Questions tagged [recommender-system]
A recommendation engine tries to predict how much a user will enjoy certain goods (movies, books, songs, etc) and makes recommendations. They are often used by online vendors to suggest new purchases.
441
questions
0
votes
0answers
8 views
How to generate syntactic data using real data? [closed]
I have a dataset that includes 300 samples and 6 features (categorical and numerical), and also a target variable.
This dataset comes from a context where shows how much each individual person (e.g., ...
0
votes
0answers
7 views
Collaborative filtering: can I use test loss for number of features selection?
We have a big dataset of users with their item preferences and we want to give recommendations based on Collaborative filtering.
Or issue is that when selecting the number of features using precision@...
0
votes
0answers
16 views
Product cannibalization - Would affinity analysis help?
We send an email to our customers to advertise 3-4 products every day.
Supposing that today's email is advertising three products (product A, product B, product C), the customer is given four options:
...
0
votes
1answer
27 views
Unbiasing machine learning model features, is it valid?
I've been working on a RecSys model recently (using HRNNs), and when thinking about the features used for users and itens, I thought that many of them ended up being biased by the old system ...
1
vote
0answers
5 views
How to get KNN linearly hybridised by two similarities?
I'm writing a KNN (collaborative filtering) hybrid similarity recommender and I need some advice. It is based on this paper.
I've currently got 2 datasets. The first one is ...
1
vote
0answers
96 views
Adding explicit user info to matrix factorization
In the paper Matrix Factorization Techniques for Recommender Systems, it is claimed that we can incorporate extra user information into our recommender model by doing something like this:
$$
\hat{r}_{...
0
votes
0answers
10 views
How should I create rating matrix for implicit data?
My task is very simillar to kaggle: Instacart Market Basket Analysis. I need to predict which products will be in a user’s next order. So i have data wich looks like this
Now am working on building ...
0
votes
0answers
16 views
How can I use decision tree for a ranking model in a recommendation system?
I am trying to understand how ranking models work. I did some research and found that ranking models are models that can rank things to recommend for recommendations. I have found that I can use ...
0
votes
0answers
20 views
What is the right approach to bucket users for algorithms with different coverage for A/B testing
I've couple of recommendation algorithms that I want to A/B test. Algorithm A has 90% user coverage and algorithm B has 95% user coverage. That means if the algorithms are asked to provide ...
0
votes
0answers
14 views
How to calculate a similarity score between two users rating films
This is probably a standard problem for a seasoned data scientist or statistician, but I'm stuck right at the beginning:
Suppose I have two users (A and B) on a site like IMDB, where they can rate a ...
0
votes
0answers
7 views
How to use Product Matching to create Product Bundles
I am working on a product matching model.
GOAL
A store has many products like creams, perfumes, other beauty products.
Based on product properties I have to cerate bundles of it so we can sell more ...
0
votes
0answers
19 views
Addressing self-fulfilling-prophecy / feedback issues for ML models that retrain on data that they themselves previously generated labels for?
Methods for addressing self-fulfilling-prophecy / feedback / echo-chamber issues for ML models that train on data that they themselves previously generated recommendations / labels for?
Suppose we ...
0
votes
0answers
11 views
Cross validation for Collaborative filter-based recommendation systems
I am an absolute beginner and am trying to implement collaborative filter for furniture ecommerce (think wayfair). I need some guidance about cross-validation strategy.
Situation:
I am working on a ...
0
votes
0answers
11 views
What is the difference between Singular Value Decomposition and Matrix Factorization for a Recommendation System?
For collaborative Filtering method to recommend items to users, I have read articles on the Internet that we can do matrix factorization. However, I am confused about the term between Matrix ...
2
votes
0answers
14 views
Detecting behaviour change in a recommender system
Suppose I want to track a recommender system's live performance. The task isn't exactly to detect outliers, but to detect if the system started behaving differently, looking at the output only - an ...
0
votes
0answers
7 views
Ranking items by the magnitude of their effect on dissimilarity?
[reposting with more detail, after previous question was removed due to lack of detail or clarity]
I am working on getting a better understanding of my company's user base. We have distinct ...
