Zindi
CompeteJobsLearnChatLeaderboardFor Organisations
PartnersMeet the teamPressCase studiesAI4EAC
Primary competition visual

Akeed Restaurant Recommendation Challenge

Helping Oman
$3 000 USD
Completed (~6 years ago)
Prediction
Collaborative Filtering
1430 joined
242 active
InfoDataChatLeaderboard
Starti
18 May 20
Closei
16 Aug 20
Reveali
16 Aug 20
User avatar
Alsubari
IBB University
should we predict one restaurant for each customer location, or more?
Help · 17 Jul 2020, 13:15 · 4

I am getting confused, because when I increase prediction number for each customer location, the accuracy increase.

Discussion 4 answers
User avatar
johnpateha
17 Jul 2020, 14:07

in orders we could find more than one restaurant per customer location. I think that test data very similar to train, so it's a good idea to predict sometimes more than one (based on probabilities from your model).

Upvotes 0
User avatar
Alsubari
17 Jul 2020, 14:10replied to johnpateha

Thankx alot

Upvotes 0
User avatar
johnpateha
17 Jul 2020, 14:15

I suggest you to find prob threshold which maximized F1 for your validation data and mark as 1 all rest above this threshold for test

Upvotes 0
User avatar
Alsubari
17 Jul 2020, 14:27replied to johnpateha

that will be better, thanks alot sir, I will try it

Upvotes 0

Join the largest network for
data scientists and AI builders

AboutFAQsPrivacy PolicyTerms of UseRules
Status
© Zindi 2026