This is course of Northeastern University Lecture assignment.
- Assignment 1: Sentiment Analysis on Movie Reviews
- Assignment 2: Text Classification Project with TF-IDF Vectorization
- Assignment 3(On working): Implementing Skipgram and CBOW Algorithms
๐ข Introduction: In this project, I've tackled the exciting challenge of sentiment analysis, specifically focusing on movie reviews. Movies have always been a passion of mine, but diving deep into what people think about them? That's an adventure in itself!
๐ฅ About the Dataset: The dataset I've worked with comprises both positive and negative movie reviews. The contrast between these reviews, the emotions, the passion, and sometimes the sarcasm, provided a rich ground for analysis.
๐ Sentiment Analysis: Sentiment analysis is more than just finding positive or negative words in a text. It's about understanding context, tone, and underlying feelings. My goal was to assess if a given review praised the movie or was critical of it.
๐ง Tools & Methodology: For this analysis, I leveraged the Naive Bayes algorithm. Naive Bayes, with its probabilistic approach, proved to be quite effective in discerning the sentiments embedded within the reviews.
๐ฌ Outcome: Tapping into the world of movies and extracting sentiments from reviews was both challenging and rewarding. Through this project, I got a glimpse into the world of critics and movie enthusiasts alike, understanding what makes a movie resonate or fall flat for them.
๐ Gratitude: A big thank you to everyone who has contributed to the dataset and to those who are checking out this project. It was a roller coaster of emotions (pun intended), and I'm thrilled to share this journey with all of you!
๐ข Introduction: In this project, I've dived deep into the realms of text classification leveraging the power of Multinomial Naive Bayes and Neural Networks. Unlike traditional Bag of Words methods, I chose to employ TF-IDF Vectorization, ensuring a more nuanced representation of text data.
๐ Visualization: Besides just classification, I've also dedicated efforts to visualize these vectors. It's truly intriguing to see how different text data points relate to one another in a multidimensional space.
๐ Datasets: I've worked with some large datasets in this endeavor. These aren't your typical few-hundred rows of data - they are massive! And that's what made this project even more interesting (and a tad bit challenging).
โฐ Execution Time: Given the size of the datasets and the computations involved, certain parts of the project can take a considerable amount of time to run. Hence, patience is key. If you're planning on running the code or testing it out, maybe go grab a coffee in the meantime.
๐ Recommendation: I recommend starting with the exploration and preprocessing of these datasets a tad bit early if you're planning to replicate or extend this work. This isn't just due to the execution time but also because understanding the nuances of such vast data takes time.
๐ Tools & Libraries Used:
TF-IDF Vectorization Multinomial Naive Bayes Neural Networks And a few more... ๐ Acknowledgements: I want to give a shout-out to all the open-source tools and libraries that made this project possible. And of course, to all of you who are checking out this project or finding it useful in any capacity!