Machine learning is used to classify and mine tweets on multiple levels

Abstract

Research on sentiment analysis can be found under the discipline of Natural Language Processing (NLP). It is helpful in determining the feelings and meanings that lie beneath the surface of a piece of literature. When attempting to extract sentiments from Twitter data, you face a particularly unique combination of obstacles as a result of the platform's unstructured format, its relatively small size, and the presence of slang, misspellings, and acronyms. The vast majority of researchers analysed the results of multiple sentiment analysis machine learning algorithms and compared their findings. However, merging these methodologies is a topic that has received insufficient attention in the relevant academic literature. According to the findings of this research, mixing many different machine learning algorithms together produces superior outcomes compared to doing it individually. Due to the unprocessed nature of the tweets, this study employs a wide range of preprocessing techniques in order to generate data that can then be used in machine learning classification models. An inquiry into the potential benefits of combining the machine learning algorithms known as KNN and SVM is the primary focus of this work. When making an analytical observation, the classification accuracy and F-measures for each emotion class, as well as the average of those values, are used as inputs. The results of the evaluation show that the proposed hybrid approach is more accurate and has a better F-measure than individual classifiers.