Design and Implementation of a Job Skill Entity Recognition (JSER) for Custom Entity Annotation using Trie Structure

Abstract

Abstract—Named Entity Recognition in computer science is tricky because it uses terms and needs detailed annotations. Even though Named Entity Recognition has gotten better in uses it is not widely used in academic computer science. This paper presents JSER, a method that uses Natural Language Processing techniques to find and classify both known and unknown entities. Our approach uses Part-of-Speech tagging to pull out entities from unstructured text. We make annotations better and faster with techniques like tuning models adjusting thresholds and adding more data. JSER seems promising for tasks, like finding information analyzing text automatically and managing knowledge.