Social Content Opinion Mining by Grey Wolf and Random Forest Using Pattern Features

Abstract

There has been a rise in the amount of content created and shared through social media, making it necessary to find methods of analysing the opinion and sentiment of the public as a series of patterns. The new method in this study is a hybrid model of social content opinion mining, which combines Grey Wolf Optimisation (GWO), Random Forest (RF) and pattern-based feature extraction models. The framework leverages linguistic patterns, n-grams, TF-IDF representations, as well as sentiment and negation patterns to extract informative textual features. GWO is used to optimize processing of the features, meanwhile the RF is used to classify the sentiment as positive, negative and neutral. The accuracy, precision, recall, F1-score and computational efficiency are used to measure the model performance against conventional machine learning classifiers. The proposed approach aims to improve the reliability of the classification, reduce feature dimensionality, and enable robust social media opinion analysis.