Features and integration
recommendo, commendo's hybrid recommendation engine solution (including collaborative filtering and content based filtering approaches) can be integrated in online shops and distribution portals very easily.
Interfaces:
For data input/output there are 2 different interfaces available (one has to be chosen):
Requests:
Different types of recommendation requests are possible via the interfaces:
- User / Item Prediction
- Item / Item Correlation (similarity of e.g. products)
- User / User Correlation (similarity of e.g. users)
Interface calls - in general:
For data import as well as recommendation score export several interface calls are available.
Input calls (add data to the recommendo engine):
- New user
- New items
- Shopping cart content
- Click information
Output calls (get predictions and correlations out of the engine):
- Sorted top-K recommendations
- Unsorted recommendation list
- Predictions for the overall set of items (sorted & unsorted) as well as for categories
Available Data sources
Integration process:
recommendo can be integrated very fast in a few days.
Progress:
- Initial discussion, including the definition of used data sources
- Delivery of a historic data snap shot (if available) to commendo for analysis purposes
- Data analysis and set up of recommendo basic configuration, based on data sources and field of application
- Delivery of recommendo including adapted basic configuration (stand-alone software, fully functional) to customer
- Trial period
On customer side:
- Data preparation: Providing the data snap shot for initial analysis and training
- Integration: Interface connection on customer side
- Visualization of recommendations for each user (front end)
For further questions or technical details please contact us via the
contact form.
