As final project for CS 410 Text Information System, we participated in Text Classification Competition to detect Twitter Sarcasm. We were given both Training and Test datasets.

label: SARCASM or NOT_SARCASM response: The Tweet to be classified context: The conversation context of the response example: {"label": "SARCASM", "response": "@USER @USER @USER I don't get this .. obviously you do care or you would've moved right along .. instead you decided to care and troll her ..", "context": ["A minor child deserves privacy and should be kept out of politics . Pamela Karlan , you should be ashamed of your very angry and obviously biased public pandering , and using a child to do it .", "@USER If your child isn't named Barron ... #BeBest Melania couldn't care less . Fact . 💯"]}

id: String identifier for sample. This id is required for project submission and grading. response: The Tweet to be classified context: The conversation context of the response example: {"id": "twitter_1", "response": "@USER @USER @USER My 3 year old , that just finished reading Nietzsche and then asked me : " ayo papa why these people always trying to cancel someone on Twitter , trying to pretend like that makes them better themselves ? " . To which I replied " idk " , and he just " cuz hoes mad " . Im so proud . ", "context": ["Well now that \u2019 s problematic AF ", "@USER @USER My 5 year old ... asked me why they are making fun of Native Americans ..", "@USER @USER @USER I will take shit that didn't happen for $ 100", "@USER @USER @USER No .. he actually in the gifted program and reads on second grade level . ... and he knows Kansas City is in Missouri"]}

Train Test 5000 1800

Our project objective is to learn from the Training dataset and predict the labels of Test dataset (SARCASM or NOT_SARCASM).

1. Python==3.8.3
2. nltk==3.5
3. pandas==1.0.5
4. scikit_learn==0.23.2

1. conda create -n "project_demo" python=3.8.3
2. conda activate project_demo
3. git clone https://github.com/subhasishb-coder/CourseProject.git
4. cd CourseProject
5. pip install nltk==3.5
6. pip install pandas==1.0.5
7. pip install scikit_learn==0.23.2
8. cd code
9. python TestClassficationCompetion_Sarcasm_Detection.py

https://mediaspace.illinois.edu/media/t/1_xwb0wmzt