This frameworks allows the user to query news articles based on their sentiment.Sentiment Analysis is a technique used for determining the overall emotional reaction to a document, interaction or an event. In our case with news articles, by analysing the sentiment of the article's text, we can tell whether it expresses a positive, negative or neutral feeling to the reader.
The are two main modules in this framework. The engine and the crawler.
A RESTfull service for querying through the Elasticsearch engine implemented in python: server.py
After receiving a news article from the crawler, it performs the sentiment analysis and then indexes the analyzed article using the functionality of the ElasticSearch engine.
The Sentiment analysis is performed via the sentiment_analysis.py python file. It uses VADER a lexicon and rule-based sentiment analysis tool that is specifically attuned to sentiments expressed in social media, and works well on texts from other domains.
python server.pyThe server starts on port 5000. It makes a connection to elasticsearch engine on 10.16.3.12. Note that this is a private IP, so you must first setup elastic search and change the IP accordingly.
A java program that retrieves headlines from BBC RSS feed and produces a log file containing the following information for each article:
- Title
- Content
- Date
- Author
- List of topics that the article belongs to