Author Email: firstname.lastname@example.org
Microblogging sites have gained very importance nowadays. We can know sentiments of people and predict things from their sentiments. In this paper, we are detecting public safety events by analyzing the sentiments via Microblogging text messages. In this the ratings have been given i.e, positive, negative or neutral to the tweets, where there is a pre selection of topics causing riots and also some random tweets based on it. In this we will calculate the subjectivity and polarity confidence to capture the sentiments. Three classes of tweet are considered here that are positive, negative and neutral. There are predicated topics that have high chances of riots. For a particular event related with that topic we will analyze the tweets about that topic and also the tweets taken for analyzing purpose will be taken from the microblogging site users whose locations are same as that of the event. If the tweets negative comments are more than a particular threshold then the event will be said that it requires public safety.
Microblogging, Classification, Sentiment
Microblog mining has become a very important research field .It can be used to know commercial and political view of the people. A sentiment analysis can be considered as a two type classification problem. The negative and positive classes are considered under it.
The vector space model is very much used for text representation , in this it takes a vector to represent a document. An important preprocessing step for many tasks in text mining, is to transform natural language message text to numerical vector which includes cluster analysis and sentiment classification. 
In the previous paper, the support vector machine has been taken as the classifier. It is because its excellence in solving two-class classification problem. In this, they included three dictionaries from Dalian University, self-established dictionary and CNKI-Hownet emotional .[2,4,5]
The analysis of the sentiments can help in various fields for reviews of the people to predict things from the future like events that need public safety. We can know that how the sentiments play an important role in different field like politics business and many more.
We have taken 14000 tweets for sentiment analysis of 10 different topics .The 10 topics include some riot causing topics and some random topics.By considering tweets of particular area for example Gujrat and also by choosing the topic Padmavat,we will perform sentiment analysis on tweets . The total number of negative comments will me converted into percentage. If the percent of total negative comments goes above the threshold then the Events related to Padmavat in Gujrat will need Public Safety Arrangements.
149 total views, 2 views today