Matjaž Vončina (2012) A presentation of web news from multiple sources. EngD thesis.
We built a website, where visitors can find and read current news from Slovenia from multiple sources. We presented news articles in groups of similar news to shorten the time to find important news and to spare visitors browsing of several websites. To achieve this we built a database of news and news processor. We developed a system to read and parse news from multiple sources, news normalization with lemmatization, weighting of words in the news and presenting the news using a vector space model. We used our model to calculate similarity between news, which enabled us to clusters similar news. We built a prototype website to display relevant news clusters.
Actions (login required)