Improving Data Analytics during a Pandemic Episode: A Review

Jeffrey E. Jarrett^*
*Correspondence: Dr. Jeffrey E. Jarrett, Department of Management Science, University of Rhode Island, Rhode Island, United States, Tel: +17142345677, Email:

Keywords

Decision statistics • Capabilities • EWMA • Multivariate data and pandemics

Introduction

The pandemic in America and other nations produced a panic in federal and state health institutions. The panic did not occur because the federal and state health care and diagnostic institutions knew very little about the COVID-19 virus nor had the necessary equipment to maintain health standards in the face of the growing and spreading of the virus. Our purpose in this manuscript is not to analyze the politics of the care and diagnostics in the preparation, reaction and the execution of the health treatments execution of health agencies in the response to the spreading of the virus to the vulnerable communities across the United States.

The goal is to consider better analytical data tool which could have very much improve the interpretation of data retrieval and interpretation to assist in the decisions and analysis to improve the methodology to assist health professionals in determining health care diagnosis and in turn health strategies in therapeutic care. If the real option the following trade-off arises: if information is timely, the therapeutic strategies decision can be based on better information to arrive a better decision making in the uncertain world of treating a pandemic virus. Incorrect and untimely information is the product of the misreporting of factual events associated with diagnostics and error associated with poorly utilized statistical analytics and eventual predications.

Professionals information to determine, detect, investigate and respond to evidence concerning disease and the outbreaks of epidemics and pandemics. A predominant method of analysis in this monitoring in classical bio surveillance is the movement of classical disease surveillance utilizing multivariate quality control methods original developed in industrial applications. Applied multivariate statistical quality control (MSQC) methods in addition, MSQC applications by continued earlier studies by focusing on direction all sensitive procedures in bio surveillance [1,2].

Previously, the usefulness of monitoring social and economic processes is an excellent application of Statistical Quality Control, SQC. They created new ways to analyze data by exponentially weighted moving average (EWMA) control charts for residuals [3]. This analysis was said to be very useful in the detection of shifts in variability. Furthermore, the newly improved methods by expanding the analysis to the case of multivariate methods. In turn, OKSH suggested examining the works [4-9]. The multivariate form of the MEWMA control chart simultaneous monitors two or more related processes in an exponentially weighted moving-average (EWMA) control chart. Contains information from all the previous subgroups or observations.

As an illustration, a MEWMA chart monitors temperature and pressure in a plastic injection molding process. Each observation (or point) contains information from all the previous subgroups or observations in combinations with a user defined weighting factor. Thus, MEWMA charts aid the analyst by detecting small process shifts faster than other forms of multivariate control charts including the well-known T2 Chart. In addition, MEWMA charts are neither significantly influenced by changes having very large or very small values entering the calculation. Last, MEWMA charts have the ability to tailor detect any size shift in a process. Hence, they can be utilized to monitor in-control processes for detecting tiny shifts away from the target of goal of the process.

MSQC monitors two or more data sets in combination. This is very important in health diagnostics which includes a large number of previous studies qualities in multivariate quality monitoring literature. All of these models employ the multivariate testing statistic entitled Hoteling T2 [10-24].

As an alternative, I propose the using an alternative scheme developed which builds on the Hoteling model before the formation of the exponential weighted moving-average (EWMA) [22].Presented a substantial improvement entitled MEWMA [24]. Utilized Markov chain analysis to calculate the average return length (ARL) of MEWMA and designed the new MEWMA under the assumption of serial independence [12]. Again, discussed the robustness of the MEWMA and indicated how to implement this method in monitoring health time series data where two or more data sets of health data have some serial correlation of value greater than zero and up to 0.60 p-values of correlation [21-23]. Utilizing the criterion of ARL, MEWMA in bio surveillance in the age of a Pandemic having cross-correlated variables provide a better analytical framework than simple using seven day moving-average daily statistics where the velocity and variation in change have proven to be unpredictable and often providing insufficient forecasts and prediction for data driven decision making.

Conclusion

To conclude and sum up, the use of diagnostic analysis will not easily prevent and cure COVID-19. In particular, the use of a seven period moving averages a simple but very restrictive statistic in its use and interpretation. Thorough use of sophisticated analyses such as MEWMA utilizing often quality analysis of two or more cross correlated diagnostic and health metrics variables coupled continuous improvement will enable a better future for controlling the environment of the COVID-19 Pandemic that we are attempting to survive and live through. The methodology promulgated in this document suggests better use of medical strategies to solve the problems of the current Pandemic. The response by the current administration in preparing the nation for the Pandemic is noticeably insufficient. Last, this document aims at promoting better analytical methods for interpreting the collected data by health professionals in performing their professional tasks to end the Pandemic. Another solution that may have promise is in [25-26].

