2. Case Study
The dataset consists of 1470 observations of 15 variables which are described below:
Attrition - Whether the attrition happened (Yes=1) or not (No=0). This is our independent
variable which we are trying to score.
Age
BusinessTravel – relates to travel frequency of the employee
Satisfaction_level – relates to environment satisfaction level of the employee (1-Low, 2-Medium,
3-High, 4-Very high)
Sex – gender of the employee
JobInvolvement– relates to job involvement level of the employee (1-Low, 2-Medium, 3-High, 4-
Very high)
JobSatisfaction – relates to job satisfaction level of the employee (1-Low, 2-Medium, 3-High, 4-
Very high)
MaritalStatus
MonthlyIncome
NumCompaniesWorked – how many companies has the employee worked upon
OverTime – whether the employee works overtime (Yes / No)
TrainingTimesLastYear - number of trainings received by the employee last year
YearsInCurrentRole – number of years in current role
YearsSinceLastPromotion - number of years since last promotion
YearsWithCurrManager - number of years with current manager
3. Business Problem and Methods to
control it
The main objective of the case study/business is to retain the employees of the organization
Employee Attrition is unpleasant to any organization.
The main objective of the case study/business is to reduce attrition rates of the organization
using analytical methods.
Analytics can help organizations control employee churn through predictive models which can then
be
used for developing retention strategies.
Business Problem and Methods to
control it
5. Packages used in R
Following Packages were used in order to analyze and work on
the
business problem and to come with a retention strategy.
Usdm, Fselector (for information gain to pick up important
variables
from the data), RandomForest,
e1071 (svm), tree (decision tree)
Packages used in R
7. Machine Learning Algorithms with it’s
accuracy value
SVM: with 85.03% accuracy
Decision Tree with 83.30% accuracy
Random Forest with 86.40% accuracy
Here we can see that Random Forest works best for the
given data with 86.40% accuracy.
8. CONCLUSION
Employee retention is important to an organization as it reduces the high training cost
and saves the time of an organization.
After doing a deep analysis with predictive modeling techniques, I have come up with
some Employee Retention Strategies:
Job satisfaction level and environmental satisfaction level impacts positively for retaining
the employees. Higher Job satisfaction level and environmental satisfaction level
increases the chances of employee retention.
Married Men’s and Women have more stability than single and divorced men and Women.
Hence, hiring married Men and Women is beneficial for this organization.
Low income with higher years in the organization increases the chances of attrition.
Timely promotions and various perks can be provided to retain the employees.
It’s always better to provide extra perks or promotions for the existing employees rather
than hiring new employees due to attrition, as Hiring new employees costs very high and
it is also time consuming as the new employees may not provide the quality of work
immediately when compared with existing employees in the organization.