prediction measurement. The cumulative number of COVID-19 affected people in some ASEAN countries had been collected from the Worldometers database. Three models [Holt’s method, Wright’s modified Holt’s method, and unreplicated linear functional relationship model (ULFR)] had been utilized to identify an efficient model for short-time prediction. Moreover, different smoothing parameters had been tested to find the best combination of the smoothing parameter. Nevertheless, using the day-to-day reported cumulative case data and 3-days and 7-days in advance forecasts of cumulative data. As there was no missing data, Holt’s method and Wright’s modified Holt’s method showed the same result. The text-only result corresponds to the consequences of the models discussed here, where the smoothing parameters (SP) were roughly estimated as a function of forecasting the number of affected people due to COVID-19. Additionally, the different combinations of SP showed diverse, accurate prediction results depending on data volume. Only 1-day forecasting illustrated the most efficient prediction days (1 day, 3 days, 7 days), which was validated by the Nash–Sutcliffe efficiency (NSE) model. The study also validated that ULFR was an efficient forecasting model for the efficient model identifying. Moreover, as a substitute for the traditional R-squared, the study applied NSE and R-squared (ULFR) for model selection. Finally, the result depicted that the prediction ability of ULFR was superior to Holt’s when it is compared to the actual data.

show abstract

Evaluating Temporal Bias in Time Series Event Detection Methods

Cited by 3 publications

References 33 publications

Coronavirus Disease 2019 (COVID-19): Prediction Strategy Using Sequential Deep Learning Model

Coronavirus Disease 2019 (COVID-19): Prediction Strategy Using Sequential Deep Learning Model

Multi-Scale Event Detection in Financial Time Series

Determining an effective short term COVID-19 prediction model in ASEAN countries

Contact Info

Product

Resources

About