Sentiment analysis of Leonardo's film reviews
web crawler: SeleniumScrapper.py
data preprocessing and model training: final.ipynb
This project aims to apply the machine learning models to analyse the sentiment of Leonardo’s films.
This project includes several steps:
(1) Data extraction: use WebDriver to crawl movie reviews of Leonardo from IMDB (by selenium library)
(2) Manual label: Select representative films to label Pos or Neg
(3) Text preprocessing: replace contractions, remove non-ascii, remove punctuation, etc.
(4) Model training: Use classification algorithm including random forest, SVM, deep learning, etc.
(5) Compare result of different models
dataset includes:
| movie | title | rate | author | date | content | label |
|---|---|---|---|---|---|---|
| Blood Diamond | One word: Superb | 9/10 | D_Mas | 2007/10/8 | A wonderful movie, wish I would have seen this movie earlier! Blood Diamond brought together a variety of different aspects. | pos |
result: predicted label (pos or neg)