Attrition Predictor in BPO Industry
Authors: Almarian Pailden
Abstract: This application used Random forest method to determine the chances of New hire employees to churn in less than 3 months. 80% of the data is used as the training set while 20% utilized as test set.
The impact in accuracy of each variable is also included in the view. Error rate will improve depending on the available data and setting of training and test set.
Random forest create multiple decision trees and combined together to get a more accurate and stable prediction. It is a popular machine learning algorithm.
This is a good model for predicting attrition in BPO industry.
Full Description: In BPO industry, the companies forced to constantly finance the training of new hire because of the employee retention in this industry is too fluid. Thus, it is essential to determine if the applicant will stay longer or not with the company.
This app will help the company specific to BPO industry to predict the chances of each applicant to resigned. Employees that leave the company at least within 3 months or still on training will hurt the company financially since revenue and cost will be greatly affected. With this app, hiring managers or recruitment specialist will have the a tool or indicator to hire applicants based on the company's historical data.
By hiring employees with longer retention rate, that would help generating revenue while re allocating some training budget to other productive activities of the Company.
The user just need to input exam score, interview scores and basic profile information of the applicant.
By clicking the "Submit" button, it will automatically calculate and provide the result.
Keywords: random forest, prediction, caret, attrition, hiring, BPO
Shiny app: Attrition Predictor
Repo: GitHub - pretty1020/attrition_predictor
RStudio Cloud: link to R Studio Cloud
Thumbnail:
Full image:
