Authors :
Rajat Keshri, Dr. Srividya P
Volume/Issue :
Volume 5 - 2020, Issue 4 - April
Google Scholar :
http://bitly.ws/9nMw
Scribd :
https://bit.ly/3ca2iqa
Abstract :
Many employees leave the organisation or
the company depending on various factors. This effects
the growth and production of the company in many
ways. The companies and many MNCs use machine
learning methods to predict a turnover of workers to
solve this problem. Such predictions help the company
in success planning and employee retention. The dataset
used in this paper for the above problem comes from
the Human Resource Information Systems, which are
usually different for different companies. Due to the
differences of the dataset in different organisations, it
results to a noisy data which makes the models to overfit or produce inaccurate results. This is the main issue
which this paper focuses on, and one which has not been
discussed traditionally. This paper discusses a new
algorithm called the LightGBM, released by Microsoft
in 2017. Here, we compare LighGBM with other
existing algorithms. Data from the dataset is used to
compare LightGBM and other classification algorithms
and show LightGBM’s high accuracy of prediction.
Keywords :
Machine Learning; Supervised Classification; Retention Prediction; Gradient Boosting
Many employees leave the organisation or
the company depending on various factors. This effects
the growth and production of the company in many
ways. The companies and many MNCs use machine
learning methods to predict a turnover of workers to
solve this problem. Such predictions help the company
in success planning and employee retention. The dataset
used in this paper for the above problem comes from
the Human Resource Information Systems, which are
usually different for different companies. Due to the
differences of the dataset in different organisations, it
results to a noisy data which makes the models to overfit or produce inaccurate results. This is the main issue
which this paper focuses on, and one which has not been
discussed traditionally. This paper discusses a new
algorithm called the LightGBM, released by Microsoft
in 2017. Here, we compare LighGBM with other
existing algorithms. Data from the dataset is used to
compare LightGBM and other classification algorithms
and show LightGBM’s high accuracy of prediction.
Keywords :
Machine Learning; Supervised Classification; Retention Prediction; Gradient Boosting
