Communicating Analytical Results and Interpreting Machine Learning Models with SAS Viya

Описание: 5 Tips and Tricks that will make your life as data scientist easier!

The presentation and communication of analytical results to those who have to apply and accept analytical models is often a challenging tasks. Different vocabulary, the demand for “non-analytical” explanations and the exercise to convert model details into “Aah”-experiences on the business side are tasks that data scientists are often facing. This talk illustrates how SAS Viya offers a comprehensive set of methods to create, interpret and communicate analytical results with interactive and visual tools. It also shows how SAS Model Studio, the visual modeling front end in SAS Viya, provides so called “Model Interpretability” charts that are created out of the box that allow to better understand and interpret the details of even complex machine learning methods. The tips and tricks have been collected over many successful machine learning projects in various business domains.

The tricks and tips include:
1. Performing interactive cutoff analysis to visualize the outcome of predictive models (03:24)

2. Use a decision tree to „explain“ complex machine learning models (11:38)

3. Turn on the model interpretability charts like partial dependence plots, ICE plot and LIME charts (16:36)

4. Quantify the importance of explanatory variables in € and $ (22:45)

5. Display the (hidden) regression coefficient of categorical input variables (32:47)

Downloads
• Presentation 107 at Github:
https://github.com/gerhard1050/DataSc...
• Book "Data Quality for Analytics" for business case calculation in tip 4: https://github.com/gerhard1050/Data-Q...
• Book "Applying Data Science" for the employee survival case study in tip 5: https://github.com/gerhard1050/Applyi...
• Article @SAS Communities: Display the hidden estimate for the reference category in EFFECT coding for better interpretability
https://communities.sas.com/t5/SAS-Co...
• Article @SAS Communities: Encoding of CLASS Variables in Regression Analysis - Better understand the ORDINAL encoding :
https://communities.sas.com/t5/SAS-Co...

Links
• More webinars - "Home Alone" Data Science Webinar from Gerhard: LinkedIn   / home-alone-invest-now-your-data-science-sk...   and Medium   / home-alone  
• Blogs on LinkedIn:   / gerhardsvolba  
• Twitter:   / gsvolba  
• Content on Github: https://github.com/gerhard1050
• Books @SAS-Press: https://support.sas.com/svolba

#DataScienceClass #DataScience #SASviya

SAS SOFTWARE D-A-CH ABONNIEREN
   / @sassoftwaredach  

ÜBER SAS
SAS ist weltweit Marktführer im Bereich Analytics. Kunden weltweit setzen innovative Software und Services von SAS ein, um Daten in Wissen zu verwandeln und intelligente Geschäftsentscheidungen zu treffen. Seit 1976 verschafft SAS Kunden rund um den Globus THE POWER TO KNOW.

http://www.sas.com/de_de/home.html

Facebook►   / sasdach  
Twitter►   / sas_dach  
LinkedIn►   / sas  
Blog► http://blogs.sas.com