DAT 520 Milestone Three Guidelines and Rubric
In this milestone, you will perform an evaluation of your decision model and revise your decision model as needed. Evaluation examples are if you are
performing a bottom-up style recursive partitioning analysis, and you should report on the error rate and variable selection. You might also consider alternative
variable categorizations to improve your model. If you are performing a top -down decision tree modeling exercise, what are the threshold values that cause the
tree to flip? You should perform sensitivi ty analysis on the critical variables in your tree and report what those sensitivity analyses are telling you. For either sty le
of modeling, what makes your tree stronger? What breaks the model? For more information on completing this milestone, please ref er to the Final Project
Notes in the Assignment Guidelines and Rubrics folder.
Specifically, the following critical elements must be addressed in your final submission:
Include the structure of your revised decision tree, with a clear description.
Evaluate the results of your revised model, including analysis that is specific to your revised model. In your evaluation, reflect on the appropriateness
and adjustments of the revised model, as well as the accuracy of the results you obtained.
Suitable diagnostics should be incorporated into the model.
Guidelines for Submission: This milestone should be 2 to 3 double-spaced pages of text, with tree model images and any other supporting material appended.
Review your work to ensure that there are no major errors in writing mechanics. If you have citations, include the sources at the end and cite them APA format.
Critical Elements Proficient (100%) Needs Improvement (70%) Not Evident (0%) Value
Structure Deci s i on tree and des cri ption are cl early
s tructured
Deci s i on tree and des cri ption are
s omewhat cl early s tructured
Deci s i on tree and des cri pti on are not
adequatel y s tructured
30
Evaluation of Results Eval uati on cons i ders reas onablenes s ,
accuracy, mi s s ing/extraneous el ements ,
and error i n the model
Eval uati on does not ful l y cons i der
reas onabl enes s , accuracy,
mi s s i ng/extraneous el ements , and error
i n the model
Eval uati on does not cons i der
reas onabl enes s , accuracy,
mi s s i ng/extraneous el ements , and error
i n the model
30
Model Diagnostics Model i ncl udes cl ear us e of di agnos ti cs Model bui l ds i n parti al us e of
di agnos ti cs
Model does not i ncl ude di agnos ti cs 30
Articulation of
Response
Submi s s i on has no major errors rel ated
to grammar, s pel l i ng, s yntax, or
organi zati on
Submi s s i on has major errors rel ated to
grammar, s pel l i ng, s yntax, or
organi zati on that negati vel y i mpact
readabi l ity and arti culation of mai n
i deas
Submi s s i on has criti cal errors rel ated to
grammar, s pel l.
History Class XII Ch. 3 Kinship, Caste and Class (1).pptx
DAT 520 Milestone Three Guidelines and Rubric In this m.docx
1. DAT 520 Milestone Three Guidelines and Rubric
In this milestone, you will perform an evaluation of your
decision model and revise your decision model as needed.
Evaluation examples are if you are
performing a bottom-up style recursive partitioning analysis,
and you should report on the error rate and variable selection.
You might also consider alternative
variable categorizations to improve your model. If you are
performing a top -down decision tree modeling exercise, what
are the threshold values that cause the
tree to flip? You should perform sensitivi ty analysis on the
critical variables in your tree and report what those sensitivity
analyses are telling you. For either sty le
of modeling, what makes your tree stronger? What breaks the
model? For more information on completing this milestone,
please ref er to the Final Project
Notes in the Assignment Guidelines and Rubrics folder.
Specifically, the following critical elements must be addressed
in your final submission:
clear description.
that is specific to your revised model. In your evaluation,
reflect on the appropriateness
and adjustments of the revised model, as well as the accuracy of
the results you obtained.
2. Guidelines for Submission: This milestone should be 2 to 3
double-spaced pages of text, with tree model images and any
other supporting material appended.
Review your work to ensure that there are no major errors in
writing mechanics. If you have citations, include the sources at
the end and cite them APA format.
