Develop an automatic sentiment analysis system and evaluate it using the IMDB Movie Dataset (http://ai.stanford.edu/~amaas/data/sentiment/) (http://ai.stanford.edu/~amaas/data/sentiment/) or your own corpus built for this course. Your documentation should include

• a literature survey on the task and similar datasets,
• details of your implementation, including the evaluation method applied,
• an error analysis. (Analyze cases where your system generated incorrect results, classify those cases, and identify any observed patterns in the erroneous outcomes.)