The IMDb dataset is a collection of 50,000 reviews from the Internet Movie Database (IMDb). The reviews are labeled as either positive or negative and are split into two sets of 25,000 reviews for training and testing. Each set contains an equal number of positive and negative reviews.

The IMDb dataset is a binary sentiment analysis dataset for natural language processing or text analytics. It contains more data than previous benchmark datasets.

IMDb is a rich source of film data that includes cast and crew lists, movie release dates, box office information, plot summaries, trailers, actor and director biographies, and other trivia. Information on IMDb comes from a variety of sources, such as filmmakers, film studios, on-screen credits, and other official sources.