IMDB Movie Review Sentiment Classifier
This model classifies English movie reviews into Positive or Negative sentiment using TF-IDF and Logistic Regression trained on the 50,000 IMDB dataset.
- Accuracy: ~90.95%
- Algorithm: Logistic Regression with
TfidfVectorizer(unigrams + bigrams)
How to Use
Download and run predictions with Python using huggingface_hub:
import os
import joblib
from huggingface_hub import hf_hub_download
# Download model and vectorizer from this repository
REPO_ID = "YOUR-USERNAME/imdb-sentiment-analysis"
model_path = hf_hub_download(repo_id=REPO_ID, filename="sentiment_model.pkl")
vectorizer_path = hf_hub_download(repo_id=REPO_ID, filename="tfidf_vectorizer.pkl")
# Load model and vectorizer
model = joblib.load(model_path)
vectorizer = joblib.load(vectorizer_path)
# Test review
reviews = ["This movie was magnificent!", "Terrible waste of time."]
reviews_tfidf = vectorizer.transform(reviews)
predictions = model.predict(reviews_tfidf)
for review, pred in zip(reviews, predictions):
sentiment = "Positive" if pred == 1 else "Negative"
print(f"{review} -> {sentiment}")