A Supervised Approach to Analyse and Simplify Micro-texts

Analysis of micro-text presents a challenging task due to the incompleteness of its corpus in the domain of Natural Language Processing (NLP). Primarily, micro-text refers to the limited textual content in the form of letters and words, collected from various web-based resources. In the present paper, we are motivated to build a supervised model for analysing micro-text. The model assists in simplifying the texts and extracting the important knowledge from unstructured corpora. Additionally, we have prepared an experimental dataset to validate the proposed model. The validation process offers 94% accuracy to identify the micro-text from the unstructured corpus like Twitter. The proposed model helps to design various applications such as annotation system and prediction system for micro-texts in the future.