DAV CHAP 6 Flashcards
Data mining definition
Process of Discovering new and meaningful correlations patterns trends by “mining” large amount of stored data using pattern recognition tech and statistical and mathematical techniques
known as knowledge discovery, data surfing, data harvesting
What is driving data mining
- Change in technology - increase use of internet and computing power + data warehouses + better modelling approaches
- Change in customer behavior - more informed, demanding, willing to switch to competitor, harder to satisfy needs as it gets more complex
- Change in competition - Evolution of strategy like one to one marketing and mass marketing, more competition, faster pace, niche players.
Crisp DM model
- Business understanding - understand business objective
- Data understanding - collect, identify, describe data
- Data preparation - select, clean data
- Modelling - select modelling technique, build model and assess
- Evaluation - Evaluate results and review
- Deployment - plan deployment + presentation + review
Decision tree
Supervised learning method - classification method that uses value of input variable to predict value of categorical variable
Decision tree benefits
- Make readily understandable rules to predict customer behavior
- Evaluate values of different outcomes and probability to reach them
- Produce graphical representation of how different factors affect the outcome
- Make segmentation scheme based on decision tree results
Clustering
Creates groups of records that are similar to each other within a particular group and very different across different groups
Association between members determined by characteristics specified in the analysis
Explore large amount of data and organize it
Define data understanding
identify source of data, collect, describe explore and verify data quality
Data preparation
select data to clean, integrate the data and format it
Modelling
Select modelling technique, build, fine tune and assess
Business understanding
understand business objective, and goal of data mining, work with stakeholders to produce project plan
Decision tree practical applications
- Reduce customer fraud
- Who is likely to stop buying from us
- Who’s likely to be a credit risk