A generic framework for sentiment analysis: Leveraging opinion-bearing data to inform decision making

摘要

The increased exposure of the average citizen and customer to polarised content from various sources has been of significant consequence for companies and governmental organisations. Such content has, for example, served as a catalyst for violent uprisings and shifts in stock market prices. The collection and study of opinion have therefore become a necessity in many industries. Due to the vast nature of such data, manual approaches to this problem are no longer feasible. Several computational approaches have been proposed within the field of sentiment analysis, which successfully address many aspects of this problem, such as the classification of data into one of several sentiment categories. The research in the field is lacking, however, with respect to the integration and application of these techniques in practice, as well as their incorporation into the decision-making process of affected entities. In this paper, a generic framework for sentiment analysis is proposed, with a focus on facilitating the model development process for a user in a manner such that good performance may be achieved irrespective of the problem domain, as well as facilitating a flexible, exploratory analysis of model results in combination with existing structured attributes in order to gain actionable insights. The objective of the framework is to aid organisations in successfully leveraging unstructured, opinion-bearing data in combination with structured data sources to inform decision making.