MIS372 - Predictive Analytics
Questions and Answers.
Data mining - Answer - Finding/extracting patterns in data
- Understanding previously unknown patterns
- Combines statistics, machine learning and computing
Data mining enablers - Answer - Large volumes of data
- Different dimensions of data
- Complex questions that require more than just traditional statistical analysis
Data mining types - Answer 1. Descriptive modelling
2. Predictive analysis
Descriptive modelling and example - Answer - Historical data
- Summaries what has happened (dashboards/scorecards)
- What was the best selling product of Apple between 2016-2017?
Predictive analyses and example - Answer - Future outcomes given historical data
- What is likely to occur in the future
- What is the likely retention rate in the IS course this year?
Descriptive modelling features - Answer - Extracts/presents main features from the data
- Summarises data regarding data dimension e.g. time, manufacturer
- Co-occurrences of events/patterns
- Find associations in data elements
- Makes no assumptions about data prior to modelling
Descriptive modelling techniques - Answer - Correlation analysis
- Data clustering (segmentation)
- Association rule mining/sequential pattern mining
Predictive modelling features - Answer - Models existing historical data to predict likely
future outcomes or events
- Represents how different variables relate to each other
Questions and Answers.
Data mining - Answer - Finding/extracting patterns in data
- Understanding previously unknown patterns
- Combines statistics, machine learning and computing
Data mining enablers - Answer - Large volumes of data
- Different dimensions of data
- Complex questions that require more than just traditional statistical analysis
Data mining types - Answer 1. Descriptive modelling
2. Predictive analysis
Descriptive modelling and example - Answer - Historical data
- Summaries what has happened (dashboards/scorecards)
- What was the best selling product of Apple between 2016-2017?
Predictive analyses and example - Answer - Future outcomes given historical data
- What is likely to occur in the future
- What is the likely retention rate in the IS course this year?
Descriptive modelling features - Answer - Extracts/presents main features from the data
- Summarises data regarding data dimension e.g. time, manufacturer
- Co-occurrences of events/patterns
- Find associations in data elements
- Makes no assumptions about data prior to modelling
Descriptive modelling techniques - Answer - Correlation analysis
- Data clustering (segmentation)
- Association rule mining/sequential pattern mining
Predictive modelling features - Answer - Models existing historical data to predict likely
future outcomes or events
- Represents how different variables relate to each other
