klaudiosinani/moviebox

Machine learning movie recommending system

What it solves

It provides a way to find movies with similar stories and meanings to a specific movie based on its plot summary, rather than relying on generic categories.

How it works

The system uses a dataset of 5,000 movie titles and plots from the Carnegie Mellon Movie Summary Corpus. It employs TF-IDF (Term Frequency-Inverse Document Frequency) to analyze and vectorize the plot summaries. To find recommendations, it calculates the cosine similarity between the vector of the input movie and the vectors of other movies in the dataset, selecting those with the closest match.

Who it’s for

Developers or researchers interested in content-based recommendation systems and the application of TF-IDF and cosine similarity for text analysis.

Highlights

  • Content-based recommendation engine using TF-IDF and cosine similarity.
  • Includes a dataset of 5,000 movie titles, categories, and plots.
  • Accessible via both a command-line interface (CLI) and a Python API.
  • Supports customizable recommendation counts and optional plot display.

Related

  • Project
    Bergvca/string_grouper

    A fast Python library for finding and grouping similar strings using tf-idf and cosine similarity, designed for efficient fuzzy matching and deduplication of large datasets.

  • Project
    sdl60660/letterboxd_recommendations

    A movie recommendation system that uses collaborative filtering and SVD to suggest films to Letterboxd users based on their rating patterns.

  • Project
    BUPTLdy/Sentiment-Analysis

    A Chinese product review sentiment classifier that provides a modern TF-IDF and Linear SVM baseline alongside legacy Word2Vec and LSTM implementations.

  • Project
    jnMetaCode/ai-shortfilm-prompts

    A cinematic video prompting system and Claude Code plugin that uses a 5-stage cinematography-based framework to create professional-grade AI short films.

  • Dispatch
    Image Similarity with Hugging Face Datasets and Transformers

    Hugging Face demonstrates how to build an image similarity system using the Transformers and Datasets libraries by computing dense vector embeddings and measuring cosine similarity.