Cross-Industry Process Standardization for Text Analytics

doi:10.1016/j.bdr.2021.100274

2022-02-01

PRE

Christina G. Skarpathiotaki

Kostas E. Psannis *

DOI：10.1016/j.bdr.2021.100274

原文链接

原文求助

摘要

En 中文

We are living in a world where everything computes, everyone and everything is connected and sharing data. Going beyond just capturing and managing data, enterprises are tapping into IoT and Artificial Intelligence (AI) to create insights and intelligence in a revolutionary way that was not possible before. For instance, by analyzing unstructured data (such as text), call centers can extract entities, concepts, themes which can enable them to get faster insights that only few years back was not feasible. Public safety and law enforcement are only few of the examples that benefit from text analytics used to strengthen crime investigation. Sentiment Analysis, Content Classification, Language Detection and Intent Detection are just some of the Text Classification applications. The overall process model of such applications considering the complexity of the unstructured data, can be definitely challenging. In response to the chaotic emerging science of unstructured data analysis, the main goal of this paper is to first contribute to the gap of no existing methodology approach for Text Analytics projects, by introducing a methodology approach based on one of the most widely accepted and used methodology approach of CRISP-DM. (C) 2021 Elsevier Inc. All rights reserved.

Keyword:

Big data analytics

Advanced analysis

Artificial intelligence

Machine learning

Text analytics

Cross-industry processes