Real Time Data based Automated Resume Classification and Job Matching using SVC, Jaccard Index and Cosine Similarity

With ever increasing demand for online recruitment processes, traditional hiring processes have taken a leap into an automated hiring procedure. Document classification is fundamentally implemented using Convolution Neural Networks (CNN), Deep Learning, Natural Language Processing (NLP) etc. This paper presents improvement to the current approach by introducing real-time data scraped from LinkedIn. In previous papers, either cosine similarity or Jaccard coefficient were used. However, in this paper both cosine similarity and Jaccard index are used to get most optimized result. Moreover, the accuracy of this paper is more than the previous papers. This paper proposes a NLP based automated approach for resume classification primarily for technical roles using SVC algorithm and resume matching with a given job description using text distance and token based similarity. The machine learning model is then trained using this data and a comparative analysis of various algorithms is done using accuracy, precision, recall and F1 score to get the best suitable classification algorithm.