Location-Specific Sentiment Analysis with Xxxx- ford CoreNLP Sample Clauses

Location-Specific Sentiment Analysis with Xxxx- ford CoreNLP. After preliminary computations and comparisons of tweets’ sentiment score, the researcher notices that sentiment-based event detection model would always fail. Considering that the dataset contains tweets posted across the United States, the researcher attributes the ineffectiveness of the model to the excessive locations covered by the dataset. Xxxx’s study validates that social event detection should sometimes be restricted to certain geographic areas to avoid the diluting effect of posts made in other places [5], and thus a location- specification of the dataset is introduced. The location filtering process is added to the pipeline and inserted between Step1. Language Filtering and Step3. Sentiment Analysis. Figure 4.3: Selection of Tweets about New York Ever since the state of New York had its first confirmed case on March 1st, 2020, both confirmed cases and death cases have been escalating in the state and NY soon becomes the epicenter of the nation. In early April, New York even has more confirmed cases than any other country in the world. Therefore, hoping to achieve event detection in the state facing the most severe threats of the outbreak, this research focuses on tweets related to New York. As shown in Figure 4.3, tweets will be filtered by their location information. To begin with, since some tweets have coordinate data embedded if the Twitter user chooses to attach their location to the tweet, all data will be categorized into a) have coordinate information, and b) not have coordinate information. Secondly, group (a) will be broken into two subgroups