Recommendation Model Based on Semantic Features and a Knowledge Graph

Algorithm 3

Recommendation algorithm based on semantic features and DBpedia.

Input: The user (u) to be recommended, short texts of items;

Output: The prediction rating of the item (i) by the user .

1: For each item i(i∈I), extend short texts of the item based on the DBpedia, get extension feature set W⁽ⁱ⁾.

2: Calculate the item vector of i according to W⁽ⁱ⁾.

3: Construct the user-item rating matrix for dataset.

4: Construct user semantic similarity matrix.

5: Select the set KNN(u) composed of the first K users most similar to u.

6: According to formula (5), calculate prediction rating of i by u (p_ui).