An ensemble learning model for predicting the intention to quit among employees using classification algorithms

Abstract

Employees are often more likely to use social media for job searching, which sometimes causes withdrawal behaviour. This study proposes an ensemble learning model for predicting the intention to quit (IQ) based on selected features, such as job Involvement (JI), organizational commitment (OC), activities on professional networking sites (APNS), and updating profiles on job portals (PJP). The Receiver Operator Curve (ROC) examines the model’s accuracy. We show the best relationship to predict intention to quit is between activities on professional networking sites and updating one’s profiles on job portals on social media. Seven classification algorithms of Gradient Boosting, Random Forest, K-Nearest Neighbour, Logistic Regression, Neural Network, Support Vector Machine, and Naïve Bayes are used to build the classification model. In addition, four combinations of the above-mentioned methods are used to construct an ensemble learning classification model. The performance comparison indicates that the combination of Gradient Boosting, Logistic Regression, and K-Nearest Neighbour produced the best ensemble learning model for predicting the intention to quit. The study’s contribution incorporates the stimulus organism response theory to predict the intention to quit through information gain, emphasizing social media features. Based on these features, the classification tool is utilized to identify between those who intend to resign and those who do not.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 33 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.