Here is my first Data Science project about Sarcasm Detection based on Kaggle dataset.

We have a dataset of sarcastic headlines from two news websites. There is each dataset, which consists of three attributes:

- is_sarcastic: 1 if the dataset is sarcastic, 0 otherwise
- headline: the headline of the news article
- article_link: Link to the original news article.

We need to find out which news items are sarcastic and which are not. I solved this problem in two tasks. In 1 task I worked with the following plan:

• Read and clean the data
• Find and visualise most common words
• Identify and visualise stop words
• Pre-process text
• Define classification models
• Retrieve and save the best model

In 2 task I worked with plan bellow:

• Load and pre-process data
• Define the Deep Learning models
• Get the best model to solve the problem
• Save model.