OPINION MINING IN RESOURCE-POOR LANGUAGES: TECHNIQUES AND CHALLENGES

Abstract

Human emotions automatically arise from the experience of gratification, sorrow, failure, or success. They are a reflection of behavior and cerebral thoughts. Opinion mining (OM) is crucial in all fields of work today, such as cyber security, smart cities, medicine, education, e-commerce, governance, and agriculture. Consumer understanding is required to increase business profits and strategy building as they are valuable assets. OM is an inevitable task to understand users' beliefs. Ample resources are available for the English dialect, but there is not enough for code-mixed and non-English dialects. This survey focuses on the OM of resource-poor languages. It depicts the various techniques, datasets, and polarities of emotion mining used by scholars in previous research. It illuminates the various practical challenges with suitable examples and possible solutions to develop an intelligent system. It will help to make noteworthy strides in discovering and filling the OM research gaps. We have also analyzed the applications of OM. Our study will hopefully guide future researchers, academicians, industrialists, and learners in accomplishing their tasks.