Stack Overflow
Stack Overflow

Questions tagged [lightfm]

30 questions
3
votes
1 answer

LightFM train_interactions shared among train and test sets: This will cause incorrect evaluation, check your data split

tl;dr: Working with Yelp Dataset to make a recommendation System but running into Test interactions matrix and train interactions matrix share 68 interactions. This will cause incorrect evaluation, check your data split. error when running the…
Shyu
  • 71
  • 6
2
votes
1 answer

Loss functions in LightFM

I recently came across LightFM while learning to train a recommender system. And so far what I know is that it utilizes loss functions which are logistic, BPR, WARP and k-OS WARP. I did not go through the math behind all these functions. Now what I…
Hasham Mirza
  • 23
  • 1
  • 5
1
vote
1 answer

LightFM how to make predictions for new users (cold start) - user id 8 not in user id mappings

I am building a recommendation system in order to recommend training to employees based on user features and item features which LightFM according to the documentation its a great algorithm. my user dataframe: User-Id name age los ou gender …
Luis Valencia
  • 32,619
  • 93
  • 286
  • 506
1
vote
0 answers

How to build the dataset.build_item_features for LightFM recommendation model?

I am working on modeling a recommendation system using LightFM and trying to build the dataset on Steam data. My initial part of the code is as below: dataset = Dataset() dataset.fit(users = users_data['userID'], items = games_data['id']) …
user11619814
  • 389
  • 1
  • 6
  • 17
1
vote
0 answers

How to build interaction for LightFM fit?

I want to train LightFM with Dataframe shown below. FromUserID Title PostID Views Likes 1 Cricket 34 12 8 6 Mobile 37 11 6 3 tv 34 8 5 5 WTC final 30 6 5 I already trained the model with Views column (which is working fine) but…
Mahipal Singh
  • 144
  • 1
  • 11
1
vote
0 answers

Obtaining feature importance/Sensitivity for model interpretability

I'm new to recommender models and I'm using LightFM for a project. I'm creating model for customer like/dislike recommendations (no ratings involved). Are there any options for model interpretability in such cases? I understand LightFM is a hybrid…
user9291966
  • 11
  • 3
1
vote
1 answer

Lightfm Incorrect number of features in user_features

I am building recommender system - hybrid in Lightfm. My data has 39326 unique users and 2569 unique game titles(items). My train interaction sparce matrix has shape: <39326x2569 sparse matrix of type '' with 758931 stored…
Ellie Biessek
  • 148
  • 2
  • 14
1
vote
1 answer

How can I do cross validation on user-item interactions matrix for LightFM movie recommender system?

I have an interactions matrix (scipy.sparse.csr_matrix) from movielens dataset with movie ratings from users, and I am building a LightFM model with item_features. Right now I have the matrix divided into train and test, but how do I do…
xxx
  • 41
  • 5
0
votes
0 answers

My kernel dies when I fit my LightFm model from Microsoft Recommenders

I downloaded the code to run it in my Jupyter Notebook https://github.com/microsoft/recommenders/blob/main/examples/02_model_hybrid/lightfm_deep_dive.ipynb When I try to run the LightFm model the kernel dies (On the exact same code from the…
KwstisXD
  • 1
0
votes
0 answers

Why am I not getting anything printed out in the terminal while using lightfm in python?

So I am trying to build a recommender system and found out that the library lightfm offers the functionalities to build it. I went to their site and looked into the documentation and I saw some examples that I copied to test and to see what they do.…
All For One
  • 1
  • 1
0
votes
0 answers

I would like to know how I can obtain, if possible, the MSE and MAE metrics of the lightFM model in python

To evaluate the model, I need to obtain the lightFM model error metrics, specifically the MAE and MSE metrics. I have found extensive literature on obtaining error metrics for models implemented with scikit-learn and tensorflow libraries, but in the…
CAMILO RUBIO
  • 3
  • 4
0
votes
0 answers

GridSearchCV with LightFM

I'm a newbie so my apologies if something I ask might be to obvious and my english is not quite good. I'm stuck in doing a custom grid search with cross validation with LightFM which does not come with those functions. It seem the way I split the…
Gianluca Calò
  • 11
  • 1
0
votes
0 answers

How do i convert a pd df to a sparse matrix for a lightFM recommender without first making a memory-busting non-sparse matrix?

I am building a LightFM recommender in python. I have a pandas dataframe of transactions: each row is a purchase and has a customer and product id. I realise that to build my model I need to convert this to a sparse matrix with 1 row per customer…
Evolving_Richie
  • 23
  • 5
0
votes
0 answers

LightFM- alternative to save model

I created a recommendation system using lightFM on Google Colab. I want to deploy this model using Node JS. I read this article and tried the method recommended here. It recommended using tensorflow.js. For that the model needed to be saved and this…
Sara
  • 1
0
votes
0 answers

Lightfm: how to predict on new user who has interactions with items?

I'm using Lightfm for movie recommendation, my question is quite simple but I can't find the solution to it: I have trained a model using : User Feature matrix (Age, gender etc..) Item Feature Matrix (movie category etc ..) User Item interactions:…
Bonbon
  • 1
1
2