Video Clustering Algorithm

The algorithm is part of a series of repositories for a Machine Learning-based video recomendation system for a web applications users. The three fundamental purposes of this algorithm are:

1. To create a set of video clusters based on the AWS Rekognition tags for the videos using an unsupervised machine learning algorithm.
2. To create a cluster classifier for new (out of training) videos using an ensemble learning model.
3. To create a bit-wise dataframe to perform quick and efficient vertorized operations.

Features

• Import a csv file with the video tags
• Export the model and results to a S3 bucket
• Write the video clusters in a Postgress SQL database

Installation

For testing purposes we porpose using python 3.8.* with a 64-bits machine.

Clone the repository and Install the dependencies

git clone/git pull
pip3 install -r requirements.txt

Development

Want to contribute? Great!

License

MIT

Free Software, Hell Yeah!