study unit 1 crisp-dm...crisp-dm framework • the crossthe cross -industry standard process for...

13
Study Unit 1 CRISP-DM ANL 309 | Business Analytics Applications © 2014 SIM University. All rights reserved.

Upload: others

Post on 03-Aug-2020

4 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Study Unit 1 CRISP-DM...CRISP-DM Framework • The CrossThe Cross -industry Standard Process for Data Miningindustry Standard Process for Data Mining (CRISP-DM): This is a framework

Study Unit 1

CRISP-DM

ANL 309 | Business Analytics Applications

© 2014 SIM University. All rights reserved.

Page 2: Study Unit 1 CRISP-DM...CRISP-DM Framework • The CrossThe Cross -industry Standard Process for Data Miningindustry Standard Process for Data Mining (CRISP-DM): This is a framework

Introduction

• There are various steps in the cross-industry t d d f d t i i (CRISP DM)standard process for data mining (CRISP-DM)

framework

© 2014 SIM University. All rights reserved.

Page 3: Study Unit 1 CRISP-DM...CRISP-DM Framework • The CrossThe Cross -industry Standard Process for Data Miningindustry Standard Process for Data Mining (CRISP-DM): This is a framework

Why Is There a Need for a Framework?

• To be effective, business analytics must be part of the business process.

• This implies that the results of a business analytics project must be p y p jactionable.

• To exact the optimum benefits of applying business analytics, someTo exact the optimum benefits of applying business analytics, some process must be in placed to integrate the business analytics into the business process.

© 2014 SIM University. All rights reserved.

Page 4: Study Unit 1 CRISP-DM...CRISP-DM Framework • The CrossThe Cross -industry Standard Process for Data Miningindustry Standard Process for Data Mining (CRISP-DM): This is a framework

Virtuous Cycle of Data Mining

• Berry and Linoff (2000,2004) term this process as the “virtuousBerry and Linoff (2000,2004) term this process as the virtuous cycle of data mining” and define it as “an iterative learning process that builds on results over time” (Berry and Linoff, 2004, p22).

• Business analytics is continuous learning process integrated into the business process and produces results, actedinto the business process and produces results, acted upon by the companies to increase its customer value.

© 2014 SIM University. All rights reserved.

Page 5: Study Unit 1 CRISP-DM...CRISP-DM Framework • The CrossThe Cross -industry Standard Process for Data Miningindustry Standard Process for Data Mining (CRISP-DM): This is a framework

CRISP-DM Framework

• The Cross industry Standard Process for Data Mining• The Cross-industry Standard Process for Data Mining (CRISP-DM): This is a framework adopted in the application of business analytics to customer relationship management

• The CRISP-DM integrates business analytics into the business process by focusing data mining or business analytics technology on specific business problemstechnology on specific business problems.

© 2014 SIM University. All rights reserved.

Page 6: Study Unit 1 CRISP-DM...CRISP-DM Framework • The CrossThe Cross -industry Standard Process for Data Miningindustry Standard Process for Data Mining (CRISP-DM): This is a framework

Six Phases in CRISP-DM

1. Business Understanding

2. Data Understanding

3. Data Preparation

4 Modeling4. Modeling

5. Evaluation

6. Deployment

© 2014 SIM University. All rights reserved.

Page 7: Study Unit 1 CRISP-DM...CRISP-DM Framework • The CrossThe Cross -industry Standard Process for Data Miningindustry Standard Process for Data Mining (CRISP-DM): This is a framework

CRISP – DM Framework

© 2014 SIM University. All rights reserved.

Page 8: Study Unit 1 CRISP-DM...CRISP-DM Framework • The CrossThe Cross -industry Standard Process for Data Miningindustry Standard Process for Data Mining (CRISP-DM): This is a framework

Data Understanding

Collect the data

Perform exploratory data analysis to understand the dataPerform exploratory data analysis to understand the data

Evaluate the quality of the data

Identify subset of the data for modelling

© 2014 SIM University. All rights reserved.

Page 9: Study Unit 1 CRISP-DM...CRISP-DM Framework • The CrossThe Cross -industry Standard Process for Data Miningindustry Standard Process for Data Mining (CRISP-DM): This is a framework

Data Preparation

Prepare the raw data setPrepare the raw data set

Select the relevant variables and cases for data mining objectives

Apply the necessary transformations to the data

Deal with missing values

Most labour-intensive phaseost abou te s e p ase

© 2014 SIM University. All rights reserved.

Page 10: Study Unit 1 CRISP-DM...CRISP-DM Framework • The CrossThe Cross -industry Standard Process for Data Miningindustry Standard Process for Data Mining (CRISP-DM): This is a framework

Modelling

• Select and apply the appropriate modeling techniques.

• Build and calibrate several models where appropriate.pp p

• Loop back to the data preparation phase if necessary.

© 2014 SIM University. All rights reserved.

Page 11: Study Unit 1 CRISP-DM...CRISP-DM Framework • The CrossThe Cross -industry Standard Process for Data Miningindustry Standard Process for Data Mining (CRISP-DM): This is a framework

Evaluation

E l t th d l i t f th d t i i bj ti d• Evaluate the models in terms of the data mining objectives and the business goals as identified in the business understanding phase.

• Review the process and decide on the follow-up action.

© 2014 SIM University. All rights reserved.

Page 12: Study Unit 1 CRISP-DM...CRISP-DM Framework • The CrossThe Cross -industry Standard Process for Data Miningindustry Standard Process for Data Mining (CRISP-DM): This is a framework

Deployment

Prod ce final report and presentation• Produce final report and presentation.

• Identify how the results can be translated into appropriate business strategies to achieve the business goals.

© 2014 SIM University. All rights reserved.

Page 13: Study Unit 1 CRISP-DM...CRISP-DM Framework • The CrossThe Cross -industry Standard Process for Data Miningindustry Standard Process for Data Mining (CRISP-DM): This is a framework

Business Analytics: An Iterative Process

• The sequence of the phases is not rigid. In fact, moving back and forth among the phases is always required.

• The business analytics process is iterative, requiring the lessons learnt during the processes be it inputs into the existing or new and more focused business questions and processes

© 2014 SIM University. All rights reserved.