yoganandaraol / employeeattritionclassification Goto Github PK

• They want to try a new initiative to retain employees.
• The idea is to use data to predict whether an employee is likely to leave.
• Once these employees are identified, HR can be more proactive in reaching out to them before it's too late.
• They only want to deal with the data that is related to permanent employees.

Once an employee leaves, he or she is taken an interview with the name “exit interview” and shares reasons for leaving. The HR Department then tries and learns insights from the interview and makes changes accordingly.*

• This approach is that it's too haphazard. The quality of insight gained from an interview depends heavily on the skill of the interviewer.
• The second problem is these insights can't be aggregated and interlaced across all employees who have left.
• The third is that it is too late by the time the proposed policy changes take effect.

The HR department has hired you as data science consultants. They want to supplement their exit interviews with a more proactive approach.

• You are given datasets of past employees and their status (still employed or already left).
• Your task is to build a classification model using the datasets.
• Because there was no machine learning model for this problem in the company, you don’t have quantifiable win condition. You need to build the best possible model.

Deliverable: Predict whether an employee will stay or leave.
Machine learning task: Classification
Target variable: Status (Employed/Left)
Win condition: N/A (best possible model)