This project conducts sentiment analysis on 100 articles from The City A.M. The articles were taken from the Industrials sections of the online site. Articles were chosen where they were written specifically about a company on the London Stock Exchange. This meant share prices could be looked up and taken from the London Stock Exchange website.
Sentiment analysis was broken down into 4 components. These components were sentiment and subjectivity values; NaiveBayes classifier for overall article polarity classification; Textblob object constructor for overall article polarity classification; and Textblob object constructor for sentence level classification.
After each article was analysed, each sentence and their polarity classification were added to the corpus and training data. This allowed for each subsequent to be analysed on the training data in the corpus.