Topics: Coronavirus | AI | Data Science | Deep Learning | Machine Learning | Python | R | Statistics

KDnuggets Home » News » 2015 » Dec » Opinions, Interviews, Reports » Kanri Distance approach for translating Predictive Models to Actions ( 15:n42 )

Kanri Distance approach for translating Predictive Models to Actions


Kanri proprietary combination of patented statistical and process methods provides a uniquely powerful and insightful ability to evaluate large data sets with multiple variables.



Case Study 2

Understanding Drivers of Transactional Net Promoter Scores (tNPS).   This study was done in a Major Healthcare Company

Objective:

  • To measure distances of dissatisfied customers from the satisfied customers and understand the drivers of dissatisfaction (higher distances) on individual basis from the information based on 1394 variables.

Approach:

  • Construction of scale based 1394 variables and 19000 satisfied customers.
  • Validation of the scale 17225 dissatisfied customers (abnormals).
  • Optimization of the scale – first level 125 variables, second level – 30 variables and third level-16 variables.

Root cause analysis (RCA):

  • Drivers with contribution ratios for all 17225 abnormals.

Success metrics:

Ease of use – the methodology is simple to execute and outputs are presented in straightforward manner.  Contextual flagging provides simple variable/root cause prioritization.

Speed:  Entire analysis performed in less than 2 business days.

Quality of insights: Distances and their drivers at individual level, ability to add new variables and new customers.

Table 2 below shows sample KDC output for this study.

kdc-table2
Table 2: Sample KDC output for tNPS study

Kanri Distance Calculator and Classical Methods

Problem of multivariate diagnosis with a different philosophy.

  • Classification Vs. Measurement.
  • Healthy or Target group Vs.  Abnormals.
  • Probabilistic Vs. Data Analytic.
  • Individualized root cause analysis to identify key contributors.
  • Localized root cause analysis if needed to identify key contributors.
  • KDC tests multiple simultaneous hypotheses.
  • Randomization (Not an issue with KDC).
  • KDC works well with both continuous and non-continuous data.

Creating a Management System with KDC

  • Single distance calculation run including all available variables.
  • Easily interpreted distance contributions on continuous scale – purely quantitative.
  • Distance matrix translates directly into highly individualized management system.
  • Includes Statistical Process Control built.

For more information visit: www.kanri-insights.com or contact Don Gray at don@kanri-insights.com .

Bio: Rajesh Jugulum is a Co-Founder of Kanri and has held several leadership positions in engineering, process improvement, robust design and data quality related areas including Cigna, Bank of America and Citi Group. Before joining financial industry, Rajesh was a lecturer and research engineer at MIT and was a Master Consultant at Six Sigma Academy.

Don Gray is a Co-Founder of Kanri and has held executive operational, engineering, financial, reengineering, and data management positions for companies including Cigna, Citi Group, Bank of America, and NEC. In these roles Don has led breakthrough performance improvements and been responsible for design, implementation, and ongoing execution.

Related


Sign Up

By subscribing you accept KDnuggets Privacy Policy