Resume2Role: Intelligent Resume Parsing and Domain Prediction Using NLP and Classification Models

Accurate and effective processing of a high volume of resumes poses significant challenges in the dynamic field of digital recruitment. Traditional manual screening methods are highly inefficient, inaccurate, and susceptible to biases. To overcome these limitations, this paper introduces the Resume2Role solution, which utilizes machine learning algorithms and NLP techniques to automatically parse and classify resumes based on job roles. The process involves the use of Python for text extraction from resumes formatted in either TXT or PDF form. The data is cleaned and normalised using preprocessing methods such as lemmatisation, lowercasing, and stop-word removal. Term Frequency-Inverse Document Frequency (TFIDF), which successfully transforms unstructured resume data into structured numerical representations, is then used to vectorise the processed text. Several models were evaluated for classification, and the Random Forest Classifier produced the best accuracy in identifying the appropriate job domain (e.g., IT, HR, Finance, Sales). Further-more, methods such as oversampling were used to address class imbalance, which improved the model’s capacity for generalisation.The creation of a completely functional Flask-based web application, which enables real-time resume uploads, parses important information (such as name, email, phone number, and skills), and provides rapid domain classification with job recommendations, is a significant contribution of this effort. The application performed well in a variety of professional areas and résumé formats. The system’s efficacy is validated by evaluation measures such as confusion matrix analysis, recall, accuracy, and precision.