Workflows

SVD user-based collaborative filtering rec... (1)

• Download

This workflow takes user-item matrix A as a input. Then it calculates reduced SVD decomposition A_k by taking only k greatest singular values and corresponding singular vectors. This worfkflow calculates recommendations and predictions for particular user %{id} from matrix A. Particular row %{id} is taken from original matrix A and replaced with %{id} row in A_k matrix. Predictions are made for %{id} user based on another users A_k. Note: This workflow uses R-script operator with R library ...

Created: 2011-05-09 | Last updated: 2011-05-09

Credits: Ninoaf Matko Bošnjak

Attributions: User-based collaborative filtering recommender system template Datasets for the pack: RCOMM2011 recommender systems workflow templates

LSI content based recommender system template (1)

• Download

This workflow performs LSI text-mining content based recommendation. We use SVD to capture latent semantics between items and words and to obtain low-dimensional representation of items. Latent Semantic Indexing (LSI) takes k greatest singular values and left and right singular vectors to obtain matrix  A_k=U_k * S_k * V_k^T. Items are represented as word-vectors in the original space, where each row in matrix A represents word-vector of particular item. Matrix U_k, on the other hand ...

Created: 2011-05-06 | Last updated: 2011-05-09

Credits: Ninoaf Matko Bošnjak

Attributions: Content based recommender system template Datasets for the pack: RCOMM2011 recommender systems workflow templates

Simple Recommender System Web Service (1)

• Download

 This workflow takes macro value of user request and computes top-n recommendations from already learned model on train set. Macro value is transformed to example set with the appropriate \textit{user identification} role within \textit{ProcessInput} subprocess operator.   Simple recommendation web service workflow takes three inputs: \begin{enumerate} \item Macro value \textit{user} - identification number of user request \item Train set from RM repository \item Learned Mo...

Created: 2012-05-10 | Last updated: 2012-05-10

Front-end Recommendation Web Service (1)

• Download

 Front-end recommendation web service has a simple job to query the cached recommendations from the \textit{item recommendation table}

Created: 2012-05-11 | Last updated: 2012-05-11

Write activity web service (1)

• Download

 When the specific user $i$ "consumes" certain item $j$ the write activity web service is called, which writes activity $(i,j)$ to \textit{train set table} and removes recommendation $j$ for user $i$ from \textit{item recommendation table} in SQL database.

Created: 2012-05-11 | Last updated: 2012-05-11

Online update recommendation web service (1)

• Download

 After some arbitrary number of recommendations to specific users, system has to update recommendations in \textit{item recommendation table}. This is accomplished by calling the online update recommendation web service, which updates the recommendation model in RapidAnalytics repository and updates the recommendations for specific users in \textit{item recommendation table}.

Created: 2012-05-11 | Last updated: 2012-05-11

Offline update recommendation web service (1)

• Download

No description

Created: 2012-05-11

TextualAttributeExtraction for Recommender... (1)

• Download

This workflow takes conent data from VideoLectures.Net Recommender System Challenge and extractes word-vectors for each lecture. Latent semantic analysis with Singular Value Decomposition is done on item-word binary matrix. The last step is the binomializaiton of dataset.

Created: 2012-06-03

Offline update recommendation web service V2 (1)

• Download

  Periodically we have to do a full re-training on whole train set by the offline update recommendation web service.

Created: 2012-06-20 | Last updated: 2012-06-20

Online update recommendation web service V2 (1)

• Download

 After some arbitrary number of recommendations to specific users, system has to update recommendations in item recommendation table. This is accomplished by calling the online update recommendation web service, which updates the recommendation model in RapidAnalytics repository and updates the recommendations for specific users in item recommendation table.

Created: 2012-06-20 | Last updated: 2012-06-20