Recommendation system
We have developed a recommendation system based on our deep metadata and the largest multimedia content database. The recommendation system can also work with big data, which is useful for a "cold" subscriber.
Peculiar properties:
Ratings:
- Best/Worst/Popular
- Everyone beloved/Ambiguous
- Male / female
- For any audience segments
- Metadata ratings
- Ratings of the relevance
Info about the user:
- Users with similar tastes
- Consumption statistics
- Favorite genres/year/countries
- Distribution of ratings
- Portrait of the user by taste factors
Content info:
- Statistics on the object
- Distribution of ratings
- Gender preferences
- Age audience
- Similarity to other objects
- Assessment of the availability of content
Predictive estimate:
- Accurate prediction, forecast 8.68
- Binary tag, like/dislike
- Subjective assessments
- Explanation of forecast