Titan Recommendation Algorithm

The Titan Recommendation Algorithm is a powerful and flexible video recommendation system designed to generate personalized video feeds based on user preferences, behaviors, and interactions. This package leverages a modular design, worker threads, and a robust scoring system to deliver highly relevant content to users.

Features

User-Centric Recommendations: Personalize feeds using user preferences, location, feedback, and interaction history. Dynamic Scoring: Calculate video relevance based on engagement, novelty, social proof, and more. Modular Configuration: Easily customize weights, thresholds, and scoring logic through the Config interface. Worker Threads: Efficiently process large datasets with asynchronous workers. Diversity Support: Promote diverse recommendations by penalizing overused tags or creators.

Installation

To install the package, use npm:

npm install titan-recommendation-algorithm

Usage

Generate a Personalized Feed

import { generateFeed } from 'titan-recommendation-algorithm'; 
import { UserPreferences, UserData, Video } from './types';

const userPreferences: UserPreferences = { 
  preferredTags: ['comedy', 'education'], 
  minDuration: 10, 
  maxDuration: 300, 
};

const userData: UserData = { 
  userLocation: 'New York', 
  rewatchedVideos: ['video1'], 
};

const videos: Video[] = [ 
  { 
    id: 'video1', 
    tags: ['comedy'], 
    duration: 120, 
    views: 1000, 
    engagementRate: 0.8, 
    uploadDate: '2023-12-01' 
  }, { 
    id: 'video2', 
    tags: ['sports'], 
    duration: 90, 
    views: 2000, 
    engagementRate: 0.6, 
    uploadDate: '2023-12-02' 
  }, 
];

generateFeed(
    userPreferences, 
    userData, 
    videos
  ).then((feed) => { 
      console.log(feed)
  }); // Outputs an array of sorted video IDs });

Update User Feedback

import { updateFeedback } from 'titan-recommendation-algorithm'; import { UserData } from './types';

const userData: UserData = { 
  rewatchedVideos: ['video1'], 
  feedback: { video1: 4 }, 
};

const updatedData = updateFeedback(userData, 'video2', 5); console.log(updatedData.feedback); // Outputs: { video1: 4, video2: 5 }

Adjust Configuration

import { updateConfig, getConfig } from 'titan-recommendation-algorithm';

updateConfig({ noveltyScoreWeight: 0.5 }); console.log(getConfig().noveltyScoreWeight); // Outputs: 0.5

Documentation

For detailed documentation on each function, interface, and configuration option, refer to the TSDoc comments in the source files.

Contributing

We welcome contributions! See the CONTRIBUTING.md file for more details.

License

This project is licensed under the MIT License. See the LICENSE file for details.