0
votes
0answers
8 views
which recommendation system is used for social media
I'm a beginner in machine learning. Actually, I'm developing a social media app where the user will be shown different images that other users post on the app, the images are shown to the users based ...
0
votes
0answers
15 views
Normalizing relevancy score in recommendation system
I am designing a recommendation system that gets some data from the user, calculates a score per each of the products, and based on that recommends some products. Since there are lots of entry points, ...
0
votes
1answer
40 views
Top-N ranking loss is difficult to optimize directly. Why?
Variational Autoencoders for Collaborative Filtering paper tells:
"Recommender systems are often evaluated using ranking-based measures, such as mean average precision and normalized discounted ...
1
vote
1answer
16 views
If range of the activation function is smaller than that of data, the last layer of the decoder should be kept linear. Why?
Training Deep AutoEncoders for Collaborative Filtering paper tells that
"If range of the activation function is smaller than that of data,
the last layer of the decoder should be kept linear.&...
1
vote
1answer
23 views
A challenge training autoencoders is non-convexity of the objective. Why?
AutoRec: Autoencoders Meet Collaborative Filtering paper tells:
"A challenge training autoencoders is non-convexity of the objective."
Can someone explain and elaborate this statement?
2
votes
1answer
77 views
Should we use regression or classification for recommendation systems?
Lets say that based on some user available ratings i want to predict the ratings of unrated products for some specific user. I totally understand the fact that classifier models are for categorical ...
0
votes
0answers
14 views
Multiclass or Multilabel classifier for recommender engine
Context: Build recommender engine beeing able to recommend items to new users (in cold start scenarios)
X is dataframe with:
30 items
100k users
For each user features are: age, gender, income, ...
...
0
votes
0answers
10 views
Tag based recommendation system
Please bear with me cause I am not an expert in the subject. I want to build a recommendation system. It is not based on what others liked (collective based) or what I have liked (content based). The ...
0
votes
0answers
6 views
raw_id vs inner_id in Surprise Lib?
What is the difference between raw_id vs inner_id in Surprise Lib.
The documentation is not clear enough (at least for me). Why ...
1
vote
0answers
25 views
Embeddings model: out of sample prediction with Keras for Collaborative Filtering
I have been trying to play with an example on Collaborative Filtering for Movie Recommendations (keras.io), which builds embedding layers for movies and users.
Now, in a regular pre-trained word- and ...
0
votes
0answers
32 views
What is anti test set?
What is the exact meaning of the term Anti Test set? Why and where is it used?
I recently came across the link build_anti_testset(fill=None) and it is a bit ...
1
vote
1answer
29 views
Find optimal weights for a regression model with some restrictions
I have created a hybrid recommendation system which contains 4 recommendation models.
In my case i am trying to predict the ratings of the products and after that recommend the high rated (predictions)...
0
votes
0answers
14 views
Scoring metric for recommendation system
I'm working on a project that involves building a news recommendation system. I've come as far as quantifying user interaction with different articles on the site into user's affinity towards atopic ...
1
vote
0answers
39 views
Alternating least squares — what are the limitations?
I am taking an introductory course to Machine Learning and we learned alternating least squares for recommender systems. I learned that this method has some advantages --- easy to parallelize, and ...
2
votes
0answers
25 views
What is the difference between pairwise kernels and pairwise distances?
What is the difference between pairwise kernels and pairwise distances?
I frequently came across terms like pairwise kernels and pairwise distances while learning about Pairwise metrics, Affinities, ...
1
vote
0answers
34 views
optimal k for knn algorithm in item based recommendation system using cross-validation
I'm trying to make a recommendation system in my Graduation project app using k-nearest neighbor algorithm.
I make an item recommendation system to recommend products to the active user based on ...
0
votes
0answers
8 views
What are metrics for evaluating recommender systems if most users only transacts once?
What are metrics for evaluating recommender systems if most users only transacts once? 80% only bought once, 10% twice, and so on
If I use Precision@k=5, because most users are buying only once, then ...
1
vote
0answers
23 views
Recommender System - Predict ratings with Random Forest Regressor or Classifier?
My dataset has clothes that contain features for every cloth. Let me show you the dataframe to understand.
