Pack: e-LICO recommender workflows

Description:

 A test file for Collaborative filtering recommender

Description:

No description available

Description:

This process does a random item recommendation; for a given item ID, from the example set of items, it randomly recommends a desired number of items. The purpose of this workflow is to produce a random recommendation baseline for comparison with different recommendation solutions, on different retrieval measures.

The inputs to the process are context defined macros: %{id} defines an item ID for which we would like to obtain recommendation and %{recommender_no} defines the required number of recommendations. The process internally uses an example set of items containing at least item IDs.

The process essentially creates a random subsample of the item IDs and by means of Generate ID operator it creates the score of the recommendation.

The output of the process is an example set consisting of two attributes: recommendation and score of the recommendation.

Description:

This process executes the recommendation based on item to item similarity matrix.

The inputs to the process are context defined macros: %{id} defines an item ID for which we would like to obtain recommendation and %{recommender_no} defines the required number of recommendations. The process internally uses an item to item similarity matrix written in pairwise form (id1, id2, similarity).

The process essentially filters out appearances of the required ID in both of the columns of the pairwise score matrix and then limits the output to the first top %{recommender_no} items based on the score (e.g. similarity) written in the matrix.

The output of the process is an example set consisting of two attributes: recommendation and score of the recommendation.

Description:

This process is a special case of the item to item similarity matrix based recommender where the item to item similarity is calculated as cosine similarity over TF-IDF word vectors obtained from the textual analysis over all the available textual data.

The inputs to the process are context defined macros: %{id} defines an item ID for which we would like to obtain recommendation and %{recommender_no} defines the required number of recommendations. The process internally uses an example set of items containing item ID and an arbitrary amount of textual attributes.

This process essentially selects only textual attributes which are then used as an input for text mining operator, Process Documents from Data. This operator lowers the case of the text, tokenizes it, filters out short and long tokens, filters out stopwords and in the end does stemming based on Porter’s algorithm. The resulting tokens are then filtered for their appearance in the data: tokens appearing in more than 30% or less than 1% are filtered out. The result of the analysis is an example set of TF-IDF word vectors and a bag of words. The bag of words is used to create a TF-IDF vector for the requested item. Afterwards, using the cosine similarity/distance, we calculate the distance between the requested item TF-IDF vector and all other items vectors. First %{recommender_no} items with their distance score are outputted as a final result.

The output of the process is an example set consisting of two attributes: recommendation and score of the recommendation.

Description:

This process executes a collaborative filtering recommender based on user to item score matrix. This recommender predicts one user’s score on some of his non scored items based on similarity with other users.

The inputs to the process are context defined macros: %{id} defines an item ID for which we would like to obtain recommendation and %{recommender_no} defines the required number of recommendations and %{number_of_neighbors} defines the number of the most similar users taken into account while calculating the predicted score. The process internally uses a user to item matrix in a form of an example set where attributes of the matrix denote items; examples represent a single user while the attribute value denotes a user’s score on the item.

This process calculates Euclidean distance of the requested item to all items calculated only over the items for which the requested item has a score inputted. The resulting distance is recalculated into similarity and used as a weight while calculating weighted average of all the scores. The aggregated scores are then transposed to obtain the list of items and their scores and outputted as the final result.

The output of the process is an example set consisting of two attributes: recommendation and score of the recommendation.