An Approach for News Trend Analysis Based on Online Bangladeshi Newspapers

Abstract

In the present world, there are many people who are talking in bangla language. We the Bangladeshi people use this language as a mother language. In our system we work with this bangla language. There are many online bangla newspapers. We collect the articles from these online newspapers using web crawler program. In our case we use the prothom-alo online newspaper. Category wise we collect those articles. The main task of our thesis is to find the trend of the newspaper. That is, what type of things the daily newspaper having every day? What‟s the number of positive article and negative article? For this purpose we have gone through the newspaper‟s articles to find some method or rule that can give us the view about the article. We have used step by step process. We try to collect the feedback for each of the bangla words from the general people. Here feedback is positive, negative and neutral value. Then using these feedbacks and proposed rules we try to find the weight of the sentences and articles. For weighting each word we have collected the feedback that I mentioned earlier. But for finding the weight of the sentences and articles we use some set of rules. In this thesis approach we have faced many problems with weighting the words, sentences and articles. This weighting technique is the core technique here. We have analyzed many newspaper articles to finally come with a technique called “addition rule” to weight the sentences. We have mentioned many problems and ambiguities here. Through our proposed algorithm we have tried to solve those problems and show the trend of the newspaper.