NTU » Sentiment Analysis

This project is about twitter. User's post in twitter is known as "tweet". We want to create an algorithm that detects user's emotion contained in tweet. For example, one tweet may reveal that the user is happy, another may suggeset that he/she is sad.

To make the algorithm detect the user's emotion automatically, training data is needed. That is to say, the algorithm needs some examples to learn from. You may help us to collect the examples.

Now, what we have is some tweets collected from twitter. What you need to do is judge whether a tweet contains positive emotion (e.g., the writer was happy), or negative emotion (e.g., the writer was sad), or the tweet is neutral (i.e., the writer didn't express any emotion in this tweet).

There are 385 pieces of tweet in the database. Please label them into three categories: positive emotion, negative emotion, or neutral.

Please note that these 385 tweets are regarded as ambiguous by an automatic labeling program who labeled the tweet by checking whether it contains certain pre-defined positive / negative emotion bearing words. These 385 tweets contain both positive and negative emotion bearing words. So it's hard for the program to make decision. Please help to label the 385 twees according to your judgement.

We can move on to other dataset after you finish labeling these 385 tweets. Let's see how long does it take first.