INfORmER : identifying local news articles related to an organisational entity and its remit

Abstract

Nowadays, the general public has the benefi t to access an immense amount of documents and articles on the internet. There are several organisations that on a daily basis must go through all local online newspapers, in order to check whether there are any articles that are relevant to their organisation in some way. This is a very time consuming and frustrating job that is prone to many human errors when done manually. It is for this reason that there is an ever growing need for a reliable and efficient article recommender system that takes care of the tedious job of going through local news articles and choosing which are relevant to an organisation based on its interests and remit. Throughout this report we investigate similar article recommender systems and also develop a system to help users in recommending articles from local newspapers without having to go through the hassle of reading all the local news articles. Hence, we created INfORmER, a system that uses an induction wrapper algorithm to scrape local newspapers and using several different pre-processing techniques, such as random oversampling combined with a hybrid ensemble classifi er to evaluate which articles to recommend. The use of `N' different classifi ers in a system is evaluated, therefore, tests were run on different number of classifi ers which were recorded so that the optimal number of classifiers and their combination was recorded. INfORmER also has the option to automatically send an email with the articles it deems relevant. During the evaluation of our system we found that with the tool that was implemented, INfORmER, surpassed the traditional methods of using the cosine similarity techniques. The developed system gives sufficiently good recommendations when compared to an actual real life data on human annotated dataset and recommends articles which the human annotator deemed to be irrelevant but infact were relevant.