Among text-oriented applications, Natural Language Processing (NLP) plays a significant role in managing and identifying a particular text data. NLP is broadly used in text mining domains as well. In general, text mining is the process of combining diverse techniques to characterise and transform the text. Hence, the syntactic information and semantic information are utilised together in the NLP model to assist the process of analysing, or extracting the text. On the other hand, the text mining model is examined by different standard measurements, which also vary concerning text objectives or applications. Due to the advent of machine and deep learning models, text mining has become the hot research area used in various domains like classification, recognition, sentiment analysis, and speech-related topics. Though the models are not time-consuming and effective, certain factors are considered for further enhancement of these models. Thus, this survey paper elucidates the evaluation metrics used in text mining approaches by deploying standard algorithms. It explores the literature work on the formerly implemented text mining approaches for analysing the evaluation metrics in text mining. In addition to this, the proposed model to perform text mining in every survey paper is analysed. Further, it provides the algorithmic categorisation of existing research works, discussion on different datasets used with the consideration of various evaluation metrics, and finally categorises the metrics used for analysing the performance of such text mining approaches. This survey work also illustrates the merits and demerits of the existing text-mining approaches. Finally, the research gaps and challenging issues are given to direct future work.