Critical Elements Proficient (100%) Needs Improvement (70%)
Not Evident (0%) Value
Structure Deci s i on tree and des cri ption are cl early
s tructured
Deci s i on tree and des cri ption are
s omewhat cl early s tructured
Deci s i on tree and des cri pti on are not
adequatel y s tructured
30
Evaluation of Results Eval uati on cons i ders reas onablenes s ,
accuracy, mi s s ing/extraneous el ements ,
and error i n the model
Eval uati on does not ful l y cons i der
reas onabl enes s , accuracy,
mi s s i ng/extraneous el ements , and error
i n the model
3. Eval uati on does not cons i der
reas onabl enes s , accuracy,
mi s s i ng/extraneous el ements , and error
i n the model
30
Model Diagnostics Model i ncl udes cl ear us e of di agnos ti
cs Model bui l ds i n parti al us e of
di agnos ti cs
Model does not i ncl ude di agnos ti cs 30
Articulation of
Response
Submi s s i on has no major errors rel ated
to grammar, s pel l i ng, s yntax, or
organi zati on
Submi s s i on has major errors rel ated to
grammar, s pel l i ng, s yntax, or
organi zati on that negati vel y i mpact
readabi l ity and arti culation of mai n
i deas
Submi s s i on has criti cal errors rel ated to
grammar, s pel l i ng, s yntax, or
organi zati on that prevent
unders tandi ng of i deas
4. 10
Earned Total 100%
Decision TreeAsset-Deposit/ Loss below 0.1Grossly
UnstableAsset/ Deposit Ratio below 1.11Asset -Deposit/Loss
above 1.0Relatively stable0Asset/ Deposite Ratio above 1.1
StableIDNameValueProbPredKindNSS1S2S3S4S5RowColMark0
TreePlan000D212000251TRUE100E234000185TRUE200T0000
00325TRUE31D356700109TRUE41T000000279TRUE503T0000
00213TRUE603E289000913TRUE703E210110001913TRUE86T
000000717TRUE96T0000001217TRUE107T0000001717TRUE1
17T0000002217TRUE
x
AnalysisFDIC: HSOB Bank & Thrift FailuresTable BF03Federal
Deposit Insurance CorporationUS and Other Areas(Dollar
amounts in thousands)Effective Date(s): 2010 - 2017Insurance
Fund: ALLCharter Type: ALLTransaction Type: All
FailuresStateNumber of InstitutionsNumber of FailuresNumber
of Assistance TransactionsAssetsDepositsEstimated Loss
(12/31/2016)Ratio (Assets/Deposits)Assets-
Deposits/LossAlaska000000Alabama4403,923,5923,524,148527
,1161.1133448420.76Arkansas220258,100237,22731,9801.0879
874550.65Arizona101001,724,9111,521,634367,1011.13359125
80.55California1818010,369,8638,527,9431,180,1361.21598643
41.56Colorado7705,907,9345,033,322997,0681.1737643650.88
Connecticut11026,36825,7159,2111.0253937390.07District of
Columbia000000ERROR:#DIV/0!ERROR:#DIV/0!Delaware000
000ERROR:#DIV/0!ERROR:#DIV/0!Florida5656017,814,85515
,603,0392,734,2351.1417554620.81Georgia6161017,615,64216,
457,1705,499,6121.0703931480.21Guam000000ERROR:#DIV/0
!ERROR:#DIV/0!Hawaii000000ERROR:#DIV/0!ERROR:#DIV/
0!Iowa11091,58081,96716,0531.1172789050.6Idaho110153,361
145,8133,4871.0517649322.16Illinois4444018,464,88416,970,6
6. Virginia000000ERROR:#DIV/0!ERROR:#DIV/0!Wyoming0000
00ERROR:#DIV/0!ERROR:#DIV/0!Totals:36336301596478431
3927570428178689.591.1462720230.72
Running head: DAT 520 FINAL PROJECT MILESTONE TWO
DAT 520 FINAL PROJECT MILESTONE TWO
7DAT 520 Final Project Milestone TwoStudent
NameDecisions Methods and ModelingSouthern New Hampshire
University
Bank Failures
Structure
The model employed was a top-down structure. The focus of the
decision tree is to find states where bank failures are most
probable. It, therefore, defines a model where; the liquidity of
the bank defines their stability. The more stable a bank, the less
likely it is to fail. Using Asset to deposit ratios and further non
- current assets to loss ratio will give the best estimate of the
stability of different banks in a multitude of states. These three
variables will be the major determinants through which the
model will be used in determining the nature of the Bank
Failures.
Documentation
Different banks have in the past failed. This is often
characterized by their inability to meet depositor money. When
7. a bank receives money from a prospective client, often than not,
they decide to use the money deposited in investment projects.
