Analytics Midterm

Front 1

What are the levels of abstraction for data

Back 1

-Information
-Knowledge
-Intelligence

Front 2

What are some problems with decisions based on data by humans

Back 2

- Intuition is prone to errors
- Illusion of memory
- Illusion of knowledge

Front 3

What technologies are associated with business intelligence?

Back 3

o Data management
o Reporting – report generation
o Intelligence – automatic generation of insights

Front 4

What are challenges of Business Intelligence

Back 4

- Does not form a closed loop in it's generation of insights
A. Lack of timeliness
B. Lack of actionable process

Front 5

What is the difference between BA and BI

Back 5

BA provides system generated intelligence based on automated data analysis (i.e. closes the loop)

Front 6

What are the foundation levels of business intelligence

Back 6

- Standard Reports
- Ad hoc reports
- Query drill down
- Alerts

Front 7

What are the core BA levels of analytics

Back 7

Statistical analysis
Forecasting
Predictive modeling
Optimization

Front 8

What are the two approaches to BA solution design? What is the best practice?

Back 8

- top-down – based on user requirements
- bottom-up – based on source system structures

- best practice: do both. start with top-down and source bottom-up

Front 9

What are two computational issues during implementation of BA

Back 9

- High performance analytics (HPA)
 Speeding up the modeling stage
 Associated with Big Data (data existing beyond databases)
- Real time
 Ability to score data in a near instantaneous manner

Front 10

What is the focus of data mining?

Back 10

Automatically recognizing complex patterns from data

Front 11

What is the process and challenge of data mining

Back 11

- Learning from a finite set of sample data
- Generalizing/producing useful output on new cases

Front 12

What are the two major types of learning approaches for data mining?

Back 12

- Supervised learning
- Unsupervised learning

Front 13

That are the 3 types of supervised learning in data mining?

Back 13

o training data : to train the algorithm (model)
o validation data : to validate the model after being trained
o test data : used to evaluate final model's performance