Deep Time-Series Clustering: A Review

We present a comprehensive, detailed review of time-series data analysis, with emphasis on deep time-series clustering (DTSC), and a case study in the context of movement behavior clustering utilizing the deep clustering method. Specifically, we modified the DCAE architectures to suit time-series data at the time of our prior deep clustering work. Lately, several works have been carried out on deep clustering of time-series data. We also review these works and identify state-of-the-art, as well as present an outlook on this important field of DTSC from five important perspectives.

A versatile computational algorithm for time-series data analysis and machine-learning models

Here we introduce Local Topological Recurrence Analysis (LoTRA), a simple computational approach for analyzing time-series data. Its versatility is elucidated using simulated data, Parkinsonian gait, and in vivo brain dynamics. We also show that this algorithm can be used to build a remarkably simple machine-learning model capable of outperforming deep-learning models in detecting Parkinson’s disease from a single digital handwriting test.

Time series data analysis and ARIMA modeling to forecast the short-term trajectory of the acceleration of fatalities in Brazil caused by the corona virus (COVID-19)

Objective This paper incorporates the concept of acceleration to fatalities caused by the coronavirus in Brazil from time series data beginning on 17th March 2020 (the day of the first death) to 3rd February 2021 to explain the trajectory of the fatalities for the next six months using confirmed infections as the explanatory variable. Methods Acceleration of the cases of confirmed infection and fatalities were calculated by using the concept of derivatives. Acceleration of fatality function was then determined from multivariate linear function and calculus chain rule for composite function with confirmed infections as an explanatory variable. Different ARIMA models were fitted for each acceleration of fatality function: the de-seasonalized Auto ARIMA Model, the adjusted lag model, and the auto ARIMA model with seasonality. The ARIMA models were validated. The most realistic models were selected for each function for forecasting. Finally, the short run six-month forecast was conducted on the trajectory of the acceleration of fatalities for all the selected best ARIMA models. Results It was found that the best ARIMA model for the acceleration functions were the seasonalized models. All functions suggest a general decrease in fatalities and the pace at which this change occurs will eventually slow down over the next six months. Conclusion The decreasing fatalities over the next six-month period takes into consideration the direct impact of the confirmed infections. There is an early increase in acceleration for the forecast period, which suggests an increase in daily fatalities. The acceleration eventually reduces over the six-month period which shows that fatalities will eventually decrease. This gives health officials an idea on how the fatalities will be affected in the future as the trajectory of confirmed COVID-19 infections change.

DAMPAK PANDEMI COVID-19 TERHADAP PERTUMBUHAN DANA PIHAK KETIGA PADA BANK SYARIAH

The aim of this study is to analyze the impact of contagious of Covid-19 on the growth of Third Party Funds (DPK) in Islamic banks, to see the economic strength during the Covid-19 Pandemic. The analysis was carried out on the financial statements of two Islamic banks, are BNI Syariah and Bank Syariah Mandiri (BSM). By using time series data analysis in the aggregate time of the first semester for the period of January - June and the first month in the second semester in July 2020. The data source is taken from the financial publication monthly report from BNI Syariah and BSM. As a comparison, data for 2018 and 2019 are used in the same month period, so that obtained of the growth trend of DPK in Islamic Bank during normal times and during the Covid-19 pandemic. The results of correlation and comparison analysis, this study found that the Covid-19 pandemic has influenced the trend of public funds in Islamic banks. In choosing fund products at banks, the customer avoid investment risks by reducing deposits in the form of investment funds and prefer Wadiah as safe products. This caused DPK in Islamic banks both BNI Syariah and BSM have significant positive growth for Wadiah and decline in investment funds during the Covid-19 pandemic. This shows that Islamic banking faces financial risks in abnormal situations during the Covid-19 pandemic.

Peramalan Data Runtun Waktu menggunakan Model Hybrid Time Series Regression – Autoregressive Integrated Moving Average

Forecasting is a method used to estimate or predict a value in the future using data from the past. With the development of methods in time series data analysis, a hybrid method was developed in which a combination of several models was carried out in order to produce a more accurate forecast. The purpose of this study was to determine whether the TSR-ARIMA hybrid method has a better level of accuracy than the individual TSR method so that more accurate forecasting results are obtained. The data in this study are monthly data on the number of passengers on American airlines for the period January 1949 to December 1960. Based on the analysis, the TSR-ARIMA hybrid method produces a MAPE of 3,061% and the TSR method produces an MAPE of 7,902%.