/hermes

Recommender System Framework

Primary LanguageJupyter NotebookOtherNOASSERTION

hermes

Hermes is Lab41's foray into recommender systems. It explores how to choose a recommender system for a new application by analyzing the performance of multiple recommender system algorithms on a variety of datasets.

It also explores how recommender systems may assist a software developer of data scientist find new data, tools, and computer programs.

The Wiki associated with this project has details on many references that we utilized when implementing this framework. It also details the datasets used in this base framework, as well as some resources to help you get started in recommender systems and Spark.

For tips on how to get started, see the wiki page: Running Hermes.

##Blog Overviews There are a number of blog articles that we produced during the course of this project. They include:

Join the Hermes Running Club March 2016
Python2Vec: Word Embeddings for Source Code March 2016
TPS Report for Recommender Systems? Traditional Performance Metrics March 2016
Recommender Systems - It's Not All About the Accuracy January 2016
The Nine Must-Have Datasets for Investigating Recommender Systems February 2016
Recommending Recommendation Systems (project intro) December 2015

visualization

We are trying varied tools and concepts to visualize the results of this project.

boku

  • conda install bokeh
  • from top-level hermes folder $bokeh serve src/results/hermes_run_view.py
  • view in browser at http://localhost:5006/hermes_run_view

d3

  • easy_install web.py
  • from viz folder $python app.py
  • view in browser from location:port displayed in terminal