An Explainable NLP-Based Framework for Placement Email Understanding using TF-IDF and Linear SVM

Authors:

HARI KRISHNAN A¹ , Mrs. PRADEEPHA S² , Dr E. MARIAPPAN³ , Dr M. KALIAPPAN⁴

Student, Department of Artificial Intelligence and Data Science, Ramco Institute of Technology, Rajapalayam, Tamil Nadu, India¹

 Assistant Professor, Department of Artificial Intelligence and Data Science, Ramco Institute of Technology, Rajapalayam, Tamil Nadu, India.²

Associate Professor, Department of Artificial Intelligence and Data Science, Ramco Institute of Technology, Rajapalayam, Tamil Nadu, India.³

Associate Professor, Artificial Intelligence and Data Science, Ramco Institute of Technology, Rajapalayam, Tamil Nadu, India⁴

Abstract

In today’s placement and recruitment environment, students receive a large number of emails containing important information related to hiring processes, assessment schedules, and job role requirements. Identifying relevant placement emails within this continuous stream remains a challenging and time-consuming task for many students. To address this problem, this paper presents an automated and explainable placement email classification framework built using core Natural Language Processing and Machine Learning techniques. The proposed system applies NLP-based text preprocessing and a Linear Support Vector Machine classifier to distinguish placement-related emails from general communications. The preprocessing pipeline includes text cleaning, tokenization, stopword removal, and lemmatization, followed by feature extraction using TF-IDF representations. This approach enables efficient handling of unstructured email content and improves classification performance. In addition to classification, a lightweight information extraction mechanism is incorporated to identify essential details such as company name, job role, and interview date from relevant emails. Experimental results demonstrate that the proposed model achieves high accuracy, precision, recall, and F1-score, indicating its effectiveness in filtering placement-related communications. The system ensures transparency and interpretability through the use of classical machine learning methods, making it suitable for real-world deployment in academic environments. Overall, the proposed framework helps students efficiently identify important placement emails and reduces the risk of missing critical opportunities.

Keywords:

Campus placement, natural language processing, support vector machine, email classification, TF-IDF, information extraction, explainable systems