Weekly Update

The theme of the week was research. After reading a couple of white papers, including Dr. Ways!, we, as a team are starting to get a better understanding as to the challenges of performing sentiment analysis effectively, and also the countless ways to do so.

Some of the papers we all read were: * Dr. Way's Research Paper * Sentiment Analysis * Opinion Mining

From reading these papers, we got more familiar with a couple of NLP (natural language processing) techniques, along with some more specific uses for Python & NLTK in sentiment analysis.

Word of the Week

As with anything, understanding the terminology is very important. So, Here is our sentiement analysis word of the week! An important piece of sentiment analysis terminology is a "feature" “Features” are whatever you’re analyzing in an attempt to correlate to the labels. For example, in some cases of sentiment analysis, the features will be the words in each tweet. Other algorithms could use different types of features — some algorithms use bigrams or trigrams (strings of two or three consecutive words, respectively) as the features. It all depends on what algorithm we decide to use.

What to Do Next

We must select the features but then we must chose a classifier. Here is where the next portion of our research is going to come in. Do we use and train a Naive Bayes classifier, or do we try a SVM? There are a couple of different routes to go and we need to look more into the capabilities of NLTK and different API's to see what would be the most efficient for what we are trying to do.

We also learned of a couple of different problems predominating in the research community: different classifing methods and handling negations and sarcasm. Again, areas for more research and exploration.

Finally, we have decided that as another visualization. We are going to try and implement a heat map. Showing not only sentiement towards a particular term, but also geographic location.

Much to be done! - AttitudeAnalytiks