Tweet sentiment analysis using logistic regression
Siddhant Kumar, Narinder Kaur, Kavita Kavita, A. Joshi · IET conference proceedings. · 2023
Twitter is like a blogging platform where any users can send status and messages, called ``tweets,'' to other people. Provided a large database of so-called emotions, its separation can be done by supervised learning. In order to making supervised learning models, segmentation algorithms need a large set of data with proxy labels. However, categorized data is often hard and costly to get, promoting an unrevealed reading interest. This type of reading uses label less data to be compatible with details given with labeled data in the training program; therefore, it is especially meaningful in programs that include analyzing tweet sentiments, in which a large amount of unspecified data is accessible. While attractive, the slowpaced reading of the tweet emotional analysis is new. We present an exhaustive survey of slow-tracking methods used to tweet segregation. Such methods contain graph-based methods, conclusions and subject-based methods. Comparative study of supported algorithms in self-training, co-training, title modelling, and remote monitoring expose their bias and shed light the factors an expert should consider in real-world systems.