Deck 9: Decision Aiding Predictive Systems: a Framework for Data Mining Machine Learning for Health Systems Management

Question 1

Which of the following is most true of quality of life?

A) It is a basic human right.
B) It depends on money.
C) It depends on access to health care.
D) It is found in small communities.

Answer 1

A

Question 2

Data from EHR have been useful not just for individual patients, but for overall:

A) funding.
B) dissemination of information.
C) research.
D) governmental regulation.

Answer 2

C

Question 3

If you were doing data mining on EHR records from a town showing an abnormally high rate of testicular cancer, you would look for what?

A) Patterns correlating with cancer cases
B) Population inflow
C) Breast cancer as well
D) Ethnic connections

Answer 3

A

Question 4

A generally useful framework for DM is:

A) HITECH.
B) ADA.
C) ARRA.
D) CRISP.

Question 5

Which of the following methods is not a part of DM?

A) Clustering
B) Classification
C) Regression analysis
D) Remote sensing

Question 6

What is the last phase of CRISP-DM?

A) Deployment
B) Data understanding
C) Data preparation
D) Evaluation

Question 7

Which phase of CRISP-DM comes just before deployment?

A) Data preparation
B) Business understanding
C) Modeling
D) Evaluation

Question 8

Which of the following can DM help health care organizations predict?

A) Mostly financial items
B) A wide range of health-related issues
C) Personnel matters
D) Government regulatory changes

Question 9

If you were working to create software to predict outcomes in cases of bladder cancer, the prediction set output variable would be:

A) certain.
B) unknown.
C) known.
D) uncertain.

Question 10

K-NN classifier is also known as:

A) easy find.
B) overtime.
C) lazy learner.
D) simple-search.

Question 11

Approximately how many years ago was SVM introduced to the IT world?

A) 5
B) 10
C) 20
D) 25

Question 12

If you introduced a neural network approach in your hospital's IT, what would you be trying to improve?

A) Access
B) Interoperability
C) Speed
D) Security

Question 13

Which of the following is a severity score that has been used in predictive IT in ICUs?

A) SAPS
B) ACA
C) ARRA
D) HITECH

Question 14

The main purpose of DM-ML is to create predictive decision-aiding systems.

Question 15

If you have been hired by a health innovation company to help create DM-ML systems, you will work with the past, but focus on the future.

Question 16

Data cleaning adds outliers to datasets to serve as controls.

Question 17

Data mining cannot use raw data to produce clear results, but instead must have sorted, cleaned data.

Question 18

Classification is a minor part of predictive analysis.

Question 19

Decision trees are rarely used in health IT.

Question 20

Bayesian Belief Networks have been used extensively in hospital IT.

Question 21

Regression compares three variables at a time.

Question 22

Patterns occur frequently in datasets.

Question 23

The Weka tool has been used for predictive work with breast cancer.

Question 24

The "DM" in "CRISP-DM" stands for _________.

Question 25

Effective treatment requires _________ diagnosis.

Question 26

How is predictive analysis different from descriptive analysis? Illustrate your answers with examples from healthcare scenarios.

Question 27

Why and how do you think ML techniques can be used intelligently to deal with airborne diseases?

Question 28

Provide examples of applications where ICU recordings of patients may be utilized to provide safer, higher quality, and cost-effective care to patients.

Question 29

Early diagnosis of disease is important. State ways in which ML can be utilized for this purpose.