The Text Analysis Framework for Interactive Statistical Classification Service

Abstract

Statistical classification serves as the basis of national statistics such as the population census and economic census. Not only is it being used as the main basics for national policy establishment and execution (e.g SME support and employment stability), it is also used as a major classification criterion from the perspective of corporate analysis services such as economic, company, investment. Therefore, we innovate the statistical classification system by AI approaches, and develop accurate and timely interactive statistical services. In this paper, we introduce the methods for standardizing unstructured data processing, building a statistical classification system based on machine learning, and establishing an interactive statistical service. With this system, it is possible to implement reliable classification models in a minimum amount of time and without a complex knowledge building, only using machine learning-based methodologies and accumulated public and internal examination data. In addition, it can be implemented a web service that provides functions such as classification auto classification coding and auto-completion, and also provides a knowledge search that serves responses to the classification knowledge.