With its remarkable speed and agility, social media has caught everyone off guard. When new conditions arise, whether they are social, political, or current affairs-related, people’s sentiments are conveyed globally through their assistance, which makes them good candidates for sentiment mining. Any firm looking to analyze and improve its goods and services can benefit greatly from the application of sentiment analysis. In order to perform a sentiment analysis on their individual clients, the airline industry finds it far simpler to obtain feedback from perceptive data sources like Twitter. It is impossible for customers who seek to know who is who and what is what in daily life to obstruct the advantages of Twitter sentiment analysis. The sentiments contained in tweets led to their classification into three classes: positive, negative, and neutral. Furthermore, utilizing accuracy, precision, sensitivity, and specificity as performance criteria, a range of ML classifiers were assessed. It was also looked into how feature extraction methods, such as word2vec, TF, and term TF-IDF, affected the accuracy of classification. Furthermore, analysis was conducted on the chosen dataset from the Kaggle Web site using a hybrid ant colony optimization with artificial neural network (HACOANN). According to the results, the suggested HACOANN outperforms the other classifiers. The HACOANN can extract features using TF and TF-IDF features with an accuracy of 93.17, respectively. Ensemble classifiers outperform non-ensemble classifiers in terms of accuracy, as evidenced by the findings. TF-IDF is a better feature extraction method for machine learning classifiers, according to additional experiments. TF and TF-IDF feature extraction outperforms word2vec feature extraction. Compared to machine learning classifiers, the HACOANN achieves a lesser accuracy.