The growth of user-generated material through virtual entertainment has made evaluation mining a difficult task. Text data are having data collection opinions on products, trends, and legislative issues as a microblogging platform. Feeling analysis is aprocess for dissecting the mentality, feelings, and assessments of numerous individuals about something, and it is frequently applied on tweets to deconstruct common opinion on news, tactics, social advances, and personalities. Assessment mining can be done without personally going through tweets by using Machine Learning models. Their findings could aid state-run administrations and enterprises in implementing plans, used to recognize feelings by categorizing tweets as happy or sad. The proposed casting a ballot classifier (LR-SGD) with TF-IDF generates the most ideal outcome with 79 percent precision and 81 percentF1 score, according to an inside and out relative presentation research. To confirm the stability of the suggested approach on two additional datasets, one parallel and the othermulti-class dataset, and to get positive results.