PROSIM

Rechercher dans ce blog

How are statistics computed?

le octobre 20, 2018

Obtenir le lien
Facebook
X
Pinterest
E-mail
Autres applications

draft stats

Obtenir le lien
Facebook
X
Pinterest
E-mail
Autres applications

Commentaires

Enregistrer un commentaire

Translate

Articles les plus consultés

Architecture Map

French-spoken version with French and English subtitles. This video presents the architecture of the PROSIM database (PROduct SIMilarity) which compares products from the collaborative, free and open database of food products from around the world ( http://world.openfoodfacts.org ). PROSIM is built with Mongo running on a server, and can be queried using its dedicated API . The Tuttifrutti website ( https://tuttifrutti.alwaysdata.net ) uses this API for showing data in a graph with the following available scores: Nutriscore Nova Classification Scores are directly computed and delivered by OpenFoodFacts. The computation for the similarity between 2 products and their score is provided by a file called ComputingInstance.java (1 file per score) whose code is accessible here: Nutriscore Nova Classification

Engine's Statistics <nutriscore> on July 2020

OFF JSON data timestamp: 25/06/2020 # products: 1 398 820 # valid products: 806 780 Running on: Laptop Core i5 2430M (2.4/3 GHz) 8GB RAM, Kubuntu 18.04, 4 cores 64 bits Arch. sub-matrix WIDTH : 7 500 Arch. sub-matrix HEIGHT : 7 500 requester: oric.dev email: oric.dev@iznogoud.neomailbox.ch computing instance db nickname: nutriscore db-name: tuttifrutti_nutriscore first time creation: 16/06/2018 summary: worldwide products based on their nutriscore description (EN): Database aimed at classifying OpenFoodFacts products based on their similarity to each other along with their nutriscore description (other): Base de classification des produits OpenFoodFacts selon leur proximité avec des produits équivalents et leur score nutritionnel (nutriscore) %age of completion: 40.40 # products in db: 423 283 stats panel: not available # intersections computed: {N.(N-1)} 262 961 163 234 (on 650 893 968 400) Min. similarity %age: 75 # intersec...

Engine's Statistics <novascore> on 12.10.2019

OFF JSON data timestamp: 22.08.2019 # products: 757 828 # candidate products: 297 505 Running on: PC 3.50GHz 16GB RAM 8 cores 64 bits Status: completed Arch. sub-matrix WIDTH : 35 000 Arch. sub-matrix HEIGHT : 35 000 requester: oric.dev email: oric.dev@iznogoud.neomailbox.ch computing instance db nickname: novascore db-name: tuttifrutti_novascore first time creation: 08/10/2018 summary: worldwide products based on their Nova classification description (EN): Database aimed at classifying OpenFoodFacts products based on their similarity to each other along with their Nova classification description (other): Base de classification des produits OpenFoodFacts selon leur proximité avec des produits équivalents et leur classification Nova (novascore) %age of completion: 70.404 # products in db: 155 883 # intersections computed: {N.(N-1)} 62 314 214 770 (on 88 508 927 520) Min. similarity %age: 75 # intersections in db: 318 43...

Fourni par Blogger