A Consequential ML Intersection For Figurative Meaning Detection Using Bengali Linguistic Dataset

Abstract

Each human feeling may presently be connected to the writings we see on a day by day premise on different online stages. Such stages gives client inside the autonomy to share, connected with another clients suppositions and announcements in different terms and themes. With which case, it is basic for such settings to have had a framework that can recognize between honest to goodness feelings and threatening vibe. A prime illustration of how people employ imaginative linguistic methods in social communication is humor. In addition to exchanging information or conveying implicit meaning, humor fosters interpersonal connections among individuals who are exposed to it. It can assist people in separating themselves from tensions and assist them in finding the humorous side of issues. Additionally, it aids in controlling our emotions. Additionally, the ways in which people create humorous content offer information into their genre and character attributes. Which is why we have chosen to center our endeavors on distinguishing one of the foremost interesting sorts of all time: humor. For tasks like predicting sentiment polarity at the document level, we think an n-gram model in conjunction with latent representation will result in a more appropriate embedding. For the purpose of creating a quick and effective embedding for brief text segments, our suggested embedding mixes n-gram encoding with such a latent model. We utilized Direct Relapse, Choice Tree, Arbitrary Timberland, KNN, Multinomial Credulous Bayes, RBF SVM, and Direct SVM with in division strategy to get the most extreme exact findings. The findings which may well be fundamental in case of deciding between the cases between the conceptional intrigued for those crossing points of categorization of the animosity and veritable commenters who proposed for joke or per say humors conjunctions. The best accuracy was achieved by the Unigram was MNB ( Multinumia Naiv bias) at 93%, by biagram it was same as 93% by MNB, with the triagram the feature with the MNB at 94% accuracy in toal.