One of the most significant communications for knowledge exchange is social media. The detrimental effect of false information on everyone using social networking sites is growing at the same time. For countries all around the world, the growth of the travel and tourism sector is essential to their economies. The sharing of tourist reviews on social media is exponentially growing at the same time. This sharing impacts psychological decisions regarding travel. Over the decades, a variety of social networking methods for analysis have been developed. However, they came into a number of issues as a result of short-text social media comments, which left them with a serious shortage of data. This research proposed a recommendation system for social networks to identify if the content is false or true using a hybrid SVM classifier using the LSTM in order to get around these challenges. The recommended method begins with gathering data using a tool like Kaggle and producing a dataset. Pre-processing of the acquired data includes segmentation, stop word elimination, lemmatization, and spelling correction. Following pre-processing, a count vectorizer is used to extract features from the data and transform them into binary. A hybrid SVM classifier using the LSTM model is used to further classify the collected features. Data that has already been processed and contains real information is compared to predicted data. It will be clear if the message is true or false if the predicted data matches the pre-processed data. To attain greater performance, the recommended approach is put into action. The overall projected result of the hybrid SVM classifier using the LSTM approach used for recommendations is superior to the current methods. The hybrid SVM classifier using the LSTM model was more accurate at identifying untrue from real communications.

#