Big Data Analytics Based Sentiment Analysis Using Superior Expectation-Maximization Vector Neural Network in Tourism

Tourism experience shared through social media has become a highly influential source of information and has a multi-faceted impact on tourism. With the vast development of the Internet, text data has become one of the leading formats of big tourism data. Text analytics of such data has great potential to express tourists' opinions effectively. Sentiment analysis is an essential component of tourism big data because it can detect positive and negative opinions in texts. Tourist comments are essential for the development of tourism but still, the number of comments complicates the analysis of essential aspects of the comments by the owner. Big data-based sentiment analysis is one of the most challenging problems globally, and the amount of data is enormous. To resolve this problem, the proposed big data approaches can help detect new words, especially with sentiment analysis and detection of proper nouns and emotional words useful for subsequent tasks as word vectors. The proposed system follows the three steps: text analysis and cleaning, Word vector similarity analysis, and final sentiment classification. First step is used to remove the noise of the data and detect the symbols. The next step is the ID3 (Iterative Dichotomiser) Maximum Word Vector Dimensionality Posteriorl method, which discovers all travel review corpora's main problem and uses it to enrich the vocabulary vector representation of words in context. Attention mechanisms are used to learn words and the overall meaning of different weights text attributes. According to the classification, the final Superior Expectation-Maximization Vector Neural Network (SEMVNN) is used for classifying sentiment analysis level. The SEMVNN method gives accuracy, time complexity, precision, recall and F-measure values to achieve better results than the previous system.