A Approach to parse resume and recommend Job Modified K-mers and Firefly algorithm (FFA) in Resume Parsing

Abstract

Named Entity Recognition (NER) thrives as key component in Natural Language processing (NLP) system for information retrieval, question answering, relation extraction, machine translation and text summarization. NER aims in classifying the text from corpus of document to certain pre-defined categories such as location, Name, person name, month, date and so on.  The Earlier NER systems yields a considerable success in performance, however the human engineering cost to design domain specific rules and features are high with variations in precision outcomes and speed as well. Hence to address these issues, the NER is developed based on NLP using Text segmentation and resume matching phenomena on the resume files. After Text pre-processing based on the entities, the Semantic relations between the entities were discovered using text segmentation using lexico-syntactic patterns as query. In resume matching aspect, the hybrid of K-mers algorithm and Firefly algorithm (FFA) implemented to efficiently fetching out matched feature representation. The K-mers algorithm, identifies all K-mers prevailing in more occurrences as sub-strings, utilises dictionary, that stors all observed K-mers for reducing the memory requirements. The K-mers were checked for matching with match criterion on all the resume profiles. The Global best solution gained through FFA to address the optimisation issues and for parallel execution.  This FFA utilises randomness property with guess in searching mechanism and optimisation attempts to enhance solution quality, matching resume profiles, in high precision rate, using similarity scores. Based on the similarity Scores, of each resume, the related shortlisted candidate profiles are obtained using NER approach. The comparative assessment and internal outcomes in entity recognition for resume matching, explicated the efficiency of proposed NER model.