While carrying out the storm casualty’s project, the statistical
analyses techniques and/or the machine learning techniques that
were investigated were evaluated in detail. The first dataset
contained the storm details and it encompassed 51 variables and
56,921 cases. The data set was examined with the help of R. It
helped to identify that there were numerous missing values. For
example, in ‘category’ there were 56,000 missing variables. For
addressing the issue, I eliminated the variables with over 40 %
missing values. The second database contained information relating
to storm fatalities during the year 2017. It was much smaller than
the first dataset as it contained only 775 cases and 11 variables.
While examining the data it was revealed that out of 775 victims,
490 were male. The third dataset used in the project contained
information relating to the location along with individual storm
episodes. It contained 43,579 cases and 11 variables. R was used as
the main tool to merge the three datasets together so that a
comprehensive insight into the storms and associated casualties
could be captured. R was used as one of the main techniques in the
project as it provided a robust framework to carry out tasks such as
handling of missing variables as well as feature creation. In addition
to R, SAS Enterprise Miner was also used as a useful technique to
compute diverse models in an effective and efficient manner. The
evaluation of the diverse techniques in the project was extremely
useful as it helped to merge the available data and get a bigger
picture relating to the storm casualties pattern. The use of these
techniques acted as the core foundation upon which the five
predictive models were based upon. The predictive models used in
the study include logistic regression, neural network, support vector
machine, random forest and ensemble model.
References
Catalyst, N. E. J. M. (2020).
What Is Pay for Performance in
Healthcare?
NEJM Catalyst.
https://catalyst.nejm.org/doi/full/10.1056/CAT.18.0245.
Manley, R., & Satiani, B. (2009). Revenue cycle management.
Journal
of vascular surgery
,
50
(5), 1232-1238.
World Health Organization. (2018).
Delivering Quality Health
Services: A Global Imperative
. OECD Publishing.
What Do Health Information Managers Do?: University of Wisconsin
HIMT
. UW Health Information Management & Technology. (2020,
January 21). https://himt.wisconsin.edu/about-himt/what-him-
managers-do/.
