Main Article Content
Tourists want to know the good and bad aspects before going to tourist place of a city or country. Often they search in social network websites to read previous visitors opinions. However, due to the large amount of reviews tourists find it extremely difficult to obtain useful opinions to make a decision about destinations, accommodation, restaurants, tours, and attractions. Unfortunately, some reviews are irrelevant and become noisy data. It finds difficult for the people to analyze the reviews. In this situation Aspect-based sentiment analysis summarizes likes and dislikes of the people from reviews. The main purpose of the project is to collect the reviews from various sources and preprocess it to find the polarity and categorize each tourist reviews into different sentiments and the aspect term related with the each tourist reviews are being extracted to check the accuracy using models. This approach adopts language processing techniques, policies, and lexicons to address several sentiment evaluation challenges, and convey summarized results. According to the said results, the thing extraction accuracy improves considerably when the implicit elements are considered. Also, when using the identical dataset, the pro- posed approach outperforms gadget mastering methods that use Naive Bayes (NB). However, the use of those lexicons and guidelines as input capabilities to the NB version has achieved better accuracy.