Every cloth is rated between [0,10] with step 0.5
In bibliography and paper conferences some ...
0
votes
1answer
61 views
mathematical proof that if the vectors 𝑞⃗ and 𝑑𝑖 all are normalized unit vectors [closed]
In a vector space model when we are given a query as a vector 𝑞⃗ and documents 𝑑1
, 𝑑2 …, we usually rank
the documents in relevance to the query vector using cosine similarity. Show by a ...
0
votes
0answers
137 views
Offline Precision@k and Recall@k for recommender system
How can I evaluate offline Precision@k and Recall@k metrics for recommender system if I only have items-users matrix? I think I can't just compare recommendations and user data because it will be ...
0
votes
0answers
7 views
Physical significance of non-negative factors of a matrix?
I was trying to make a recommender system using matrix factorization techniques on rating data.
I came across 2 algorithms - SVD and NMF. While the basic difference is very clear , I was wondering ...
5
votes
3answers
841 views
Can I make recommendations using inferences from multiple linear regression?
I have a question that doesn't seem to fall into classic "recommender systems", but may fall into the category of causal inference (though I'm not sure).
Say I have dataset of sales for a ...
0
votes
0answers
48 views
How can I use UV decomposition with Gradient Descent?
I have read Matrix and Tensor Factorization Techniques for Recommender Systems and I have seen a technique for Recommender System which is UV decomposition with objective funtion to minimize real ...
2
votes
0answers
20 views
Data Analysis, find the best city for given individuals
I'm trying to solve a data analysis question where I'm given data on ~1000 cities. For each city, I have: the average property tax, rating of schools nearby, crime rate, avg number of rooms per appt ...
0
votes
0answers
11 views
How to evaluate document similarity / content-based recommender system
I'm planning on building a basic content-based recommender system with word2vec and cosine similarity. The data consists of 300k documents in varying length.
How do I evaluate my model if I have no ...
2
votes
0answers
75 views
How do I calculate the metric Cumulative Gain,…?
I am currently in the process of creating a recommender system. Here I use a neural network and then I use k-means to find the closest neighbors and thus show the user 20 recommended articles.
I would ...
0
votes
0answers
39 views
DeepFM - how are numeric features embedded?
In the deepFM model architecture (https://arxiv.org/pdf/1804.04950.pdf), the following information is given about the output of the embedding layer (which feeds both into the NN and the FM parts of ...
0
votes
0answers
20 views
Recommending item category based on customer shopping behavior?
I'm trying to build some recommender system for online wine shop.
What I'm thinking is recommending item category (ex: red_pinot, ...
2
votes
1answer
45 views
How can we perform matrix factorisation for three dimensional matrix?
I am working on the recommendation system in which I have three factors user_id, time and ...
2
votes
2answers
155 views
Optimization for the autoRec algorithm
In its original paper, autoRec proposes the following algorithm:
$min_{\theta} \sum_{i = 1}^{n} \|r^{i} - h(r^{i}, \theta)\|^2_{\mathcal{O}} + \frac{\lambda}{2} (\|W\|_{F}^2 + \|V\|_{F}^{2})$
where $\|...
0
votes
0answers
3 views
Recommendation Model Performance Analysis
Assume I have to recommendation models A and B. I used two evaluation metrics RMSE and F-score. RMSE values of A is lower than B. However when I was evaluating the F-score I noticed the F-score value ...
0
votes
0answers
17 views
Recommendation Algorithm Performance Tracking?
I work for jobs portal which operates very similarly to uber, just not real time.
Employers create jobs. Workers apply for these jobs.
We have a rudimentary recommendation algorithm which sorts ...
0
votes
0answers
53 views
Implicit feedback ALS algorithm: the alpha parameter
I'm creating a recommender system for a video streaming service. My only knowledge about the user preference on a video is the watched percentage of that video. I'm using the implicit feedback ALS ...
0
votes
0answers
29 views
How to predict for test set when training a recommender by decomposing the utility matrix X=UV?
This probably sounds stupid but I don't get the workflow of building a recommending system by the utility matrix: X[i,j] = how much the ith user likes the jth object. For practical issues I refer to ...
