News classification using light gradient boosted machine algorithm

News classification is a complex issue as people are easily convinced of misleading information and lack control over the spread fake news. However, we ca n break problem spreading news with artificial intelligence (AI), which has developed rapidly. This study proposes model using light gradient boosted machine (LightGBM) algorithm. The analyzed two feature extraction techniques, count vectorizer Tfidf vectorize r compared deep learning long - short term memory (LSTM). experimental evaluation showed that all LightGBM models outperform LSTM. best Li ghtGBM, achieves an accuracy value 0.9933 area under curve (AUC) score 0.9999.