Should they be in a position to meet the obligations to their
depositors, then they are continuing operations, however, in the
event, their investments do not return favorable profits, a bank
may lose its stability and be declared to have failed (Bruce,
2017). A bank may also be unable to meet its obligations to its
creditors. Such instance often leads to an unstable economic
and financial environment and have in the past led to the need
for banks to receive bailouts through which they can meet their
obligations to their depositors (Bruce, 2017). For this analysis,
a summary, of different states and the corresponding failures
was used to determine the trends between bank failures and
states.
To determine the probability of a bank being capable of
offsetting some of its debt, the first comparison that will be
made will be the ratio between the assets the bank holds, and
the total amount made in deposits. This should give a rough
estimate of the capacity of the bank to meet its obligations to its
main clientele. The higher the ratio, the more stable the bank.
Secondly, the banks capacity to mitigates itself from loss is
another measure that can be used to determine the stability of
the bank. The difference between the Assets and amount
Deposited can give a good picture of the overall liquidity of the
company. With this figure, finding its ratio against the losses
incurred in the last fiscal year (2016) can give a good picture of
the stability of the bank and hence the overall probability of it
incurring losses.
Evaluation
Data on Bank Failures between 2010 and 2017 was used as the
primary information on the trends in bank failures. From the
analysis, nine states appear to have experienced a lot of failures
over the past seven years ("FDIC: HSOB Commercial Banks,"
2017). The states of Arizona, California, Florida, Georgia,
Illinois, Minnesota, Missouri, South Carolina, and Washington
have noted the highest propensity of bank failures. Georgia
8. ranked the highest with a total of 61 bank failures in this
period. Florida then followed with 56 bank failures and then
Illinois with 44("FDIC: HSOB Commercial Banks," 2017).
Looking at the ratio between assets and deposit for these three
banks, all were above 1, which is a sign of stability. However,
the banks in Georgia and Illinois recorded significantly lower
ratios. The state of Georgia had the least with 1.07("FDIC:
HSOB Commercial Banks," 2017). It is important to note that
states like Connecticut, Idaho, and Minnesota also recorded low
Asset/Deposit ratios.
The second measure of overall failure was to determine the ratio
between the difference between the banks capacity to liquidate
its assets and the losses that it made. This would make for a
clearer picture of the stability of the bank as a recent figure was
used in this instance. From the analysis, Connecticut recorded
the lowest figure at 0.07("FDIC: HSOB Commercial Banks,"
2017). Still, comparatively, the number of deposits for the state
was significantly lower compared to that of Georgia. It can be
difficult to predict the geographical location of banks that will
experience the most loss in the future. It, however, can be
assumed that the states of Georgia, Florida, and Illinois present
the largest risk for bank failures ("FDIC: HSOB Commercial
Banks," 2017). Consequently, states like Connecticut and
Minnesota present some of the tales of a dwindling trust and
investment into local banks. These states, therefore, present
with the highest risk.
To summarize the steps are followed, first, the data is pulled
from the site provided by Federal Deposit Insurance
Corporation that includes the summary of assets, deposits and
loses in banks across fifty different states in the United States.
Following that, the ratio of assets/deposits per state, and assets-
deposits to loss are used to determine the stability of the banks
in different states. As the decision model presents in relation
with the defined ratio for stability, banks that scored a ratio
above 1.1 in the first instance (assets/deposit ratio) are
considered relatively stable, and therefore, they are the ones
9. identified as not likely to fail (see Appendix A for decision tree
model). The financial ratio above 1.1 indicates that the banks in
the specified area can meet their obligations to depositors, since
they have more assets than they do deposits. However, the
banks that score below the figure were at a high risk of failing.
In the excel attachment, ‘’I’’ and ‘’J’’ columns provide the
clear picture of the rational state by state. (see Appendix B for
excel analysis). The financial ratio, which is below 1.1, shows
that the bank is barely capable of meetings its obligations to
depositors, and following that this is considered as not a good
sign and a clue of failure.
References
Bruce, L. (2017). What happens to your accounts if the bank
fails?. Retrieved from https://www.bankrate.com/banking/what-
happens-if-your-bank-fails/
FDIC: HSOB Commercial Banks. (n.d.). Retrieved from
https://www5.fdic.gov/hsob/hsobRpt.asp