CRSP-DM

More like this

Designing The Team: Task, Ppl, Process

by Ethan Kwan

Online Shopping

by Dennis Wong

SOFTWARE QUALITY METRICS

by mkti 2012

Enterprise Architect

by José Arévalo

CRSP-DM

Deployment

If deployment in the scope, the implementations should be described

Final report or a software component

Develops concrete actions based on the results of the data mining models.

User guide

Maintenance.

Monitoring

Evaluation

The process should be reviewed in general.

Results are checked against the defined business objectives

Visualization of model and metrics

Metrics are also visualized to illustrate the results

The visualization of the trained models means, that e. g. a decision tree are shown to explain and evaluate the models

Defining metrics

Use metrics to evaluate the quality of trained models

Modeling

All data mining techniques can be used

Select the modeling technique

Training data

Building test and training sets

Derived attributes have to be constructed

Data preparation

Inclusion and exclusion criteria

Describing input and output data

Cleaning data

Bad data quality can be handled

Data selection

Transformation

Data understanding

Collecting data from data sources

Checking the data quality

Exploring and describing it

Entitle the concrete data sources and explain from where the data has been collected during the data understanding phase

Descriptive statistic

The most important method of data understanding

Statistical analysis and determining attributes and their correlations.

In terms of data mining, it is a classification with the binary target attribute.

Business Understanding

Textual description of the business goals and why data mining is useful for the specific use cases, the concrete data mining goal is specified in a structured manner

Required project plan

Data mining goal

Data mining success criteria

Data mining type