Data Visualisation and Improving Accuracy of Attrition Using Stacked Classifier

  Deep Sanghavi,  Jay Parekh,  Shaunak Sompura,  Pratik Kanani

In any organization, managing Human Resources is an important task. Loss of employees lowers the overall productivity of the team and is also financially costly. Attrition of employees leaves behind a void that is costly to fill [2]. Machine Learning can be utilized for predicting an employee’s attrition. This paper evaluates the algorithms which can be used to predict the employee attrition on the IBM HR Analytics Employee Attrition & Performance dataset [1] taken from Kaggle with 35 attributes like Job Satisfaction, Percentage Salary Hike, Work Life Balance, etc. taking into consideration all aspects right from distance from home to the number of working hours. In order to predict Attrition, which is the dependent variable, classification algorithms under supervised learning are used. This paper provides the most optimal solution with the Stacked Classifier, an ensemble model which in this case averages Adaptive Boosting, Decision Tree Classifier and Support Vector Machine algorithms ultimately giving a high accuracy of 90.65.

Keywords- Stacked classfication, IBM,attrition,data visualisation

Unique Identification Number - IJEDR1804054
Page Number(s) - 284-293
Pubished in - Volume 6 | Issue 4 | November 2018
DOI (Digital Object Identifier) -   
Publisher - IJEDR (ISSN - 2321-9939)

  Deep Sanghavi,  Jay Parekh,  Shaunak Sompura,  Pratik Kanani,   "Data Visualisation and Improving Accuracy of Attrition Using Stacked Classifier", International Journal of Engineering Development and Research (IJEDR), ISSN:2321-9939, Volume.6, Issue 4, pp.284-293, November 2018, Available at :http://www.ijedr.org/papers/IJEDR1804054.pdf