The media industry is one of the multiple segments that has very nicely made its way from print to digitization age which today we live in. This shift also caused the change in fake news as a global problem where deceiving information can be easily created and delivered. This paper looks at what to do about this. In light of increased disinformation, our research centers on the development of a model to assess credibility in news articles. An evaluation process is then performed; using text vectorization involving the Bag of Words technique with both Count Vectorizer as well as TF-IDF vectorizer. We then use these representations to train two classification algorithms Multinomial Naive Bayes Classifier and Passive Aggressive Classifier for the fake news detection. An analysis of the effect on model performance resulting from text pre-processing is also part of this study. It uses 67.7% of the curated information for training and makes predictions on test data (untrained raw data) from remaining 33.3%. In particular, this model achieves a high efficiency of 93.78% when optimal measures are taken. Although this good result proves that the proposed method is excellent at distinguishing between true and false news.