gorse changelog

Changelog History

• v0.1.3 Changes

  December 21, 2019

  ⚡️ This is a minor update of gorse 0.1.2

  🔋 Features

  • Storage :
    • Store items with popularity.
    • Store items with dates.
  • Service :
    • List the latest items.
    • Recommend items with a combination of popularity and dates.

  ⬆️ Upgrade Guide

  • Reimport data

    Export data from old versiongorse export-items database.db items.csv gorse export-feedback database.db feedback.csv# Import data to new versiongorse import-items database.db items.csv gorse import-feedback database.db feedback.csv

• v0.1.2 Changes

  December 14, 2019

  ⚡️ This is a minor update of gorse 0.1.1

  🔋 Features

  • 👉 Use string type for user ID and item ID: #23

  ⬆️ Upgrade Guide

  • Reimport data

    Export data from old versiongorse export-items database.db items.csv gorse export-feedback database.db feedback.csv# Import data to new versiongorse import-items database.db items.csv gorse import-feedback database.db feedback.csv

• v0.1.1 Changes

  November 24, 2019

  This is a fix version of gorse 0.1

  🛠 Fix

  • 🛠 Fix typos in the gorse test command.
  • 🔄 Change KNN type names and similarity names in configurations to lowercase.
  • 🔄 Change default similarity function to implicit for neighbors recommendation.

  ⬆️ Upgrade Guide

  • 🔄 Change configurations if following names are used:
    • KNN type names: Centered ➡️centered, ZScore ➡️z_score, Baseline ➡️baseline
    • similarity names: Pearson ➡️pearson, Cosine ➡️cosine, MSD ➡️msd

• v0.1 Changes

  November 13, 2019

  🚀 This is the first released version.

  🔋 Features

  • Recommendation Models :
    • KNN (Explicit/Implicit)
    • CoClustering
    • SlopeOne
    • SVD
    • SVD++
    • NMF
    • ItemPop
    • BPR
    • WRMF
  • Dateset Utils :
    • Load data from buildt-in datasets or CSV files
    • Split data by k-fold or ratio
  • Model Evaluation Utils :
    • Evaluate models by rating metrics: MAE/RMSE
    • Evaluate models by ranking metrics: Precision/Recall/NDCG/MAP/MRR
  • RESTful Recommender Service :
    • Provide RESTful API
    • Store and manage feedvack based on BoltDB
    • Fit model and generate recommendations for each user
  • CLI tools :
    • Data import/export tools for recommender server
    • Model evaluator