DISCUSSION
Text features are extracted from the text to simplify the classification of text data. Through the process of feature extraction, we reduce the dimensionality of the text, and thus eliminate irrelevant features from text data. As a result, classifiers become more accurate and the reduction of noise is improved. Statistically, the difference between these two methods (Nagashri and Sangeetha 2021) is not statistically significant. Term frequency–inverse document frequency (TFIDF) as well as count vector techniques is used separately for text preprocessing.Six Machine learning algorithms namely passive-aggressive classifier (PAC), naive Bayes (NB), random forest (RF), logistic regression (LR), support vector machine (SVM), and stochastic gradient descent (SGD) are thought about utilizing assessment measurements like precision, accuracy, recall, and F1 score, The outcomes have shown that the TFIDF is a superior text preprocessing method. PAC and SVM calculations show the best presentation for the considered dataset. (Choudhary et al. 2021)manages an audit of existing Machine Learning algorithms Naïve Bayes, Convolutional Neural Network, LSTM, Neural Network, Support Vector Machine proposed for recognizing and decreasing phony news from various online media stages like Facebook, whatsapp, twitter, and so forth This survey gives a far reaching point of interest including information mining viewpoint, assessment measurements, and agent datasheets. (Smitha and Bharath 2020)paper represents model and system to distinguish counterfeit news from news story with the help of Machine learning and Natural language preparing.Seven distinctive Machine learning Classification algorithms are prepared to group news as phony or genuine and are analyzed thinking about precision, F1 Score, review, accuracy and best one is chosen to fabricate a model to arrange news as phony or genuine. (Jiang et al. 2021)proposed our novel stacking model which accomplished testing accuracy of 99.94% and 96.05 % individually on the ISOT dataset and KDnugget dataset. Assessed the exhibition of five Machine Learning models and three deep learning models on two phony and genuine news datasets of various sizes with hold out cross validation. Different classifiers are used for identifying the fake news ,and the approach is executed on two datasets of phony and genuine news. In the wake of playing out the examination, it is seen that Passive-Aggressive Classifier gives the best outcome(Gupta and Meel 2021).