Since COVID-19 has been a worldwide serious problem, experts from different fields access COVID 19 with diverse methodologies. One of the representative methods is the machine learning approach with big data [1].
Most experts increasingly access the Coronavirus with the big data analysis methods such as crowdsourcing, machine learning, and Google’s data gathering tool. These methodologies could help improve the relationship between patients and physicians regarding physical value by improving the efficiency and quality of physicians treating patients.
Applying this big data method, users could be active participants rather than passive participants. Big data analysis both has strengths and weakness.
Strengths of Big Data Analysis
- For strengths, big data analysis is highly effective for predicting the futures based on the current situation.
- Diverse types of data could be gathered easily and rapidly.
- Dealing with large amount of data is easier and possible.
- Real-time monitoring could be possible in this fast-paced market.
In case of the COVID 19, rapid responses and analysis were needed to actively prevent the spread of virus. Regarding the amount of data, large amount of data could be monitored at once, repetitively with the same method. The expenses will be reduced, too. Detection of errors and frauds are possible with increased productivity.
Weaknesses of Big Data Analysis
Big data analysis methodology also has some weaknesses.
- The big data itself could be inaccurate.
- It is impossible to entirely avoid the possibility of data errors. Dewey defeats Truman[2] could be one of the examples. Google’s biased data analysis[3] is another example.
- There could be the overestimation of data and it is very difficult to verify the collected data. For instance, if there are too much data, it is difficult to filter the appropriate amount of data based on the researchers’ needs. This phenomenon could be explained by the concept called “automation bias[4]”.
- We have the tendency to over-rely on automation. Regarding a large amount of data, privacy issue also stands out. This issue is very complex and complicated because publicly opening the data could also be the possible solution to solve the problems of big data analysis. However, Opening the data source publicly could harm people’s privacy[5].
Combinations could be the solution
Some solutions to resolve the problems of big data analysis could be possible. Combining different sources of data and methodologies could be one of the solutions to supplement the weaknesses of methodologies. Another solution could be to combine the big data with small data. Since big data analysis is the quantitative method, applying the qualitative methods to the results of the quantitative methods could be another solution. Using traditional data with high tech data could be the other solution.
[1] https://www.nature.com/articles/s41746-018-0045-1
[2] https://en.wikipedia.org/wiki/Dewey_Defeats_Truman
[3] https://www.technologyreview.com/f/609959/google-photos-still-has-a-problem-with-gorillas/
[4] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3240751/
[5] https://www.nytimes.com/interactive/2019/12/19/opinion/location-tracking-cell-phone.html