References

1. Fricker, RD. “Directionally Sensitive Multivariate Statistical Process Control Methods with Application to Syndromic Surveillance.” Advances in Disease Surveillance. 3 (2007): 1-17.
2. Fricker, Ron, Matthew C. Knitt  and Cecilia X. Hu. “Comparing Directionally Sensitive MCUSUM and MEWMA Procedures with Application to Bio Surveillance.” Qual. Eng. 3(2004): 478-494.
3. André, Paul,  Andrei Filip and Luc Paugam. “Impact of Mandatory IFRS Adoption on Conditional Conservatism in Europe.” ESSEC Working Papers. Working Papers, (2013).
4. Lowry, CA, William H. Woodall, Charles W. Champ and Steven E. Rigdon. “A Multivariate Exponentially Weighted Moving Average Control Chart.” Technometrics. 34 (1992): 46-53.
5. Pan, Xia, and E.Jeffrey Jarrett. “Applying State Space to SPC: Monitoring Multivariate Time Series.” J. Appl. Stat. 31 (2004): 397-418.
6. Jarrett, JE and XiaPan. “Using Vector Autoregressive Residuals to Monitor Multivariate Processes in the Presence of Serial Correlation.” Int. J. Prod. Econ. 106 (2007): 204-216.
7. McDonald TL. “Review of Environmental Monitoring Methods: Survey Designs.” Environ. Monit. Assess. 85 (2003): 277–292.
8. Roberts, DC and Edna N. Roberts. “Exact Determination of Earnings Risk by the Coefficient of Variation.” J Finance. 25 (1970): 1161-1165.
9. Runger, GC, Russell R. Barton, Enrique del Castillo, and William H. Woodall. “Optimal Monitoring of Multivariate Data for Fault Patterns.” J. Qual. Technol. 39 (2007): 159-172.
10. Lowry, CA and Douglas C. Montgomery. “A Review of Multivariate Control Charts.” IIE Transactions. 27 (1995): 800-810.
11. Sullivan, JH and William H. Woodall. “A Comparison of Multivariate Quality Control Charts for Individual Observations.” J. Qual. Technol. 28 (1996): 398-408.
12. Djauhari, MA. “Improved Monitoring of Multivariate Process Variability.” J. Qual. Technol. 37 (2005): 32-39.
13. Chong BKN and S. H. Quah. “Multivariate Quality Chart for Process Dispersion Based on Individual Observations.” Qual. Eng. 15 (2007): 639-642.
14. Christopher, Kruegel, Fredrik Valeur and Vigna Giovanni. “Intrusion Detection and Correlation, Challenges and Solutions.” New York: Springer, United States, 14 (2005): 118.
15. Ye, Nong and Qiang Chen. “An Anomaly Detection Technique Based on a Chi Square Statistics for Detecting Intrusion into Information System.” Qual. Reliab. Eng. Int. 17 (2007): 105-112.
16. Ye, Nong, Joseph Giordano and John Feldman. “A Process Control Approach to Cyber Attack Detection.” Communication of the ACM. 44 (2007): 76-82.
17. Ye, Nong, Qiang Chen and C. Borror. “EWMA Forecast of Normal System Activity for Computer Intrusion Detection.” IEEE Transactions on Reliability. 53 (2004): 557-566.
18. Bersimis S, S. Psarakis and J. Panaretos. “Multivariate Statistical Process Control Charts: An Overview.” Qual. Reliab. Eng. Int. 23(2007): 517-543.
19. Yang, SF, and M.A.Rahim. “Economic Statistical Process Control for Multivariate Quality Characteristics under Weibull shock model.” Int. J. Prod. Econ. 98 (2005): 215-226.
20. Jarrett, JE and XiaPan. “the Quality Control Chart for Monitoring Multivariate Auto correlated Processes.” Computational Statistics and Data Analysis. 51 (2007): 3862-3870.
21. Jarrett, JE. “Analyzing Data utilized in Process Control and Continuous Improvement.” Int J Econ Manag Sci. 5(2016).
22. Jarrett, JE. “Examiningthe Monitoring of Health Surveillance During a Pandemic: A Review of Methodology.” Academy of Social Science Journals. 5 (2020): 1722-1728.
23. Jarrett, JE. “Bias in Adjusting Asset values for Changes in the Price Level: An Application of Estimation Theory.” J. Account. Res. 12 (1974): 63-66.
24. Jarrett,  JE. “Intellectual Property Valuation and Accounting.” Intel Prop Rights. 5 (2017).
25. Pan, Xia. “An Alternative Approach to Multivariate EWMA Chart.” J. Appl. Stat. 32 (2005): 695-705.
26. Makridakis Spyros, Evangelos Spiliotis and Vassilios Assimakopoulos. “Statistical and Machine Learning forecasting methods: Concerns and ways forward.” 19 (2018).