Ensemble Learning Based Voting Model for Dynamic Profile Classification and Project Allotment
Suhas Tangadle Gopalakrishna, Vijayaraghavan Varadharajan · 2018
Every year, lakhs of students right from college enter professional life through various recruitment activities conducted by the organization.The allotment of projects to the new recruits, carried out by the HR team of the organization is usually a manual affair.It is a time consuming and a tedious process as it involves manually opening each resume and analysing it one by one in order to assign a project.Companies round the globe are leveraging the power of artificial intelligence and machine learning to increase their productivity.In this paper, we present one such use case wherein artificial intelligence is leveraged by the organisation in allotment of projects to the new recruits.Current machine learning tools help in the allotment of projects to a few known popular domains on which the classifier has been trained explicitly.We tackle the problem with an ensemble learning based voting classifier consisting of 5 individual machine learning classifiers, voting to classify the profile of the candidate into the relevant domain.The knowledge extracted from the profiles for which there is no majority consensus among the individual classifiers is used to retrain the model.The proposed model achieves a higher accuracy in classifying resumes to proper domains than a standard machine learning classifier which is solely dependent on the training set for classification.Overall, emphasis is laid out on building a dynamic machine learning automation tool which is not solely dependent on the training data in allotment of projects to the new recruits.