Recommendation Engines algorithms are typically based on collaborative and content based filtering methods and / or combination of both. In our project we are working on Next Best Action and Offer with Dynamic Recommendation Systems. Details can be found at:
Article Emotion Based Music Recommendation System Using Wearable Phy...
Thesis Emotion Aware Artificial Intelligence for Cognitive Systems
Project https://www.researchgate.net/project/Next-Best-Action-and-Offer-Prediction-with-Dynamic-Recommendation-Systems
However we would like to hear suggestions and alternative recommendations from the researchers who are working on recommendation engines (aka recommender systems). What are the software platforms, software packages, approaches, algorithms and tools do you recommend for Recommendation Engines? How do you relate your solution to big data analysis, machine learning or deep learning frameworks?
A dissertation of a new recommender system targets to suggest appropriate journals for submitting fresh manuscripts is located at:
https://edoc.hu-berlin.de/handle/18452/20889
The study is based on content-based filtering and knowledge-based filtering. As algorithms, BM25 has been utilized for content-based filtering, while another similarity/distance measure (Gower's measure) is utilized for knowledge-based filtering. This system uses Lucene search engine library as the major software tool to implement the system. As far as I know, there is no any approach of incorporating deep-learning techniques in the field of journal recommendation.
I would use: Tcl/Tk (fast prototyping including GUI, gluing various software, the best threading implementation) as basis, SQLite (super embeddable, simple and mature) for the big data management, and MXNet for ML.
In my experience best approach would be to use ... tensorflow for training recommendation model by deep learning based method. Using tensorlfow.js to convert the model to be used in browser... and using java-script interface to avail recommendation to the user.... with this method you con port your engine to mobile using cordova and to pc using Node webkit .. just need to program once...
easy peasy... no server no data warehousing....
- Badges
- Science topic